Journal No 65

Amfiteatru Economic Journal
The Bucharest University of Economic Studies

Faculty of Business and Tourism
Volume XXVI ●February 2024 ● No. 65
Quarterly publication
Amfiteatru Economic Journal

is recognized and classified category A by
National Council of Scientific Research from Romania
The theme of the next issue
Issue no. 66/2024 – Innovative Application of AI in Business Impacting Socio-

Economic Progress
The Journal is indexed/abstracted by the following international economic

literature services:
Clarivate Analytics (2008)

- Social Sciences Citation Index
- Social Scisearch
- Journal Citation Reports/Social Sciences Edition
EBSCO Publishing (2009)
ProQuest LLC (2012)
DOAJ – Directory of Open Access Journals (2010)
EconLIT – Journal of Economic Literature (2006)
SCOPUS – Elsevier B.V. Bibliographic Databases (2008)
IBSS – International Bibliography of the Social Sciences (2006)
RePEc – Research Papers in Economics (2004)
Cabell’s Directory of Publishing Opportunities (2006)

(Business Directories – Economics and Finance)
ISSN 1582-9146 www.amfiteatrueconomic.ro
Responsibility for the opinions expressed by the articles lies with the authors
Edited with the support of the Ministry of Research, Innovation and Digitization
EDITORIAL BOARD
Editor-in-Chief
Vasile Dinu, The Bucharest University of Economic Studies, Romania; Academy
of Romanian Scientists
Managing Editor
Laurenţiu Tăchiciu, The Bucharest University of Economic Studies, Romania
Associate Editors
Cristina Circa, West University of Timișoara, Romania
Dan-Cristian Dabija, ”Babeş-Bolyai” University, Cluj-Napoca, Romania
Violeta Mihaela Dincă, The Bucharest University of Economic Studies, Romania
Cristian-Mihai Dragoş, ”Babeş-Bolyai” University, Cluj-Napoca, Romania
Madălina Dumitru, The Bucharest University of Economic Studies, Romania
Valentin Dumitru, The Bucharest University of Economic Studies, Romania
Raluca-Gina Gușe, The Bucharest University of Economic Studies, Romania
Romero Isidoro, University of Seville, Spain
Borut Jereb, University of Maribor, Sloven
Octavian-Dragomir Jora, The Bucharest University of Economic Studies, Romania
Nicolae Lupu, The Bucharest University of Economic Studies, Romania
Čudanov Mladen, University of Belgrade, Belgrade, Serbia
Judit Oláh, University of Debrecen, Hungary
Corina Pelau, The Bucharest University of Economic Studies, Romania
Cătălina Soriana Sitnikov, University of Craiova, Romania
George-Sorin Toma, University Bucharest, Romania
Aharon Tziner, The Academic College of Netanya, Netanya, Israel
Cristinel Vasiliu, The Bucharest University of Economic Studies, Romania
Milena-Rodica Zaharia, The Bucharest University of Economic Studies, Romania
Editorial Office
Irina Albăstroiu, The Bucharest University of Economic Studies, Romania
Alexandra Barbu, The Bucharest University of Economic Studies, Romania
Mihaela Bucur, The Bucharest University of Economic Studies, Romania
Simona Margareta Bușoi, ASE Publishing House, The Bucharest University
of Economic Studies, Romania
Răzvan Dina, The Bucharest University of Economic Studies, Romania
Raluca Mariana Grosu (Assistant Editor), The Bucharest University of Economic
Studies, Romania
Silvia Răcaru, ASE Publishing House, The Bucharest University of Economic Studies,
Romania
Violeta Rogojan, ASE Publishing House, The Bucharest University of Economic
Studies, Romania
Daniel-Ion Zgură, The Bucharest University of Economic Studies, Romania
Founders
Vasile Dinu, The Bucharest University of Economic Studies, Romania
Sandu Costache, The Bucharest University of Economic Studies, Romania
Editorial Advisory Board

Andrej Bertoncelj, University of Primorska, Koper, Slovenia
Yuriy Bilan, University of Szczecin, Szczecin, Poland
Slobodan Čerović, University Singidunum, Belgrade, Serbia
Ung-il Chung, The University of Tokyo, Tokyo, Japan
Lóránt Dénes Dávid, Szent István University, Gödöllő, Hungary; ordinary member
of the European Academy of Sciences and Arts
Emilian Dobrescu, Romanian Academy, Bucharest, Romania
Valeriu Ioan-Franc, Romanian Academy, Bucharest, Romania
Romualdas Ginevicius, Vilnius Gediminas Technical University, Vilnius, Lithuania
Nicolae Istudor, The Bucharest University of Economic Studies, Romania
Delgado Francisco Jose, University of Oviedo, Spain
Dumitru Miron, The Bucharest University of Economic Studies, Romania
Puiu Nistoreanu, The Bucharest University of Economic Studies, Romania
Bogdan Onete, The Bucharest University of Economic Studies, Romania
Rodica Pamfilie, The Bucharest University of Economic Studies, Romania
József Popp, Corresponding Member of the Hungarian Academy of Sciences; John von
Neumann University, Hungarian National Bank
Roxana Sârbu, The Bucharest University of Economic Studies, Romania
Dalia Streimikiene, Vilnius University, Vilnius, Lithuania
Bernhard Swoboda, University of Trier, Germany
Gabriela Țigu, The Bucharest University of Economic Studies, Romania
Contents
Challenges for Competence-Oriented Education in the Context of the Development
of Artificial Intelligence Systems......................................................................................... 6
Rodica Milena Zaharia

of Artificial Intelligence Systems
The Development of Educational Competences for Romanian Students
in the Context of the Evolution of Data Science and Artificial Intelligence .................. 14
Giani Ionel Grădinaru, Vasile Dinu, Cătălin-Laurențiu Rotaru and Andreea Toma
Identifying Sufficient and Necessary Competencies in the Effective Use

of Artificial Intelligence Technologies .............................................................................. 33
Ion Popa, Marian Mihai Cioc, Andreea Breazu and Cătălina Florentina Popa
Challenges of Artificial Intelligence on the Learning Process in Higher Education .... 53

Nicoleta Sîrghi, Mirela-Catrinel Voicu, Grațiela Georgiana Noja
and Oana-Ramona Socoliuc (Guriță)
Exploring University Students’ Perceptions of Generative Artificial Intelligence

in Education ........................................................................................................................ 71
Mădălina Lavinia Țală, Cătălina Nicoleta Müller, Irina Albăstroiu Năstase,
Olimpia State and Georgică Gheorghe
The Social Impact of Using Artificial Intelligence in Education .................................... 89

Andreea Simona Săseanu, Rodica-Manuela Gogonea and Simona Ioana Ghiţă
Exploring the Antecedents of Artificial Intelligence Products’ Usage.

The Case of Business Students ........................................................................................ 106
Rodney Duffett, Rodica Milena Zaharia, Tudor Edu, Raluca Constantinescu
and Costel Negricea
Artificial Intelligence Adoption in the Workplace and Its Impact

on the Upskilling and Reskilling Strategies.................................................................... 126
Constanța-Nicoleta Bodea, Mario Paparic, Radu Ioan Mogoș and Maria-Iuliana Dascălu
Artificial Intelligence in Education: Next-Gen Teacher Perspectives ......................... 145

Nicoleta Sămărescu, Robert Bumbac, Alin Zamfiroiu and Maria-Cristina Iorgulescu
Generation Z Students’ Perceptions on the Abilities, Skills and Competencies

Required in the Age of Artificial Intelligence Systems ................................................. 162
Sorin George Toma and Oana Simona Hudea
Artificial Intelligence and the Modelling of Teachers’ Competencies ......................... 181

Ramona Simuț, Ciprian Simuț, Daniel Bădulescu and Alina Bădulescu
4 Amfiteatru Economic
A Systematic Analysis of New Approaches to Digital Economic Education Based
on the Use of AI Technologies ......................................................................................... 201
Laurențiu Mihai, Leonardo-Geo Mănescu, Laura Vasilescu, Anca Băndoi
and Cătălina Sitnikov
Artificial Intelligence and Competency-Based Education: A Bibliometric Analysis . 220

Cătălina Radu, Carmen Nadia Ciocoiu, Cristina Veith and Răzvan Cătălin Dobrea
Artificial Intelligence in Business Education: Benefits and Tools ................................ 241

Camelia Surugiu, Cătălin Grădinaru and Marius - Răzvan Surugiu
Quantitative Evaluation of Willingness to Use Artificial Intelligence

within Business and Economic Academic Environment ............................................... 259
Daniela Șerban, Silvia Elena Cristache, Narcisa Georgeta Ciobotar,
Laurențiu Gabriel Frâncu and Jiries Mansou
Redesign of Accounting Education to Meet the Challenges of Artificial

Intelligence – A Literature Review ................................................................................. 275
Valeriu Brabete, Cătălin Mihail Barbu, Daniel Cîrciumaru, Daniel Goagără
and Dorel Berceanu
Students’ Perceptions of the Use of Artificial Intelligence in Educational Service .... 294
Ines Djokic, Nikola Milicevic, Nenad Djokic, Borka Malcic and Branimir Kalas
Economic Interferences
Measuring the Level of Performance of Country’s Cultural Economics:
The Case of EU11 ............................................................................................................. 312
Kristina Astikė and Viktorija Skvarciany
Research on Tax Compliance Incentive Effects of Platform Companies

from the Perspective of Incomplete Contract – An Empirical Study Based
on China ............................................................................................................................ 330
Xuefeng Shao and Shi Chen
Amfiteatru Economic recommends

Cultured Meat – From Scientific Challenge to Consumer Acceptance
as Sustainable Food Source ............................................................................................. 346
Lelia Voinea, Dorin Vicențiu Popescu, Teodor Mihai Negrea and Răzvan Dina
Architecture to Transform Classic Academic Courses into Adaptive Learning

Flows with Artificial Intelligence .................................................................................... 363
Andrei Bobocea, Răzvan Bologa, Lorena Bătăgan and Bogdan-Ștefan Posedaru
Victor Tufescu or the Perfect Elegance of Man and Spirit ........................................... 381
Silviu Neguţ
Vol. 26 • No. 65• February 2024 5

Challenges for Competence-Oriented Education in the Context
of the Development of Artificial Intelligence Systems
Please cite this article as:

Zaharia, R. M., 2023. Challenges for Competence-Oriented Education in the Context of the
Development of Artificial Intelligence systems. Amfiteatru Economic, 26(65), pp. 6-11.
DOI: https://doi.org/10.24818/EA/2024/65/6
The debates about the challenges launched to the educational process by Artificial
Intelligence systems are more acute than ever. On one hand, education, in general, must be
that process that ensures the skills that allow the creation of Artificial Intelligence.
Algorithms, chatbots, learning machines were created as a result of knowledge, skills and
abilities acquired through education. They will continue to be the result of the creativity,
inventiveness and daring of those who are educated in this spirit, of those who are
cultivated with these skills. “Working” with Artificial Intelligence systems also requires
knowledge, skills and abilities that the entire educational system must provide. Operating
with Artificial Intelligence systems needs not only specific skills, but also their permanent
adaptation as Artificial Intelligence systems become more complex and more performant.
Academic research is intended not only to identify the challenges that education faces and
will face with the development of Artificial Intelligence systems, but also to find solutions,
to propose measures to manage the consequences of the penetration of Artificial
Intelligence into life of society.
The importance of education seems even more evident in ensuring the balance between the
advantages and disadvantages generated by Artificial Intelligence. Artificial Intelligence is
a tool that education, as a complex process of developing and substantiating knowledge,
skills and abilities, can use in building critical thinking, in accelerating scientific
discoveries, in strengthening the ethical spirit. The temptations generated by Artificial
Intelligence, which can lead to the opposite of this fundamental goal, can be corrected
through the educational process. Ultimately, we must be careful that through education,
Artificial Intelligence does not turn from a blessing into a curse.
In the Amfiteatru Economic Journal, no. 65/2024, you will find articles that respond to the
challenges of education in the context of Artificial Intelligence and that improve the
specialized literature in the field, such as:
The article The development of educational competences for Romanian students in the
context of the evolution of data science and Artificial Intelligence explores key academic
competencies and professional skills in data science in the context of the development of
Artificial Intelligence, highlighting their importance in the business environment. Using the
“2022 Stack Overflow Annual Developer Survey” dataset and machine learning methods
such as principal component analysis, K-means clustering, and logistic regression, this
study looks at professional skills in data science. This research looks at the distribution of
jobs in the field, the level of experience, the languages and analysis programs used, the
support offered by companies and the dynamics of data science teams, as well as at the
impact that Artificial Intelligence is having on the field, providing a comprehensive
understanding of the impact of academic training on career opportunities in data science,
contributing to the development of the profile of a qualified specialist in this field. The
research also provides relevant directions and recommendations for enhancing the skills
required in data science, in order to outline a skilled profile and to meet the demands of the
business environment in a world dominated by data analytics and Artificial Intelligence.
Starting from the idea that technological changes have also created challenges, including a
gap between the skills available and the skills needed in the use of Artificial Intelligence
technologies, the article Identifying sufficient and necessary competencies in the effective
use of Artificial Intelligence technologies aims to analyze the relationships between
employee skills and effectiveness in using Artificial Intelligence tools, to highlight the set
of skills essential to effective interaction with Artificial Intelligence technology. An online
questionnaire was completed by 209 employees from Romania, the data analysis being
carried out on the basis of two advanced techniques: structural equation modeling and
analysis of necessary conditions, using the SmartPLS v4 program. The results suggest a
significant association between employee competencies and the effectiveness of using
Artificial Intelligence tools, optimism and inventiveness positively mediating this
relationship. The article outlines the significant implications for organizations, for the
formal and non-formal education system, as well as research directions on the managerial
implications of using Artificial Intelligence tools.
The paper Challenges of Artificial Intelligence on the learning process in Higher Education
had as its objective the identification of the skills necessary for the digital learning
environment specific to Romanian higher education, skills intended to encourage the
adoption of Artificial Intelligence technologies by students as beneficiaries of the
educational act. Using the structural equation model applied to an original set of data
collected on the basis of a questionnaire addressed to undergraduate students in higher
economic education, the research results emphasize that the intention to adopt applications
that use Artificial Intelligence, among students, directly depends of constructs such as:
perceived usefulness, attitude towards these technologies, perceived hedonic value,
expected performance, or degree of compatibility, while the interactivity of the applications
also has an important but indirect influence.
The research Exploring university students’ perceptions of generative Artificial Intelligence
in education focused on a certain form of Artificial Intelligence, namely the generative one,
which facilitates the acquisition (generation) of content in all its forms (text, image, video,
audio, programming codes, etc.), thus having a crucial role in education, enabling the
personalization of educational content and facilitating the learning process. The article
starts from the conceptual delimitations regarding Artificial Intelligence and the ways of
using it in education, its advantages and limits, showing that the adoption of generative
Artificial Intelligence applications, such as ChatGPT, in economic higher education, is a
topic too little addressed in the literature. Then, adopting exploratory research, carried out
on a sample of 364 undergraduate and master students of the Faculty of Business and
Tourism within the Bucharest University of Economic Studies, their perception regarding
these applications was analyzed. The research results indicated a high level of awareness
and interest in content generation models and highlighted the fact that users who have
favorable perceptions of the quality of content generated by such applications believe that
their integration into academic endeavors can stimulate creativity and can improve
employment prospects.
Vol. 26 • No. 65 • February 2024 7

The social impact of Artificial Intelligence was the starting point of the article The social
impact of using Artificial Intelligence in education which, starting from the data collected
through a survey from the students of some prestigious universities in Romania and
processed by applying three statistical-econometric logistic regression models, revealed
interesting results. The results of the first binary logistic model reflect the respondents'
views on the need and on the importance of improving the learning experience through the
use of Artificial Intelligence in education, taking into account their gender and their level of
education. Also, in relation to the two characteristics considered the most significant for the
research objective, the following multinominal logistic models were developed. The
findings show that the use of Artificial Intelligence in education influences, on the one
hand, graduates' employment prospects and, on the other hand, society as a whole.
Starting from a theory increasingly used in the literature, the Unified Theory of Acceptance
and Use of Technology (UTAUT), the article Exploring the antecedents of Artificial
Intelligence Products’ usage. The case of business students investigates the extent to which
extent facilitating conditions (those means that users consider necessary to use for a certain
technology) and other predictors (perceived risk and lack of trust in technology, gender,
education, income, technology proficiency and equipment used to access the Internet)
influence the use of Artificial Intelligence Products (AIPs) in general and for educational
purposes. Data from a sample of 450 Romanian business students were examined using
principal component analysis and logistic regression. Facilitating conditions showed a
direct (positive) effect on the dependent variables, and the combination “perceived risk-
perceived lack of trust in technology” showed an inverse (negative) effect on the dependent
variable. This study advances theory by exploring the actual use of AIPs for educational
purposes, developing the UTAUT model by isolating facilitating conditions and using
descriptive variables as predictors. Also, the study expands the empirical evidence related
to UTAUT regarding the acceptance and use of technology in Romania and formulates
practical recommendations for universities, as current and potential providers of AIPs, to
make the educational process more efficient.
The article Artificial Intelligence adoption in the workplace and its impact on the upskilling
and reskilling strategies presents the results of an ethnographic research aimed at
identifying the impact of the adoption of Artificial Intelligence in the workplace on the
requirements related to professional knowledge and skills and, based on these, the impact
on the upskilling and reskilling strategies. The research covers three areas of activity,
namely: information technology, education and scientific research. A relevant conclusion of
this research is the need to study the competence requirements from multiple perspectives,
not only from the perspective of technological innovation. The research originality mainly
consists in the way in which the concept of the level of upskilling/reskilling importance is
defined and applied, based on professional knowledge and skills development
requirements. By using the assessed level of upskilling/reskilling importance, strategies and
related actions may be defined and undertaken.
The role of educators in shaping skills, in the context of the expansion of Artificial
Intelligence, is particularly important. Three studies focus on this aspect: (i) AI in
education: next-gen teacher perspectives which mentions the fact that the progress made by
Artificial Intelligence at the world level leads UNESCO, through the Beijing Consensus, to
recommend in 2019 to governments the inclusion of Artificial Intelligence in educational
policies and processes. The research analyzes the factors that influence the behavioral
intention to use Artificial Intelligence from the training stage of future from the training
stage of future primary and secondary teachers in Romania. Through quantitative
exploratory research, carried out on a sample of 270 students from the Faculty of
Education, Social Sciences and Psychology, the subjects' interaction with Artificial
Intelligence and the intention to integrate Artificial Intelligence in education were
investigated, using binary logistic regression. The analysis shows that, among the six
variables of the model, “confidence in one's ability to use Artificial Intelligence” and
“perception of a greater number of advantages” have a positive and significant impact on
the willingness to use Artificial Intelligence in the educational process, more than 'previous
use', 'level of knowledge' or 'student requirements'. These results are of particular
importance for the revision of teacher education programs and the development of
educational policies that increase future teachers' confidence in the ability to use Artificial
Intelligence, eliminating fears or misconceptions about Artificial Intelligence in education.
(ii) Artificial Intelligence and the modelling of teachers’ competencies explores educators'
perspective on their own role in shaping skills and presents educators' perceived challenges
and key measures in the context of expanding Artificial Intelligence. Thus, teachers'
positive attitudes toward Artificial Intelligence significantly influence cognitive,
fundamental, and educational management competencies. The research highlights key
challenges to integrate Artificial Intelligence into education, including the imperative of
professional development for educators and ensuring equitable access to educational
resources and technology. The study advocates for initiatives to bridge the digital divide
and integrate Artificial Intelligence education into school curricula. (iii) Quantitative
evaluation of willingness to use Artificial Intelligence within business and economic
academic environment analyzes the state of information, use and availability of use of
Artificial Intelligence in the economic and business university environment, according to
Romanian teaching staff opinions. The research aims to identify the advantages,
disadvantages and how Artificial Intelligence is used on the teachers’ personal initiative in
research, teaching, and evaluation activities. The results of the study identify the aspects
that can optimize the processes of education - research, teaching, evaluation and learning to
meet the increased dynamics of the use of Artificial Intelligence in the economic academic
environment in Romania. Also, the advantages associated with the use of Artificial
Intelligence systems and the solutions proposed to maximize the benefits brought by
Artificial Intelligence in research, teaching, evaluation activities in the opinion of teachers
are highlighted.
Two other articles investigate aspects of student perceptions of the use of Artificial
Intelligence. (i) Generation z students’ perceptions on the abilities, skills and competencies
required in the age of Artificial Intelligence systems analyzes one of the most challenging
generations in relation to the use of new technologies. A total of 352 questionnaires were
validated from students who graduated from full-time bachelor's programs in a Romanian
faculty. The data were processed using the statistical program SPSS 17.0, the viability tests
and subsequent analyzes involving Cronbach Alpha coefficient, principal components
analysis, respectively correlational analysis. The conclusions of the article offer a new
perspective on the structure of skills, abilities and competencies needed in the digital
society and propose solutions to improve the educational process, by adapting academic
programs to the demands of the digital society. (ii) Students’ perceptions of the use of
Artificial Intelligence in educational services investigates Serbian students' conceptions of
the use of new technologies in education. The research represents the first implementation
of an eight-factor instrument, modeled as a reflexive-formative hierarchical construct. Thus,
all eight factors contribute positively and significantly to the use of Artificial Intelligence in
Vol. 26 • No. 65 • February 2024 9

education from the students' perspective. The biggest contribution could be attributed to
factors stemming from the weaknesses of traditional education, such as: sentiment analysis
in education, personalized learning and student performance prediction, and the smallest
contribution could be attributed to factors related to privacy concerns with on the use of
Artificial Intelligence and control: classroom monitoring, visual analysis, and student
grading and assessment.
Using the same research methodology, namely the bibliometric analysis, three other articles
carry out a review of the specialized academic literature for different objectives. (i) A
systematic analysis of new approaches to digital economic education based on the use of AI
technologies presents a systematic evaluation of a number of 60 scientific articles, with the
aim of studying how to approach, in recent years, the specific concepts of digital economic
education, based on Artificial Intelligence, how to use Artificial Intelligence applications in
digital economic education, all these with the aim of identifying critical success factors and
challenges facing this field. The results obtained highlighted the trend according to which
most researchers define digital education as a form of technology use with the aim of
building the necessary support for educational activities, positioning Artificial Intelligence
and its various applications as essential elements of current digital education, with
fundamental potential in revolutionizing economic processes. It has been noted that the
widespread use of digital education platforms based on Artificial Intelligence is influenced
by the technology, its superior characteristics from the perspective of the ability to
coherently correlate teaching/learning processes, the confidence generated at the level of
teachers and learners regarding the results obtained as a result of the adoption of these
digital education platforms based on Artificial Intelligence and cultural factors, in parallel
with a permanent confrontation with numerous challenges related to the resistance of users
to change, digital skills, accessibility of systems, as well as financial issues. Based on this
research endeavour, a model system of correlations and elements has been developed,
specifically for the digital economic education based on Artificial Intelligence, bringing
together both success factors and specific challenges. (ii) Artificial Intelligence and
competency-based education: a bibliometric analysis set out to identify emerging trends,
challenges and opportunities generated by the intersection of Artificial Intelligence and
Competency-Based Education. The research was carried out through the bibliometric
analysis of 1028 articles included in the Web of Science database and based on the reports
provided by the biblioshiny application, the graphical interface of the bibliometrix R
package. The results included quantitative analysis of scientific production, collaborations,
co-citations, as well as the evolution and thematic map of the field. The research revealed
an annual growth of 8.43% in publications, accelerating after 2017, and global engagement,
with the US and China leading the way. The thematic analyzes showed the evolution of the
field, from technological foundations to an interdisciplinary approach, highlighting
including the influences of global events, such as COVID-19. The study confirmed the deep
interaction between Artificial Intelligence and Competency-Based Education,
demonstrating the potential, complexity and need for collaborative and interdisciplinary
approaches. (iii) Redesign of accounting education to meet the challenges of artificial
intelligence - a literature review underlines the impact of artificial intelligence on the
accounting profession, in the sense of redefining the role and place that accounting
specialists occupy within economic entities. The systematic literature review aims to
highlight and summarize the main challenges regarding the impact that artificial
intelligence and other related technologies have on accounting professionals and the skills
they need to develop or acquire. At the same time, the aim was to outline concrete solutions
through which accounting professionals can be supported in their efforts to adapt to the new
requirements of the labor market imposed by the evolution of disruptive technologies.
The paper Artificial Intelligence in business education: benefits and tools looks at how
Artificial Intelligence can support educational activities, the key drivers and tools used in
business education. Survey data, collected from 254 learners, were analyzed using
multivariate binary logistic regression. Two research questions were formulated to verify
whether Artificial Intelligence supports educational activities and to identify Artificial
Intelligence tools that support educational activities for business. The results show that
respondents appreciate artificial intelligence because it helps teachers with administrative
tasks, allows for personalization of learning plans and saves time. However, learners are not
familiar with most of the benefits offered by artificial intelligence tools, with the exception
of computer vision, edge computing, and chatbots. The paper emphasizes the importance of
expanding the use of artificial intelligence in education, with the aim of familiarizing
students with these technologies and exploiting technology’s potential in the context of
business education.
We hope that all these articles will increase the interest of both academics and practitioners,
as students, also, in finding answers for their questions regarding the role of Artificial
Intelligence in education. Also, we are optimistic that these articles will inspire for further
researches, as the topic of Artificial Intelligence in relationship to education seems to be at
the beginning of the journey.
Associate Editor,
Rodica Milena Zaharia
Vol. 26 • No. 65 • February 2024 11

Contents
The Development of Educational Competences for Romanian Students
in the Context of the Evolution of Data Science and Artificial Intelligence .................. 14
Giani Ionel Grădinaru, Vasile Dinu, Cătălin-Laurențiu Rotaru and Andreea Toma
Identifying Sufficient and Necessary Competencies in the Effective Use

of Artificial Intelligence Technologies .............................................................................. 33
Ion Popa, Marian Mihai Cioc, Andreea Breazu and Cătălina Florentina Popa
Challenges of Artificial Intelligence on the Learning Process in Higher Education .... 53

Nicoleta Sîrghi, Mirela-Catrinel Voicu, Grațiela Georgiana Noja
and Oana-Ramona Socoliuc (Guriță)
Exploring University Students’ Perceptions of Generative Artificial Intelligence

in Education ........................................................................................................................ 71
Mădălina Lavinia Țală, Cătălina Nicoleta Müller, Irina Albăstroiu Năstase,
Olimpia State and Georgică Gheorghe
The Social Impact of Using Artificial Intelligence in Education .................................... 89

Andreea Simona Săseanu, Rodica-Manuela Gogonea and Simona Ioana Ghiţă
Exploring the Antecedents of Artificial Intelligence Products’ Usage.

The Case of Business Students ........................................................................................ 106
Rodney Duffett, Rodica Milena Zaharia, Tudor Edu, Raluca Constantinescu
and Costel Negricea
Artificial Intelligence Adoption in the Workplace and Its Impact

on the Upskilling and Reskilling Strategies.................................................................... 126
Constanța-Nicoleta Bodea, Mario Paparic, Radu Ioan Mogoș and Maria-Iuliana Dascălu
Artificial Intelligence in Education: Next-Gen Teacher Perspectives ......................... 145

Nicoleta Sămărescu, Robert Bumbac, Alin Zamfiroiu and Maria-Cristina Iorgulescu
Generation Z Students’ Perceptions on the Abilities, Skills and Competencies

Required in the Age of Artificial Intelligence Systems ................................................. 162
Sorin George Toma and Oana Simona Hudea
Artificial Intelligence and the Modelling of Teachers’ Competencies ......................... 181

Ramona Simuț, Ciprian Simuț, Daniel Bădulescu and Alina Bădulescu
A Systematic Analysis of New Approaches to Digital Economic Education Based

on the Use of AI Technologies ......................................................................................... 201
Laurențiu Mihai, Leonardo-Geo Mănescu, Laura Vasilescu, Anca Băndoi
and Cătălina Sitnikov
Artificial Intelligence and Competency-Based Education: A Bibliometric Analysis . 220
Cătălina Radu, Carmen Nadia Ciocoiu, Cristina Veith and Răzvan Cătălin Dobrea
Artificial Intelligence in Business Education: Benefits and Tools................................ 241

Camelia Surugiu, Cătălin Grădinaru and Marius - Răzvan Surugiu
Quantitative Evaluation of Willingness to Use Artificial Intelligence

within Business and Economic Academic Environment ............................................... 259
Daniela Șerban, Silvia Elena Cristache, Narcisa Georgeta Ciobotar,
Laurențiu Gabriel Frâncu and Jiries Mansou
Redesign of Accounting Education to Meet the Challenges of Artificial

Intelligence – A Literature Review ................................................................................. 275
Valeriu Brabete, Cătălin Mihail Barbu, Daniel Cîrciumaru, Daniel Goagără
and Dorel Berceanu
Students’ Perceptions of the Use of Artificial Intelligence in Educational Service .... 294
Ines Djokic, Nikola Milicevic, Nenad Djokic, Borka Malcic and Branimir Kalas
Vol. 26 • No. 65• February 2024 13

AE The Development of Educational Competences for Romanian Students
in the Context of the Evolution of Data Science and Artificial Intelligence
THE DEVELOPMENT OF EDUCATIONAL COMPETENCES FOR ROMANIAN

STUDENTS IN THE CONTEXT OF THE EVOLUTION OF DATA SCIENCE
AND ARTIFICIAL INTELLIGENCE
Giani Ionel Grădinaru1 , Vasile Dinu2 , Cătălin-Laurențiu Rotaru 3
and Andreea Toma4
1)
Bucharest University of Economic Studies, Bucharest, Romania
and Institute of National Economy, Romanian Academy,
Bucharest, Romania
2)
Bucharest University of Economic Studies, Bucharest, Romanian and Academy
of Scientists, Bucharest, Romania
3)4)
Bucharest University of Economic Studies, Bucharest
Please cite this article as: Article History

Grădinaru, G.I., Dinu, V., Rotaru, C.L. and Toma, A., Received: 17 September 2023
2024. The Development of Educational Competences Revised: 6 November 2023
for Romanian Students in the Context of the Accepted: 7 December 2023
Evolution of Data Science and Artificial Intelligence.
Amfiteatru Economic, 26(65), pp. 14-32.
Abstract
The study explores key academic competencies and professional skills in data science in the
context of the development of artificial intelligence, highlighting their importance in the
business environment. Using the “2022 Stack Overflow Annual Developer Survey” dataset
and machine learning methods such as principal component analysis, K-means clustering,
and logistic regression, professional skills in science are analysed the data. The research
targets the distribution of jobs in the field, the level of experience, the languages and
analysis programs used, the support offered by companies, and the dynamics of data
science teams, as well as the impact that artificial intelligence has on the field. With their
help, a comprehensive understanding of the impact of academic training on career
opportunities in the field of data science is provided, contributing to the development of the
profile of the qualified specialist in this field. The research also provides relevant pointers
and recommendations for enhancing the skills required in data science in order to outline a
skilled profile and fulfil the demands of the business environment in a world dominated by
data analytics and artificial intelligence. By including academic skills in the process of
 Corresponding author: Giani Ionel Grădinaru – e-mail: [email protected]
This is an Open Access article distributed under the terms of the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any
medium, provided the original work is properly cited. © 2024 The Author(s).
Challenges for Competence-Oriented Education in the Context AE
training data science specialists, the research brings innovation and highlights the skills
needed to be trained in the academic field to facilitate the employment of graduates in
specific fields of data science. This aspect is significant because, in practice, it has been
observed that most specialists working in data science rely on independent learning rather
than skills acquired in the academic field.
Keywords: data science, artificial intelligence, academic skills, professional skills.
JEL Classification: I23, J23, C49, M15.
Introduction
Data science has become one of the most exciting fields in the age of technology and
artificial intelligence (AI). With the advancement of technology and the increase in the
amount of data available, the job market has experienced a significant increase in the
number of data science jobs. Universities have recognised the importance of this field and
started building specialised programs to train future data scientists. At the same time,
specialised publications constantly and actively write about this topic, highlighting the
progress and new trends in the field.
According to Smaldone et al. (2022), the role of a data scientist usually involves the
identification of business directions, providing decision support for the requirements and
needs of the business environment. Thus, communication skills have a critical role, being
necessary to mediate the language used starting from the technical analysis to the
information transmitted to the business, management team, whether we refer to an
intelligent data visualisation or not. Curiosity and creativity are key skills, because among
large data sets, a data science specialist has to search, research, test, and draw various
hypotheses that will be the basis of the proposed solutions.
In the context of the development of artificial intelligence and data science, academic skills
play an essential role in the advancement of these fields. Both AI and data science are based
on the foundation of academic research and education to develop innovative solutions and
understand the complexity of these technologies (Leal et al., 2023).
In the context of data science and artificial intelligence, academic skills are critical to
collect, process and interpret data. Data scientists must understand the mathematics and
statistics behind analytics and machine learning techniques to create accurate and robust
models. As stated by Irizarry (2020) developing effective AI algorithms requires a solid
knowledge base in programming, mathematics, data analysis, and machine learning.
Understanding fundamental AI concepts such as neural networks, deep learning, and
machine learning algorithms enable researchers to create and optimise advanced AI models.
In parallel with the technical aspects, academic skills are closely related to the ethical and
social aspects of AI and data science (Chen et al., 2020). Professionals in these fields must
be aware of the impact that developed technologies can have on society, as a result, it is
necessary to consider privacy issues and identify possible risks and biases. They must
approach the development of AI technologies responsibly and work to ensure ethical and
fair use of these technologies.
Vol. 26 • No. 65 • February 2024 15

Specialist training programs, courses, and academic programs have evolved simultaneously
with the labour market, but the branches of the field are diverse. Whether referring to data
engineers, who ensure the proper flow of raw data needed for analysis, to data analysts,
who deal with identifying data sources, cleansing, and quality assurance of the set, as well
as intelligently visualising the results, or to specialists in data science, which builds
predictive models, they represent roles that are part of the same field. However, each role
requires rigorous training to become a specialist in the analytics process, from data
accumulation, cleaning, processing, modelling, interpretation, or even prediction. In this
context, most of the time training starts in academic programs (Irizarry, 2020). It is
highlighted that the connection between the professional and the academic environment is
very important, because in order to be able to occupy the position of specialist in data
science in a certain field, it is necessary to prepare and develop the necessary skills.
Wing (2019) states that data science appears to be a vast field defined as a set of
fundamental principles that guide the extraction and understanding of data. It is often
associated with terms like BD (Big Data), data mining, and AI. To fully understand this
field, it is necessary to identify the underlying principles and understand what it offers. This
process provides clarity in data science requirements for both academic and professional
environments.
The aim of the current research is to expose the vital link between the educational
competencies and the professional skills required in the field of data science, in the context
of the development of AI with the help of quantitative methods used in machine learning
(ML - Machine Learning, a branch of AI), such as component analysis main methods, K-
means Clustering (KM) and logistic Regression (RL), and to determine the profile of the
specialist who can be considered qualified or unqualified in the field of data science.
In the first phase, the research focuses on the professional skills sought in the field of data
science, including the technical and analytical skills needed to become a specialist in this
field and to meet the demands of the business environment. It will then identify the typical
profile of a data scientist and highlight its importance in the business environment by
exploring the role, responsibilities, and added value it brings to decision-making and the
development of solutions based on data analysis. Finally, the research focuses on
identifying the essential academic skills to become a data science specialist, by analysing
academic programs, courses, and relevant study subjects that contribute to the development
of these skills, such as programming, data management, and artificial intelligence courses.
The present research aims to explore and highlight the key professional and educational
competencies in data science, alongside their importance in the business environment,
providing a comprehensive understanding of the impact of academic training on career
opportunities in this field, as well as supporting the importance of AI as a key asset in the
development of the science profile of the data.
1. Review of the scientific literature

The use of statistical methods and models plays a crucial role in the education sector, by
revolutionising the way this field identifies the skills and knowledge that must be
introduced into the education curriculum for the training of data science specialists.
Through the use of these advanced techniques, the need to align universities with the
demands of the data science market has been highlighted.
Educational and research practices can be shaped by the AI – ML couple in the future,
leading to improvements. As stated by Alqahtani et al. (2023) in research, key applications
include text generation, data analysis and interpretation, literature review, formatting, and
editing. In the education sector, the AI – ML couple can support personalised learning,
assessment, specific learning plans, personalised career guidance, and mental health
support. Despite the advantages, the responsible use of these technologies must address
ethical and algorithmic imbalance issues. According to Chassignol et al. (2018), AI is
revolutionising education by providing personalised learning solutions, adapting teaching
methods through the emergence of individual guidance and effective assessment solutions.
Although AI-powered platforms personalise the content, concerns about AI dominance still
linger. However, researchers still draw attention to the importance of maintaining human
mentoring and social interaction in education, a fact also emphasised by Tuba et al. (2023),
who claims that AI tools do not contribute to the full development of academic skills, not
yet having the ability to systematise information. However, AI can influence the future
roles of humans and their ways of doing business. Pelău, Ene and Pop (2021) argues that
the development of the AI – ML couple represents one of the main paradigms of
contemporary society, having a significant impact on individual life and society as a whole,
as well as on the economy. The use of AI in people’s daily activities and in the interaction
between companies and consumers brings many advantages, such as increased efficiency
and an engaging interaction. However, there are also concerns about the future
development of this field. Amidst its ability to store a large amount of data about
individuals’ behaviour and process this data quickly, there is a risk that forms of AI will
become smarter than humans and influence their decisions.
Particularly interesting for scientific progress and industry evolution is the rapid transition
of the AI – ML couple from a narrow area of research in a limited number of academic
laboratories to a key topic in the business world. This process is accompanied by the
emergence of other new paradigms, such as BD and the Internet of Things (IoT). In
addition to these, a new discipline called data science gradually took center stage as the
main branch of knowledge, covering all relevant methods and work processes to translate
data into effective business solutions. Learning the main principles and capabilities of this
discipline becomes a new academic and business challenge. Another cause is represented
by the emergence of data specialists (specialist in data science), who have become
motivated factors for this transformation (Kordon, 2020).
The term data science was introduced two decades ago, bringing new topics of discussion
and analysis in various fields (Cao et al., 2016). According to Grossi et al. (2021), data
science is an important sector with applications in business, industry, and education,
encompassing areas such as AI and ML, experimental design, and data-driven modelling.
Moreover, statistics becomes an essential discipline, providing tools and methods in the
process of identifying structure, providing an overview of the field of data science. In this
context, obtaining reliable results, developing predictive models, and understanding data
structures are conferred by statistics. The idea that statistics, along with ML, plays a
fundamental role in data science is reiterated by a group of leaders of the American
Statistical Association, which issued a statement on the special powers of these disciplines,
highlighting as central elements of the current field. Dyk et al. (2015) says that, data
Vol. 26 • No. 65 • February 2024 17

science is a set of fundamental principles that support and guide the extraction of
information and knowledge from data. A concept closely related to this field is data mining,
the effective extraction of knowledge from data through technologies that integrate
proprietary principles. The applicability of data science is varied, distinguishing itself in
business, marketing, finance, sales, and more. Therefore, a successful data scientist must be
able to approach business problems from a data-driven perspective and understand the
fundamental principles of data analysis (Provost and Fawcett, 2013).
It is necessary to start from the basics of statistics and pay special attention to the evolution
of data science, in order to achieve an exhaustive presentation of this field (Donoho et al.
2015). In the context of AI development, data are today considered valuable assets, and
data mining has become a very important factor in improving the competitiveness of
enterprises. This rapidly growing field encompasses information management, social
sciences and computer science. The domain of big data is characterised by the velocity,
volume and variety of data, making it valuable for data mining. However, many companies
struggle to fully leverage the value of Big Data amid a lack of skilled professionals and
uncertainty about how to effectively implement big data analytics to deliver business value.
Thus, to maximise the potential of projects, organisations seek data science professionals
and business experts who fully understand the company’s business model (Xu et al., 2022).
The development of data science is the main driver of the new generation of AI – ML
tandem specialists. These fields are attracting increasing interest from states, companies,
and the educational sector, enjoying important initiatives from various actors such as the
United States, China, and the European Union. Wing (2019) show that the field of data
science involves a long process, which encompasses both data analysis, the work done
before and after it, the issues related to ethics and data privacy, as well as the importance of
data collection, processing, storage, management, analysis, and visualisation, alongside the
valorisation of all by users, for decision-making in various fields. The life cycle of data is a
complex one, consisting of various stages, starting from data generation to their
interpretation, with an applicability also at the level of AI techniques. Furthermore, it is
necessary to promote ethical responsibility and data protection for each phase of the data
life cycle.
The contemporary period is characterised by a rapid increase in the popularity and utility of
data science, due to its use in business, industry, and academia (Ley and Bordas, 2018). De
Veaux et al. (2016) predict in their studies the need for hundreds of thousands of jobs in the
field of data science in the next decade, leading to an increase in study programs in this
field within universities. At the level of the business environment, companies from various
branches have noticed the need to hire more people in the data science department. The
data science job market provides a detailed list of job titles such as Data Science Specialist,
Data Analyst, Data Engineer, Statistician, Data and Business Analyst, Business Intelligence
Analyst. In this sense, it is very important to clearly define the skills required for this sector,
as well as those specific to each function. As for the academic branch, there is a rapid
increase in the number of institutions wishing to develop programs to train specialists in
this field. In addition to these, there is also a promotion of the field of data science, carried
out by various publications, which presents this career as an interesting and viable option
for the future (Provost and Fawcett, 2013).
Regarding the role of a data scientist, it was described by Cao (2019) as the most attractive
workplace of the 21st century. The role of data analytics, data science and the AI-ML
couple in information and communication technology and in the disciplines of science,

engineering, and technology is essential. However, the qualifications and competencies of a
data scientist are not yet clearly defined. In this sense, there is a need to establish the duties
of a data science specialist who belongs to the new generation, able to transform science,
technology, innovation and the current and future economy. While there is a notable
increase in the number of data science courses helping to define this new profession,
employers in this segment, including entrepreneurs, salespeople, and large enterprises,
report the limited availability of qualified data specialists to assist them in strategic
development and gives them competitive advantages in the future. Cao (2019) highlighted
the discrepancies in existing professional and educational markets, the lack of
standardisation and accreditation of data science responsibilities and competencies, and the
urgent need to standardise and improve data science qualifications and education.
In the context of the job market, it is essential to define the skills required for Business Data
Analytics and Data Science specialist positions. Using a text analysis by Radovilsky et al.
(2018) on the Glassdoor.com platform, the most sought-after skills are shown to be Python
- 72% (Familiarity with Python syntax and ecosystem), R-64% (Knowledge of statistical
programming language R), SQL - 51% (Managing databases using structured queries),
Hadoop 39% (Processing Big Data with Hadoop), Java - 33% (Familiarity with Python’s
syntax and ecosystem), SAS - 30% (Knowledge of SAS statistical programming language),
Spark - 27 % (Processing Big Data with Spark), Matlab - 20% (Knowledge of numerical
computation and programming in Matlab), Hive - 17% (Processing Big Data with Hive),
and Tableau - 14% (Knowledge and concepts of statistical visualisation). As data science
professionals are increasingly in demand and experienced ones are still rare, therefore, the
idea arises that theoretical training is not enough to excel in this field. In addition to
knowledge of programs and algorithms, practical experience in solving real business
problems plays a particularly important role (Berthold et al., 2019).
Technology and the field of work have evolved and the curriculum needs to adapt to the
new requirements (Hardin et al., 2015). The literature on Big Data skills mainly focuses on
the demand side (companies), the supply side (universities and students), and the
relationship between them. Following an analysis of recruitment information from
companies (to identify the knowledge and skills required for BD positions) and an
assessment of university programs (to see if academic training aligns with industry
expectations), it was determined that the introduction of a more sustainable and efficient
system for training and managing skills in this field. In addition, students’ perspectives
should be taken into account, to better understand their level of competence and to assess
the effectiveness of university programs (Xu et al., 2022).
Currently, according to Bile Hassan et al. (2021), there are more than 530 programs in data
science, analytics, and related fields, most of which are master’s and certificate programs,
offered both online and in traditional institutions. In this sense, an example of good practice
is represented by the guide made for academic institutions wishing to develop bachelor’s
programs in this field, which describes a curriculum for the bachelor’s program in the field
of data science. The curriculum is detailed and presents courses in related fields,
emphasising the importance of collaboration between departments and faculties in the
process of developing such an interdisciplinary program. Along with these, there is at the
level of the educational offer, and a set of courses that support the training of future
specialists in data science, which emphasise disciplines such as exploratory analysis, data
Vol. 26 • No. 65 • February 2024 19

set cleaning, and transformation, the Bayesian statistical model that essentially involves
learning from data and probabilities, regression models, dimensionality reduction, ML,
model performance, data mining from online platforms, sentiment analysis, relational
databases (Hicks and Irizarry, 2018).
2. Research methodology
In accordance with the specialised literature, the research will focus on the following
pillars: Data Analysis, Data Collection, Database Administration, Ethics, Prediction,
Programming and Project Management. In order to fulfil the objective related to the
academic skills in data science that students acquire after graduating from master’s
programs in data science in Romania, the discipline sheets of the main master’s programs in
the field of science were taken from the websites of the universities from Romania. Each
discipline was included in one of the main pillars of the analysis:
 The Faculty of Mathematics and Informatics (FMI), which includes the master’s
programs “Data Science for Industry and Society”, “Data science”, “BD - data science”,
“Analytics and Technologies”, which has three main pillars of interest: Data Collection,
Data Analysis, and Programming, complemented by disciplines included in Ethics and
Prediction.
 The Faculty of Automation and Computers (FAC), with the “Machine Learning” and
“Database Administration” master’s programs, which includes most of the pillars. The most
subjects focus on Programming and Project Management, followed by Data Collection,
Database Administration, Data Analysis, and Ethics.
 The Faculty of Cybernetics, Statistics and Economic Informatics (FCSIE), with the
master’s program “Applied Statistics and data science”, includes most disciplines that
belong to the main objective of Data Analysis, as well as Programming and Ethics
activities.
 The Faculty of Economics and Business Administration (FEAA) with the Master’s Data
Mining, as previously presented, focuses on Data Analysis, Data Collection, and Ethics.
From the visualisation of the data for the pillars addressed by each faculty, the assimilated
academic skills for each individual pillar were grouped into professional skills, which could
also leave their mark on the professional career. Thus, an association of data science roles
and academic competencies pursued by each faculty in the list above has been identified.
To meet the research objectives related to professional skills in the field of data science, the
2022 Stack Overflow Annual Developer Survey (Stack Overflow, 2022) data set was used
in the current analysis. It is based on the questionnaire applied in 2022 by Stack Overflow,
created with the aim of identifying the experience of specialists in the technologies used
and their experience in professional environments, globally. After filtering the responses
based on the job held by the respondents, records of participants with a data science
background were retained, with 11071 responses. In this context, the dataset underpins the
analysis of professional data science skills and provides insights for hypotheses such as:
 What is the distribution of jobs in this field?
 How many years of experience do the specialists have?
 What are the programming languages, analysis programs used?

 What is the support offered by companies to employees?
 What are the dynamics of data science teams in companies?
Through the use of descriptive statistics, the information from the data set was synthesised, this
being presented in figure no. 1, for which the Power BI visualisation program was used.
Figure no. 1. Data Science – Professional Skills

Source: Authors representation in Power BI based
on the “2022 Stack Overflow Annual Developer Survey”
The “2022 Stack Overflow Annual Developer Survey” (Stack Overflow, 2022) provides
insight into the data science job market. The main jobs on the data science labour market
identified, according to the questionnaire, are: Database administrator, data science
specialist or machine learning specialist, Data Engineers and Data/Business Analyst. Of the
Vol. 26 • No. 65 • February 2024 21

total respondents, 70% are full-time employees, and 30% are master’s graduates. Hybrid
(45%) and remote (39%) work modes predominate, with only 15% working exclusively
from the office. Preferences for company size differ: most opt for 20-99 employees,
100-499 employees, and 2-9 employees (48% overall), while 51% go for larger companies.
As for experience in data science, those with 1-3 years, 3-5 years, and 10 years of
experience stand out. It should be noted that 71% consider their activity in the field of data
science to be a hobby, highlighting the passion. Although 80% of employers provide time
for learning, 61% of employees use company development resources such as Coursera,
Udemy, Pluralsight, Codecademy, and edX. In order to assess how education level
contributes to the qualification or non-qualification of a data science specialist for industry,
a number of steps were taken. In the first phase, the initial 40 variables were reduced, by
applying Principal Component Analysis into two principal components, resulting in two
major variables that define the profile of a data scientist. Principal component analysis
reduces the number of professional and academic components, being a method of
identifying data patterns and expressing data in a way that highlights similarities and
differences between them (Jolliffe, 2011).
The name and description of the variables extracted from the survey “2022 Stack Overflow
Annual Developer Survey” (Stack Overflow, 2022) are presented in Table no. 1.
Table no. 1. Description of variables

Name of variable Description
The main field or branch of work with which the respondent
Employment identifies
MainBranch The respondent's employment status (e.g., full-time, part-time, self-
employed, etc.).
Remote Work Form of remote work or not (e.g., Fully remote, Hybrid, NA).
Coding Activities The type of coding activities the respondent is involved in (e.g.,
Hobby, Contribution to open-source projects, etc.).
EdLevel The highest level of education attained by the respondent.
Buy New Tool Acquisition of new programming tools.
Years Code ,Years Code The number of years the respondent has coded, both as a hobby and
Pro as a professional.
Organisation Size Size of the organisation the respondent works for.
Language Want To Work Programming languages the respondent has worked with and wants
to work with.
Database Have Work Databases the respondent has worked with and wants to work with.
Platform Have Worked Platforms the respondent has worked with and wants to work with.
OfficeStackAsync Used and preferred asynchronous collaboration tools.
Age Age of the respondent.
Work Expirience The number of years of work experience the respondent has.
Knowledge_3, Questions related to the respondent's knowledge and expertise in
Knowledge_4, specific areas.
Knowledge_5
Time searching, Time Time spent searching and responding to programming tasks.
answering
True/False_1, True/False_1, True/False_2, True/False_3: true/false answers to
True/False_2, certain statements or questions.
True/False_3
Source: Authors' own processing based on the “2022 Stack Overflow Annual Developer Survey”
For each main component containing the variables in Table no. 1, the KM clustering
algorithm was applied to re-code the observations according to the respective component,
obtaining distinct groups (clusters). KM clustering is a method of partitioning a data space
into K groups, where each group has a specific average value, to be able to define the
characteristic of each main component. In KM, each data point in the set is assigned to the
group closest to its mean value (Sterling, Anderson and Brodowicz, 2017).
This procedure was repeated for the other two principal components, resulting in a new re-
coded database. The KM algorithm was then applied again to this new database to obtain a
column representing the profile of a data scientist as skilled or unskilled.
In the final step, logistic regression (RL) was used to predict scenarios and identify the
aspects that need to be improved in order to achieve the profile of a qualified data scientist.
RL is a widely used statistical model that allows multivariate analysis and modelling of a
binary-dependent variable. (Ranganathan, Pramesh and Aggarwal, 2017). As a result of the
presentation of the main components and the grouping of their characteristics using KM, it
is possible to define, with the help of logistic regression, the academic and professional
competencies that should be adjusted to support the career of a future specialist in data
science.
This review provided relevant guidance and recommendations regarding the influencing
factors and skills required to obtain a qualification in this field. Logistic regression is
mainly used because of its function of predicting the outcome of an event using binary
values. In the present case, it determines whether the respondent is skilled or unskilled in
the labour market. The current research uses the logistic regression function to demonstrate
how the identified characteristics influence the predicted outcome, namely whether a
respondent is qualified or unqualified to become employed in the business environment in a
certain field, similar to Peng, Lee and Ingersoll’s (2002) logistic regression study.
The mathematical formula of the model was:
P(y=1 | x) = 1 / (1 + exp( - (b0 + b1*x1 + b2*x2))) (1)
where:
 P(y=1 | x) – represents the probability that the dependent variable y has the value 1,
where 1 is the desired event, given the value of the predictors x;
 b0, b1 și b2 – are the estimated RL coefficients, showing the weights associated with
each predictor;
 x1 și x2 – represent the values of the predictors associated with the independent variable x.
3. Results and discussions

Studies that have looked at this topic have focused on discovering academic and
professional skills that drastically contributes to employment opportunity. Mainga et al.
(2022) addressed the topic of identifying the respondents’ skills that make them eligible for
the business environment in a certain field. The study was limited only to the descriptive
component of the respondents’ skills supporting their eligibility to be employed. The
current research used the respondents’ already acquired skills in the labour market,
predicting skilled or unskilled status, as well as how respondents can improve their skills to
become data scientists for the business environment.
Vol. 26 • No. 65 • February 2024 23

The academic skills assimilated for each individual topic were grouped following the
analysis of the subject sheets, obtaining valuable information. Thus, the academic skills
acquired for the study of the subjects included in the topic of Data Analysis were: the
ability to identify the necessary algorithms and apply the right analysis to deliver the results
needed for decision-making purposes to solve the requirements, the ability to approach a
hypothesis from multiple perspectives, and to be aware of the limitations of models. After
studying the subjects from the Data Collection topic, the academic competences
undertaken are: the knowledge needed to transform large data sets into complex projects by
applying qualitative and quantitative methods, applying exploratory techniques, detecting
extreme values or applying the correct principles for the implementation and development
of analyses for Big Data systems. For the topic of Data management, the main academic
skills acquired are to know the principles of database management and the application of
the necessary techniques. The forecasting disciplines provide students with the analytical
knowledge to develop hypotheses and identify predictions for existing trends. Following
the study Programming courses, students will acquire skills in using programming
languages to solve analysis requirements, and ethics and project management disciplines
will provide graduates with the informational foundation needed in the process of project
development, customer service delivery, and data science entrepreneurship.
The nature of each institution has also left its mark on the data science programs. Thus,
according to the professional skills described in the previous part of the paper, an
association of data science roles and academic skills pursued by each faculty was identified.
From the list above, the following can be noted:
 The Faculty of Mathematics and Informatics (FMI) prepares specialists whose skills fall
within the professional skills for the roles of Data Engineer or Database Administrator, by
combining the three: data collection and analysis, alongside programming skills.
 The Faculty of Automation and Computers (FAC), outlines the academic skills covering
a wide number of topics in data science, with attention directed to Programming and Project
Management, as a result the role that matches the professional skills described being a
specialist in data science.
 The Faculty of Cybernetics, Statistics and Economic Informatics (FCSIE) and the
Faculty of Economics and Business Administration (FEAA) have as their main objective
the outline of academic skills for Data Analysis, an aspect that is also found in the FEEA
curriculum.
From the subject sheets of each master’s program, the computer programs used in the
training of students’ skills were also extracted:
 FMI works with R, SAS and Python, and for the previously realised association
between the academic skills outlined by the master’s programs within this faculty and the
professional skills for the roles of DE or Database Administrator, according to the
requirements of the labour market, it is necessary to add in the program and the following
programs: SQL, Oracle and Microsoft Azure.
 FAC covers a wide range of programs such as Python, SQL, Linux, Matlab, Oracle, and
Tableau. As a result, to align the academic skills trained with the professional skills
required for the data scientist role, the R program and Cloud technologies could be added.
 FCSIE is the one that covers almost all the technologies required for the role of Data
Analyst and Business Analyst by using the following: R, SAS, SPSS, Tableau, Python,
Excel, and GeoDa, where the only technology that could be added is SQL.
 For FEEA, where work is mainly done with R, in order to align the academic skills
formed with the professional skills needed on the labour market for the role of Data Analyst
and Business Analyst, the integration of SAS, SQL, Python, and Tableau technologies
would be necessary.
In order to visualise the relationships between variables and the possible existence of some
groups of variables, for the database related to professional skills, following the application
of the analysis in main components, the major characteristics that a specialist in data
science must have been defined. The analysis started by standardising the data, using the
Range method. This method was chosen because there were extreme values. In order to
reduce the size of the data and to decide how many principal components it would be
advisable to keep, the eigenvalue matrix presented in Table no. 2.
Table no. 2. Eigenvalues of the Uncorrected Correlation Matrix
Eigenvalue Difference Proportion Cumulative
1 17.69 15.95 0.68 0.68
2 1.74 0.79 0.06 0.74
3 0.94 0.0364 0.78
Source: Authors’ own processing using SAS Studio on based
The first value in the array explains 17.69 variables, so it would replace 17 variables with
just one, thus having a significant reduction in space. The second value also explains almost
two variables. Next, two main components were kept in the analysis. Also, the two
principal components take up 74.7% of the information.
In order to identify the variables that determine each of the two preserved components
(Table no. 3), the two main components were interpreted.
Table no. 3. Eigenvectors

Principal Principal
The name of the variables
Component 1 Component 2
MainBranch 0.11 -0.08
Employment 0.13 -0.32
Remote Work 0.21 0.03
Coding Activities 0.22 0.07
EdLevel 0.17 0.07
Work Experience 0.12 0.55
Learn Code 0.14 -0.11
Learn Code Courses 0.17 0.08
Years Code 0.2 0.32
Organisation Size 0.2 -0.13
Vol. 26 • No. 65 • February 2024 25

Principal Principal
The name of the variables
Component 1 Component 2
Buy New Tool 0.22 0.08
Language Want To Work 0.2 0.04
Database Have Worked 0.21 0.18
Platform Have Worked 0.18 0.05
Knowledge_3 0.23 -0.11
Time Searching 0.21 -0.23
True/Fals_1 0.22 -0.16
True/Fals_2 0.22 -0.2
True/Fals_3 0.22 -0.15
Time Answering 0.21 -0.21
Years Code Pro 0.17 0.38
OfficeStackAsync 0.2 0.04
Age 0.19 0.25
The first component is positively determined by the variables: Remote Work, Coding
Activities, Organisation Size, Buy New Tool, Language Want To Work, Database used,
Can I find updated information within my organisation to help me- I do my job, I know
which system or which resource the user needs to find information and answers to the
questions I have, I know which system or which resource the user needs to find information
and answers to the questions I have, Time Searching, Time Answering, Are you involved in
supporting new hires during their onboarding?, Do you use employer-provided learning
resources?, Does your employer give you time to learn new skills?, variables related to the
technical professional skills of respondents in the data science labour market.
The second component is positively determined by the variables: Years Code, Years Code
Pro, Age, Work Experience, variables related to respondents’ experience and negatively by
the variables: Employment, Time Searching, Time Answering, Use learning resources
provided by the employer?, which refer to their development. Thus, the secondary
component will represent the respondents’ experience and development in the data science
labour market.
By applying principal component analysis, redundancy was removed and two principal
components were retained, along with significant variables for each.
In the continuation of the research, the Cluster analysis was carried out for each component
in order to identify the groups of respondents. Thus, following the Cluster Observation
algorithm, they resulted, according to Table no. 4, the following:
Table no. 4. Cluster History

First Principal Second Principal Component
Cluster Number
Component R-Square R-Square
11070 1 1
11069 1 1
11068 1 1
... ... ...
5 0.58 0.70
4 0.56 0.67
3 0.54 0.61
2 0.52 0.53
1 0.00 0.00
Table no. 4 shows that for the first component the greatest loss of homogeneity, where R-
Square (0.54)>0.5, is if the variables were divided into three clusters, and for the second
component, where R-Square (0.53)>0.5, is if we split the variables into two clusters.
To define the 3 clusters of respondents within the first component, the KM Clustering
analysis was applied, resulting in the first cluster with 3779 respondents, the second 2789,
and the last including 4503 respondents.
For the characterisation of each cluster, Table no. 5 in which only significant variables with
R-Square > 0.5 were taken into account.
Table no. 5. Cluster Means
The name of the variables Cluester no. 1 Cluester no. 2 Cluester no. 3
Knowledge_3 0.58 0.36 0.99
Knowledge_4 0.63 0.39 0.99
Knowledge_5 0.66 0.47 0.99
Time Searching 0.27 0.25 0.99
Time Answering 0.32 0.21 0.99
True/Fals_1 0.33 0.40 0.99
True/Fals_2 0.29 0.34 0.99
True/Fals_3 0.41 0.39 0.99
Cluster 1 represents the most experienced data science professionals. They quickly find up-
to-date information in the company, efficiently use systems to solve tasks, and have a short
average time to find answers. However, they engage less in the onboarding process and use
fewer learning resources because they have already developed professional skills.
Cluster 2 represents specialists with medium experience in data science. They have
difficulty finding up-to-date information in employers’ systems and do not always know
where to look for answers, but they find solutions quickly. Companies’ appreciation of
development time and resources is average.
Vol. 26 • No. 65 • February 2024 27

Cluster 3 represents early-career data scientists. They find up-to-date information and know
where to look, but the process of finding the information they need takes the longest. They
invest a lot of time in answering their work, especially during the onboarding period, and make
heavy use of learning resources provided by employers to develop their professional skills.
For the second component, the KM Clustering analysis was applied again, resulting in the
first cluster with 4555 respondents, and the second with 6516. To characterise each cluster,
table no. 6, in which only significant variables with R-Square > 0.5 were taken into
account.
Table no. 6. Cluster Means
The name of the variables Cluester no. 1 Cluester no. 2
Time Searching 0.99 0.25
Time Answering 0.99 0.27
True/Fals_2 0.98 0.31
Cluster 1 includes respondents who have a long time in the process of searching for an
answer and who use the learning resources provided by employers the most. It is the group
of data science specialists in the process of developing professional skills.
Cluster 2 includes the respondents for whom finding solutions in their work takes the least
time and they use the company’s learning resources in a small to medium percentage. Thus,
this is the cluster with specialists who have completed the process of developing
professional skills.
Also, for the two components, the KM Clustering method was applied, with the aim of
observing which cluster of the two each model belongs to. This was done to construct a
new binary variable representing the profile of the data scientist and to be further used as a
dependent variable for predictive purposes.
Logistic regression is constructed in order to model and predict the binary dependent
variable created above (the profile of the qualified/unqualified specialist in data science),
according to the independent variables, the two components identified in the previous
analyses: the first related to the technical professional skills of the respondents on the data
science job market and the second, which represents the respondents’ experience and
development on the data science job market.
The results obtained in the prediction of the binary dependent variable (skilled/unskilled
data scientist profile) were:
 For category 1 – specialists with experience in the labour market and developed
professional technical skills, from the independent variable component 1 and for category 1
– specialists in the process of development, from the independent variable component 2, the
logistic regression model predicted the value 0 – qualified profile.
 For the same category 1 - specialists with experience on the labor market and developed
professional technical skills, if category 2 is set for the second component - specialists who
have completed the development process, the logistic regression model predicted the value
1 - unqualified profile for a specialist from data science.
 For category 2 – specialists with an average level of knowledge, from the independent
variable component 1 and for category 1 – specialists in the process of learning, from the
independent variable component 2, the model framed the profile of the specialist as a
qualified one.
 For category 3 – specialists at the beginning of their career, from the independent
variable component 1 and for category 1 – professional skills in the process of
development, from the independent variable component 2, the logistic regression model
predicted the value 0 – qualified profile.
To validate the logistic regression model, the accuracy of the model was calculated for the
analysed data set. The accuracy was 100% for the model included in the above analysis,
which shows that the model correctly predicted the output values based on the input data.
Following the logistic regression model, the profile of the data science specialist was
predicted according to the characteristics of the two independent variables, and to be a
qualified specialist the following were identified: even though the level of experience is
very important, to be a specialist qualified, continuous and active development is necessary,
as a result the time and resources that employers provide are important. For an average
level of experience, by appreciating career development opportunities and interest in
learning, employers are looking for such a profile. For specialists at the beginning of their
career, the development of professional skills occupies a large part of the current activity.
Although the level of experience is not as high, openness to learning is recommended for
this category of employees.
By applying the logistic regression model, the main conclusion was that regardless of the
experience level of a data scientist, it is the continuous and active development of
professional skills that is recommended and attracts the attention of employers. Thus, both
the learning opportunities that companies offer to their employees, but also the academic
skills acquired by specialists, have a critical role for their profile.
Conclusions
The need to know the profile of the data science specialist has become increasingly
important, both for outlining the academic program and for identifying the professional
skills needed in the labour market. Thus, this research represents a useful informational
support in order to make decisions for adapting academic programs to current requirements
and understanding the key skills to practice in one of the roles under the umbrella term data
science.
Considering the fact that data science focuses on the knowledge of computer programs,
algorithms, and the accumulation of experience in their use, the first part of the research
had as its main objectives the identification of the professional skills needed on the labour
market in this field, but also the outline of the specialists’ profile. The level of experience,
technical professional skills, and interest in learning outline the profile of the data scientist.
Although the years of professional experience and the openness to the development of skills
will place a specialist in the category of those qualified and sought after on the labour
market, a significant role for their profile is played by the academic skills acquired by
specialists, an idea reiterated by Irizarry (2020).
Vol. 26 • No. 65 • February 2024 29

Completion of a bachelor’s or master’s program may be considered mandatory to practice

data science. In this sense, the educational plans for master’s programs in Romania were
analysed. One of the main results was to identify the association of the academic skills
acquired by graduating from these programs and the roles in the data science labour market.
Thus, FMI outlines the academic skills required for the roles of Data Engineers or Database
Administrator, FAC outlines the direction towards the role of Data Science Specialist, and
FEAA and FCSIE teach and train the skills required for the role of Data Analyst and
Business Analyst.
Technical professional skills are specific to each role. On the one hand, knowledge in SQL,
Python or Shell Scripting is not required for a Database Administrator, Python, Java, or
C++ are sought for DE. On the other hand, for a DA or BA the requirements included are
R, SAS, Python, SQL, Tableau or Power BI and for the Data Science Specialist role skills
in Python, AI, SQL or R are required. Equally significant, there are also communication
skills, time management, teamwork, and analytical thinking required for all roles. The
profile of the data scientist is defined by the level of experience, professional technical
skills, and interest in learning.
In conclusion, academic skills and professional skills depend on each other. By aligning
study programs with the labour market, data science is a field that will have qualified and
trained specialists.
The research is limited by the lack of data for the year 2023, as this is a year in which AI
has developed at a rapid pace, with data science participating in the development of
chatbot-like search engines and helping to open up new opportunities for data scientists. In
the future, it is necessary to study the feasibility of introducing chatbot-type AI in the field
of data science in academic institutions, even more so in the context of increasing labour
supply. As a result, there is a need to identify the main characteristics that a specialist in
data science should possess, with the aim of introducing them into academic programs.
References
Alqahtani, T., Badreldin, H.A., Alrashed, M., Alshaya, AI., Alghamdi, S.S., Bin Saleh, K.,
Alowais, S.A., Alshaya, O.A., Rahman, I., Yami, M.S. and Albekairy, A.M., 2023. The
Emergent Role of Artificial Intelligence, Natural Learning Processing,
and Large Language Models in Higher Education and Research. Research
in Social and Administrative Pharmacy, 19(8), pp. 1236-1242. https://doi.org/
10.1016/j.sapharm.2023.05.016.
Berthold, M.R., 2019. What Does It Take to Be a Successful Specialist in Data Science?
Harvard Data Science Review, 1(2). https://doi.org/10.1162/99608f92.e0eaabfc.
Bile Hassan, I., Ghanem, T., Jacobson, D., Jin, S., Johnson, K., Sulieman, D. and Wei, W.,
2021. Data Science Curriculum Design: A Case Study. In: Proceedings of the 52nd ACM
Technical Symposium on Computer Science Education. Virtual Event USA: ACM. pp.
529-534. https://doi.org/10.1145/3408877.3432443.
Cao, L., 2019. Data Science: Profession and Education. IEEE Intelligent Systems, 34(5),
pp. 35-44. https://doi.org/10.1109/MIS.2019.2936705.
Cao, L., 2016. Data Science and Analytics: A New Era. International Journal of Data
Science and Analytics, 1, pp. 1-2. https://doi.org/10.1007/s41060-016-0006-1.
Chassignol, M., Khoroshavin, A., Klimova, A. and Bilyatdinova, A., 2018. Artificial
Intelligence Trends In Education: A Narrative Overview. Procedia Computer Science,
136, pp. 16-24. https://doi.org/10.1016/j.procs.2018.08.233.
Chen, X., Xie, H., Zou, D. and Hwang, G.-J., 2020. Application and Theory Gaps During
the Rise of Artificial Intelligence in Education. Computers and Education: Artificial
Intelligence, 1, article no. 100002. https://doi.org/10.1016/j.caeai.2020.100002.
De Veaux, R.D., Agarwal, M., Averett, M., Baumer, B.S., Bray, A., Bressoud, T.C.,
Bryant, L., Cheng, L.Z., Francis, A., Gould, R., Kim, A.Y., Kretchmar, M., Lu, Q.,
Moskol, A., Nolan, D., Pelayo, R., Raleigh, S., Sethi, R.J., Sondjaja, M., Tiruviluamala,
N., Uhlig, P.X., Washington, T.M., Wesley, C.L., White, D. and Ye, P., 2017.
Curriculum Guidelines for Undergraduate Programs in Data Science. Annual Review of
Statistics and Its Application, 4(1), pp. 15-30. https://doi.org/10.1146/annurev-statistics-
060116-053930.
Donoho, D., 2017. 50 Years of Data Science. Journal of Computational and Graphical
Statistics, 26(4), pp. 745-766, https://doi.org/10.1080/10618600.2017.1384734.
Dyk, D.V., Fuentes, M., Jordan, M.I., Newton, M., Ray, B.K., Lang, D.T. and Wickham,
H., 2015. ASA Statement on the Role of Statistics in Data Science. [online] Available at:
<http://magazine.amstat.org/blog/2015/10/01/asa-statement-on-the-role-of-statistics-in-
data-science> [Accessed 15 August 2023].
Grossi, V., Giannotti, F., Pedreschi, D., Manghi, P., Pagano, P. and Assante, M., 2021. Data
Science: A Game Changer for Science and Innovation. International Journal of Data
Science and Analytics, 11, pp. 263-278. https://doi.org/10.1007/s41060-020-00240-2.
Hardin, J., Hoerl, R., Horton, N.J., Nolan, D., Baumer, B., Hall-Holt, O., Murrell, P., Peng,
R., Roback, P., Temple Lang, D. and Ward, M.D., 2015. Data Science in Statistics
Curricula: Preparing Students to “Think with Data”. The American Statistician, 69(4),
pp. 343-353. https://doi.org/10.1080/00031305.2015.1077729.
Hick, S.C. and Irizarry, R.A., 2018. A Guide to Teaching Data Science. The American
Statistician, 72(4), pp. 382-391. https://doi.org/10.1080/00031305.2017.1356747.
Irizarry, R.A., 2020. The Role of Academia in Data Science Education. Harvard Data
Science Review, 2(1). https://doi.org/10.1162/99608f92.dd363929.
Jolliffe, I., 2011. Principal Component Analysis’. Berlin: Springer. pp. 1094-1096,
https://doi.org/10.1007/978-3-642-04898-2_455.
Kordon, A.K., 2020. Data Science Based on Artificial Intelligence. Cham: Springer.
https://doi.org/10.1007/978-3-030-36375-8_1.
Leal, W., Eustachio, J.H.P.P., Nita, D.A.C., Dinis, M.A.P., Salvia, A.L., Cotton, D.R.E.,
Frizzo, K., Trevisan, L.V. and Dibbern, T., 2023. Using Data Science for Sustainable
Development in Higher Education. Sustainable Development,
https://doi.org/10.1002/sd.2638.
Ley, C. and Bordas, S.P.A., 2018. What makes Data Science different? A Discussion
Involving Statistics2.0 and Computational Sciences. International Journal of Data
Science and Analytics, 6(3), pp. 167-175. https://doi.org/10.1007/s41060-017-0090-x.
Vol. 26 • No. 65 • February 2024 31

Mainga, W., Murphy-Braynen, M.B., Moxey, R. and Quddus, S.A., 2022. Graduate
Employability of Business Students. Administrative Sciences, 12(3), article no. 72.
https://doi.org/10.3390/admsci12030072.
Pelau, C., Ene, I. and Pop, M.I., 2021. The Impact of Artificial Intelligence on Consumers’
Identity and Human Skills. Amfiteatru Economic, 23(56), pp. 33-45,
https://doi.org/10.24818/EA/2021/56/33.
Peng, C.-Y.J., Lee, K.L. and Ingersoll, G.M., 2002. An Introduction to Logistic Regression
Analysis and Reporting. The Journal of Educational Research, 96(1), pp. 8-9,
https://doi.org/10.1080/00220670209598786.
Provost, E. and Fawcett, T., 2013. Data Science and its relationship to big data and data-
driven decision making. Big Data, 1(1), pp. 51-59, https://doi.org/10.1089/
big.2013.1508.
Radovilsky, Z., Vishwanath, H., Anuja, A. and Uma, U., 2018. Skills Requirements of
Business Data Analytics and Data Science Jobs: A Comparative Analysis. Journal of
Supply Chain and Operations Management, 16(1), pp. 82-101.
Ranganathan, P., Pramesh, C.S. and Aggarwal, R., 2017. Common Pitfalls in Statistical
Analysis: Logistic Regression. Perspective in Clinique Research, 8(3), pp. 148-151.
https://doi.org/10.4103/picr.picr_87_17.
Smaldone, F., Ippolito, A., Lagger, J., and Pellicano, M., 2022. Employability Skills:
Profiling Data Scientists in the Digital Labour Market. European Management Journal,
40(5), pp. 671-684. https://doi.org/10.1016/j.emj.2022.05.005.
Stack Overflow, 2022. Stack Overflow Annual Developer Survey. [online] Available at:
<https://survey.stackoverflow.co/2022> [Accessed 8 December 2023].
Sterling, T., Anderson, M. and Brodowicz, M., 2017. High Performance Computing:
Modern Systems and Practices. New York: Elsevier.
Tuba, L. and Süheyla, A., 2023. The impact of Artificial Intelligence in academia: Views of
Turkish academics on ChatGPT, Heliyon, 9(9), article no. 19688. https://doi.org/
10.1016/j.heliyon.2023.e19688.
Xu, A., Wu, Y., Meng, F., Xu, S. and Zhu, Y., 2022. Knowledge and Skill Sets for Big
Data Professions: Analysis of Recruitment Information Based on The Latent Dirichlet
Allocation Model. Amfiteatru Economic, 24(60), pp. 464-484. https://doi.org/
10.24818/EA/2022/60/464.
Wing, J.M., 2019. The Data Life Cycle. Harvard Data Science Review, 1(1).
https://doi.org/10.1162/99608f92.e26845b4.
AE
IDENTIFYING SUFFICIENT AND NECESSARY COMPETENCIES

IN THE EFFECTIVE USE OF ARTIFICIAL INTELLIGENCE TECHNOLOGIES
Ion Popa1* , Marian Mihai Cioc2 , Andreea Breazu3
and Cătălina Florentina Popa4
1)
Bucharest University of Economic Studies, Romania and Academy
of Romanian Scientists, Romania
2)3)4)
Bucharest University of Economic Studies, Romania

Popa, I., Cioc, M.M., Breazu, A. and Popa, C.F., 2024. Received: 28 September 2023
Identifying Sufficient and Necessary Competencies in Revised: 15 November 2023
the Effective Use of Artificial Intelligence Accepted: 10 December 2023
Technologies. Amfiteatru Economic, 26(65), pp. 33-52.
Abstract
Recently, there have been significant changes in the labour market and in the lives of
employees, as modern society adapts increasingly easily to the implementation of artificial
intelligence tools. However, technological changes have also created challenges, including a
gap between available and required competencies in the use of artificial intelligence
technologies. This study aims to analyse the relationships between employee competencies
and effectiveness in the use of artificial intelligence tools, in order to highlight the set of
essential competencies in effective interaction with artificial intelligence technology.
Therefore, to achieve the purpose of the research, a questionnaire was created and completed
by 209 Romanian employees between August and September 2023. For data analysis, two
advanced techniques were applied: structural equation modelling (SEM) and necessary
conditions analysis (NCA) using the SmartPLS v4 program. The results suggest that
employee competencies are significantly associated with the effectiveness of using AI tools,
and optimism and innovativeness positively mediate this relationship. The originality of the
research stands out through the use of two advanced analysis methods (structural equation
modelling and necessary conditions analysis), with the aim of identifying the set of sufficient
and necessary skills in the use of artificial intelligence tools. These findings have significant
implications for organisations, the educational system, and future research directions on the
managerial implications of using artificial intelligence tools.
Keywords: artificial intelligence, competencies, formal education, non-formal education,
effectiveness, education
JEL classification: D83, O15, I25
*
Corresponding author, Ion POPA – e-mail: [email protected]
Attribution License, which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited. © 2024 The Author(s).
Vol. 26 • No. 65• February 2024 33

AE Identifying Sufficient and Necessary Competencies in the Effective Use
of Artificial Intelligence Technologies
Introduction
The rapid progress of computer technologies, in general and artificial intelligence (AI)
systems, in particular, has led to profound changes in the labour market. Researchers (Pan
and Froese, 2022) appreciate that AI has the potential to radically change the future of human
resource management in the context of its increasingly emphasised role in the foundation of
managerial decisions (Leyer and Schneider, 2021). The World Bank Group (2019) estimates
that half of the job growth in Europe between 1999 and 2016 was due to technological
changes implemented at the level of repetitive processes. The International Labour
Organisation (2023) shows that 24% of office tasks are highly exposed to technological
change, while 58% of them are characterised by medium exposure. In the US, for example,
approximately 47% of jobs are in the upper risk zone for potential automation (Frey and
Osborne, 2017). In Germany, although AI-based robotisation did not have a substantial effect
on employment, it did reduce youth employment (World Bank Group, 2019).
At the same time, under the conditions of the challenges launched by the progress of AI-
based technologies, a relatively new concept addressed in the specialised literature is that of
preparation for AI (AI-Readiness), reflecting the capabilities of an organisation to implement
and use AI in a way that generates added value through digital transformation (Holmstrom,
2022). Therefore, companies that succeed in implementing and scaling AI technologies
amplify their ability to become competitive and develop AI-based capabilities (Logg et al.,
2019). Employee competencies are central to AI training, directly contributing to AI maturity
levels defined by Martinez-Plumed et al. (2021) at the level of seven classes: knowledge
representation, learning, communication, perception, planning, robotics and collective
intelligence. Issa et al. (2022) define an AI training system structured on three components
under the influence of AI competencies, respectively: the attitude towards human-machine
collaboration, the ability to anticipate the strategic impact of AI, respectively, the
technological infrastructure and the capabilities of data management.
On the other hand, research (Qureshi et al., 2021) reveals a gap between available
competencies and those needed to meet the demands of AI technologies. Fareri et al. (2020)
identify not only the need to integrate existing competencies into professional models, but
also the creation of new competencies adapted to Industry 4.0 trends. At the institutional
level, the European Commission strongly encourages the Member States (2021) to improve
their portfolio of specific competencies for the development and implementation of AI
solutions. The OECD (2022) states that governments should collaborate with stakeholders in
a way that allows people to use and interact with AI, including developing the necessary
competencies. At the same time, UNESCO (2021) proposes the development of mechanisms
and tools to anticipate the current and future needs of AI-related competencies in the context
of a relevant curriculum for the labour market.
Considering the challenges mentioned above, the purpose of this research is to analyse the
relationship between employee competencies and effectiveness in the use of AI tools. In
terms of implications of an applied nature, the research aims to contribute to the identification
of possible directions of managerial intervention aimed at improving the formal (FE) and
non-formal (NFE) education system in order to reduce the gaps between the available skills
AE
and those needed in the workforce market, considering the expansion of AI-based systems.
From this perspective, the research objectives show increased relevance in the general
context of the field by addressing the current challenges of the lifelong learning system under
the influence of the development of AI technologies, thus aligning with the research efforts
of other authors (Chatterjee et al., 2012; Fernandez Sanz et al., 2017; Brougham and Haar,
2020; Chowdhury et al., 2022; Da Silva et al., 2022; Varma et al., 2022; Verma and Singh,
2022; Tilibaşa et al., 2023). From the point of view of theoretical contributions, a main
element of originality of the paper is the combined use of two advanced analysis methods,
namely structural equation modelling (PLS-SEM) and necessary conditions analysis (NCA),
in order to define the set of sufficient and necessary skills in the use of AI technologies.
Another novelty element is represented by the constructs used to evaluate the preparation for
AI at the level of optimism and innovation and, respectively, discomfort and insecurity, as
well as highlighting the indirect effects mediated by these constructs in the relationship
between employees' skills and their effectiveness.
The main research gaps that the study aims to cover relate to the lack of consensus on the
skills required for AI and the role of organisations' preparation for AI in their effective use,
aspects that are, however, justified by the novelty of this research field. Thus, by defining
sets of essential skills for AI and evaluating their impact on effectiveness, the research
provides the premises for identifying some directions of focus of FE and NFE programs for
AI, through which to ensure the maximisation of the level of training of organisations for AI
and implicitly their effective use.
The paper is structured as follows. In the first section, the theoretical framework of the paper
is described based on the analysis of the specialised literature. The second section includes
the conceptual model and research hypotheses. The third section presents the materials and
methods used, while Section four presents the results of the research. The last section of the
paper contains the conclusions.
1. The role of the education system in shaping AI competencies

The issues of AI competencies and the role of education in the development of AI
competency models have been addressed in several studies or specialist articles. In traditional
approaches, competency models for AI are primarily based on digital competency sets,
consisting of knowledge, skills, and attitudes needed to perform tasks, communicate
problems, or develop knowledge in the context of IT&C (Fernandez Sanz et al., 2017).
Digital competencies provide important competitive advantages even for non-IT occupations
(Chen et al., 2022). However, according to some authors (Gupta et al., 2022), technological
competencies do not have a significant and positive impact on the intention to adopt AI
solutions. According to the Lisbon Council (2022), modern workplaces are significantly
based on emotional intelligence and analytical capabilities. Other important cognitive skills
in the context of AI are collaborative intelligence (Chowdhury et al., 2022) or critical
evaluation (Liaw et al., 2022). On the other hand, concurrently with the development of
specific competencies for AI, an ethical issue (Varma et al., 2022) is related to the way of
managing the process of diminishing the role of some competencies, such as mainly decision-
Vol. 26 • No. 65• February 2024 35

making, by replacing these with AI-based algorithms. The specialised literature mainly
reveals two major approaches to skills models for AI: (1) the traditional one, based on skills
mostly related to the use of technology and (2) the multidisciplinary one, focused on the
development of models that integrate complex cognitive skills. The present research aligns
with the second type of approach, since three of the five essential competencies for AI
integrated in the conceptual model proposed by the authors refer to non-technological
competencies, namely leadership, personal, and social competencies. At the same time,
although the number of people aged between 25 and 34 with a higher education degree has
doubled in the last two decades (Lisbon Council, 2022), there is a skills shortage in the labour
market in the scope of using AI systems. In this context, this research aims to examine the
role of the lifelong education system in the development of AI skills, simultaneously with the
identification of possible directions of managerial intervention aimed at improving this
system, at the level of FE and NFE components.
In the FE system plan, Da Silva et al. (2022) emphasise the role of partnerships between
universities and the business environment to implement new learning methodologies in the
field of AI, based on the use of interactive digital platforms. Moreover, an important aspect
of the mission for the society of higher education institutions is the added value regarding the
skills needed in the professional activity, developed among the graduates (Marinaș et al.,
2021). The benefits of a closer link between universities and industry are exemplified by
Eriksson et al. (2020) through the eight AI competence centers created in Italy. The analysis
of specialised studies (UNESCO, 2021) also reveals a series of good practices at the
educational level on the set of AI skills: in China, the discipline ‘Algorithms and
Computational Intelligence’ was introduced as early as 2017 in the ICT curriculum standards
for high schools; in Singapore, the SkillsFuture82 initiative is focused on the development of
specific skills and digital reskilling; in Finland, Headai was developed, namely an AI solution
focused on the development of skills maps to align the university environment with the
demands of the labour market. A major initiative at European level (The Lisbon Council,
2022) is the “European Skills, Competences, Qualifications and Occupations” (ESCO)
project, launched by the European Commission in 2020, thus outlining the formal regulatory
framework for the development of a set of specific skills for AI. The need to standardise the
skills development process through ESCO or O*NET nomenclatures is also analysed in other
works (Fernandez Sanz et al., 2017; Fareri et al., 2020; European Commission, 2021).
Kabashkin et al. (2023) develop the concept of the fifth-generation university based on the
review of professional and social skills by applying the results obtained in the field of
cognitive sciences. The approaches emphasise the priority role of the public-private
partnership, as a tool to integrate the process of developing the skills necessary for AI through
EF, with the demands of the labour market.
At the same time, according to the European Commission (2018), the educational system
must create models of AI skills that ensure lifelong learning, an approach that involves
specific NFE actions. At the organisational level, it is necessary for executive managers to
implement appropriate policies in the form of training programs or support centers for the
purpose of developing AI skills (Chatterjee et al., 2021). Chowdhury et al. (2022) recommend
the implementation of coaching systems in which the leading role in the transition to AI-
AE
based systems is played by employees with relevant experience in the use of these
technologies. Also, it is important that AI skills development tools are available at the right
time, as research (Verma and Singh, 2022) shows that employees demonstrate more positive
behaviour when they are actively supported when learning and using new AI skills. Finally,
a specific characteristic of ENF relates to flexibility, in the sense that labor mobility can
increase when opportunities arise to develop skills in areas required by the labor market in
the context of AI (Brougham and Haar, 2020). Through the specific tools mentioned before,
the role of ENF is emphasised especially in the context of the assimilation of new AI
solutions, while the need for ENF is increasingly important in the conditions of rapid progress
of these technologies. The link between the development of the set of key skills for AI, based
on the integrated use of EF and ENF tools and the increase in the level of readiness of
organisations for AI with an impact on the effectiveness in using these tools, is the main
relationship on which the working hypotheses developed in the framework of the work.
2. Conceptual model and research hypotheses

The concept model proposed in this research is based on an analysis of AI competencies
systems presented in the specialised literature, relevant for the purpose of generating
increased efficiency in the use of these technologies. Younnis and Adel (2020) define five
categories of competencies needed in the context of adopting AI solutions: hard and soft,
superior cognitive (problem solving, creativity, judgment, and critical thinking), social and
emotional (teamwork, leadership, and communication), technological, and research. The
model proposed by Jaiswal et al. (2021) includes five critical competencies (data analysis,
digital, complex cognitive, decision making, and continuous learning), with the need for a
higher-level development of cognitive and technological competencies being emphasised as
a premise for perfecting the human-AI relationship. In terms of managerial competencies,
Zhang (2023) defines a model addressed to project managers involved in the integration of
AI solutions, structured on 16 competencies such as planning, control regulation, systematic
substantiation of decisions, behavioural initiative and, respectively, fairness and impartiality.
Other important competencies in the context of AI are collaborative intelligence (Chowdhury
et al., 2022) or critical evaluation (Liaw et al., 2022). Frey and Osborne (2017) show that, in
the context of advancing technology, employees with low skill levels need to acquire creative
and social competencies. Organisational agility is also a factor that facilitates the
development of AI competencies necessary for the successful adoption of these technologies
(Chatterjee et al., 2021). The proposed conceptual model includes five categories of key AI
competencies (technical, digital, leadership, personal and social) and their relationship.
Therefore, we propose the following.
Hypothesis 1 (H1): The higher the level of employees’ competencies, the more effective they
will be in using AI tools.
Previous research suggests a strong relationship between the level of AI competencies and the
attitude towards the adoption of these technologies at the employee level, characterised by
optimism, enthusiasm, and innovation. In a study looking at the use of AI in the creative
industries in the UK, Chowdhury et al. (2022) point out that collaborative intelligence and
Vol. 26 • No. 65• February 2024 37

innovation can be acquired within organisations through a better understanding of AI systems,

but also through a better utilisation of employee competencies obtained through knowledge
sharing. A recent OECD study (2023) reveals a higher level of enthusiasm for AI upskilling
among AI user groups with specialised competencies, mainly young people under 24 years of
age with higher education. At the same time, by automating repetitive activities, AI allows the
concentration of professionals in areas focused primarily on innovation, allowing the
development of specific competencies such as cognitive, social, planning, or adopting
complex decisions (Verma and Singh, 2022). On the other hand, the development of AI
competencies allows the enhancement of unrealistic optimism, such as the perception that the
implementation of AI would have a superior positive impact within the company in relation
to competitors (Weber, 2023). From the above, we formulate the following hypothesis:
Hypothesis 2 (H2): Optimism and innovativeness in the use of AI tools mediates the
relationship between employee competencies and their AI effectiveness.
The adoption of AI solutions generates numerous risks and barriers to implementation, such
as those related to discomfort and insecurity when using AI tools. The degree of awareness
of these risks is correlated with the level of development and influences the effectiveness of
the use of AI tools. A study (Brougham and Haar, 2020) conducted on a sample of 1516
employees from the US, Australia, and New Zealand highlights the impact of technological
disruption on employee well-being and occupational stress in the context of job insecurity.
In research carried out in the US and India, Lingmont and Alexiou (2020) show that the
perception of job insecurity is strongly influenced by education and the level of skill
development, in the sense that the discomfort felt is higher in the case of employees with
lower proficiency levels. Li et al. (2023) show that AI exerts a greater negative impact on
workplace learning for employees with lower levels of education, placed in positions with
low decision-making autonomy, or with little work experience. According to Brougham and
Haar (2020), training is an effective way to manage automation when it threatens job security.
Furthermore, the provision of retraining services represents an expectation of employees in
the context of their perception of job insecurity (Lingmont and Alexiou, 2020). According to
the OECD (2023), fear of job loss is a motivating factor for employees to participate in
competence development programs. Finally, Colombo et al. (2019) emphasise the role of
social and digital competencies in moderating the negative impact of technology on jobs.
Therefore, we formulate the following hypothesis:
Hypothesis 3 (H3): The discomfort and insecurity associated with the use of AI tools mediate
the relationship between employee competencies and their AI effectiveness.
Based on the three assumptions mentioned above, the proposed concept model (figure no. 1)
integrates the existing correlations between the AI competence system AI readiness and
effectiveness in the use of AI, on the one hand, and the challenges launched to the FE and
NFE systems, against the background of the continuous development of AI systems and
technologies, as well as the increase in the degree of their use, both at the level of the business
environment and within the education system.
AE
Figure no. 1. Conceptual model on relationships

between employee competencies system, AI readiness and AI effectiveness
3. Materials and methods

3.1. Data collection
To achieve the purpose of the research, an online questionnaire was implemented, that was
available to fill out between August and September 2023, using the Google Forms platform.
The target population for this survey consisted of people who had a job in Romania at the
time of the study. The questionnaire was distributed online, reaching a total of
236 participants. However, the number of eligible respondents was adjusted to 208.
The sample size was considered sufficient in the context of structural equation modelling
(PLS-SEM), both superior to that indicated by the rule suggesting that it should be greater
than ten times the number of independent variables of the most complex regression model
(30), as well as the one indicated by Hair et al. (2022) for an expected minimum value of
statistically significant path coefficients between 0.11 and 0.20 and a significance level of
5% (155).
Regarding the respondents, the majority were women (51%) and had an average tenure of
11 years, mainly in executive positions (68%). Furthermore, most of the respondents came
from public sector organisations (55%) and their fields of activity varied, including
education, public administration, health and social care, and other services.
3.2. Scales
The questionnaire used for data collection included three scales referring to “AI readiness”,
“effectiveness in using AI tools” and “Employee competencies”. Each scale was designed and
Vol. 26 • No. 65• February 2024 39

evaluated using items measured from 1 to 5. The construction of these scales was guided by an
analysis of the specialised literature, and a detailed description of them is presented below.
Competencies (COMP): This scale includes different types of competence, namely, digital
competencies (COMP_DIG), technical competencies (COMP_TEHN), managerial
competencies (COMP_MAN), personal competencies (COMP_PERS), and social
competencies (COMP_SOC). This approach was taken from the work of Kowal et al., (2022)
and was adapted to suit the specifics of the research.
AI readiness was analysed through a set of items that measured the level of optimism and
innovation (AI_OPT_INN), as well as the level of discomfort and insecurity felt when using
AI tools (AI_DIS_INS), developed based on existing scales (Parasuraman, 2000;
Parasuraman and Colby, 2015; Chen and Chang, 2023) and adapted in the context of AI use.
Effectiveness in using AI tools (SF): As suggested in the literature (Bandura, 1994; Girasoli
and Hannafin, 2006), this scale focuses on assessing the degree of effectiveness reported by
employees in the context of using AI technology. This assessment included aspects related
to their competencies in solving complex problems and learning and applying these tools
(SF_EXP). Also included in the scale were items that investigated how AI users might
manage these tools in the future (SF_SP).
3.3. Data analysis

For data analysis, two advanced techniques were used: PLS-SEM and NCA. To perform the
two analyses, SmartPLS v4 (Ringle et al., 2022), known for its ability to manage complex
models and its flexibility in analysing the relationships between latent and observed
variables, was used. The synergistic use of PLS-SEM and NCA provided the opportunity to
better support the theoretical foundation, being particularly useful in identifying the set of
competencies that contribute to increasing the effectiveness in using AI tools, but also
absolutely necessary to obtain a given level of it, in concordance with the sufficiency and
necessity logic (Richter et al., 2020).
4. Results
4.1. Evaluation of the model
To support the reliability and validity of the measurement model, we applied the suggestions
proposed by Hair et al. (2020) and analysed the model in terms of internal consistency,
convergent validity, and discriminant validity. Therefore, analysing the data in Table no. 1,
we can see that the Cronbach Alpha coefficient and the composite reliability (rho_a and
rho_c) are above the recommended value of 0.70 (Hair et al., 2019; Hair et al., 2020). We
also analysed convergent validity via average variance extracted (AVE) and external
loadings. According to the recommendations of Hair et al. (2020), AVE values exceeded the
threshold of 0.50 and external loadings (with one exception) were greater than 0.708.
AE
Table no. 1. Fidelity and validity of the measurement model

Constructs Loadings Cronbach alpha rho_a rho_c AVE
COMP* 0.853 – 0.908 0.972 0.974 0.974 0.569
COMP_DIG 0.634 – 0.878 0.917 0.926 0.933 0.636
COMP_MAN 0.892 – 0.925 0.928 0.930 0.949 0.823
COMP_PERS 0.699 – 0.906 0.932 0.935 0.946 0.715
COMP_SOC 0.895 – 0.915 0.924 0.925 0.946 0.813
COMP_TEH 0.850 – 0.907 0.936 0.937 0.949 0.757
AI_OPT_INN* 0.922 – 0.934 0.920 0.925 0.935 0.646
AI_INN 0.741 – 0.893 0.866 0.881 0.909 0.715
AI_OPT 0.870 – 0.910 0.909 0.910 0.936 0.786
AI_DIS_INS* 0.826 – 0.906 0.836 0.839 0.877 0.507
AI_DIS 0.750 – 0.834 0.823 0.825 0.883 0.654
AI_INS 0.798 – 0.874 0.776 0.782 0.870 0.690
SF* 0.946 – 0.970 0.958 0.959 0.963 0.687
SF_EXP 0.879 – 0.898 0.909 0.910 0.936 0.785
SF_SP 0.741 – 0.901 0.936 0.938 0.949 0.726
Note: * – second-order construct; rho_c – composite reliability; rho_a – adjusted composite reliability;
AVE– average variance extracted.
The discriminant validity was assessed through the Fornell Larker criterion and the
Heterotrait-Monotrait ratio (HTMT) (Table no. 2). With two exceptions, these criteria
indicate the discriminant validity of the measurement model, as evidenced by HTMT values
lower than 0.9 (Hair Jr, Howard. and Nitzl, 2020; Henseler, Ringle and Sarstedt, 2015) and
correlations between constructs lower than the square root of AVE. However, the two
constructs were retained in the model, as they represent concepts with similar meanings (Hair
et al., 2022).
Table no. 2. Discriminant validity
Constructs 1 2 3 4 5 6 7 8 9 10 11
AI_DIS 0.809 0.417 0.626 0.178 0.200 0.126 0.171 0.239 0.148 0.184 0.199
AI_INN 0.340 0.846 0.099 0.799 0.638 0.496 0.514 0.467 0.598 0.711 0.658
AI_INS 0.509 0.027 0.831 0.131 0.076 0.070 0.190 0.244 0.070 0.123 0.108
AI_OPT 0.139 0.724 -0.065 0.886 0.515 0.410 0.421 0.391 0.490 0.742 0.695
COMP_DIG 0.132 0.557 0.026 0.466 0.798 0.715 0.698 0.650 0.892 0.630 0.634
COMP_MAN 0.106 0.438 0.011 0.376 0.671 0.907 0.871 0.792 0.791 0.576 0.541
COMP_PERS 0.141 0.451 0.148 0.387 0.659 0.813 0.846 0.928 0.738 0.606 0.566
COMP_SOC 0.205 0.410 0.200 0.360 0.616 0.735 0.861 0.902 0.673 0.547 0.532
COMP_TECH 0.105 0.536 -0.037 0.453 0.827 0.740 0.690 0.629 0.870 0.598 0.609
SF_EXP 0.147 0.637 -0.035 0.676 0.573 0.527 0.557 0.503 0.552 0.886 0.915
SF_SP 0.169 0.600 -0.026 0.643 0.585 0.504 0.528 0.496 0.571 0.845 0.852
Note: bottom left – Fornell-Larcker criterion; top right – HTMT report.
As suggested by Hair Jr, Howard and Nitzl (2020), the structural model was evaluated from
the perspective of collinearity, statistical relationships between constructs and coefficients of
determination. Therefore, according to the determination coefficient R2 = 0.599,
approximately 59.90% of variance in effectiveness in the use of AI tools can be explained by
optimism and innovation, discomfort and insecurity and the competencies of the
employees. Furthermore, variance inflation factor (VIF) values below 3 for all predictors in
the structural model indicate that multicollinearity is not a concern in this case.
Vol. 26 • No. 65• February 2024 41

4.2. Assessment of research hypotheses

To evaluate the three research hypotheses, both direct effects between competencies and
effectiveness in using AI tools and indirect effects mediated by AI optimism and innovation
and AI discomfort and insecurity were analysed (Table no. 3).
Table no. 3. Direct and indirect effects
Hypotheses Relations Beta SE T statistic BCCI f2
H1 COMP → SF 0.360*** 0.065 5.491 0.251; 0.467 0.224
COMP→AI_OPT_INN 0.549*** 0.058 9.476 0.441; 0.632 0.431
H2
AI_OPT_INN→SF 0.522*** 0.061 8.486 0.414; 0.617 0.470
COMP→AI_OPT_INN→SF 0.286*** 0.046 6.275 0.216; 0.367 -
COMP→AI_DIS_INS 0.138* 0.077 1.795 0.013; 0.264 0.020
H3 AI_DIS_INS→SF -0.037 0.048 0.775 -0.115; 0 -.042 0.003
COMP→AI_DIS_INS→SF -0.005 0.008 0.664 -0.024; 0.003 -
Note: BCCI - bootstrap confidence interval; f2 - effect size.; * - p < 0.05; *** - p < 0.001.
According to the results, there is a statistically significant and positive relationship
(β = 0.360; p < 0.001) between competence and the effectiveness of the use of AI tools,
supporting hypothesis H1. The results also support H2 and show that optimism and
innovation in the use of AI tools significantly and positively mediate the relationship between
employee competencies and their effectiveness in the use of these tools (β = 0.286; p < 0.001).
In other words, employee competencies positively influence optimism and innovation in the
use of AI, which in turn contributes to a significant increase in effectiveness. At the same
time, the results do not support H3, as discomfort and insecurity in using AI tools do not
mediate the relationship between employee competencies and AI effectiveness (β = -0.005;
ns). Consequently, the results indicate that as employee competencies evolve, the level of
discomfort and insecurity in using AI tools increases, resulting in a decrease in the
effectiveness of using AI tools.
4.3. The “should have” and “must have” competencies in the interaction with AI tools
The analysis was extended to identify the set of competencies sufficient and necessary for
employees to use AI tools. Therefore, two perspectives were considered. The first perspective
had in mind the identification of those skills that are able to lead to an increase the employees’
readiness in using AI tools, thus applying the “should have” (sufficiency) logic (Table no. 4),
and the second considered the absolutely necessary level of each category of competencies,
applying the “must have” (necessity) logic (Table no. 5). The sets of “should have” and “must
have” competencies were identified for three target constructs: (1) Discomfort and insecurity,
(2) Innovativeness and optimism and (3) Effectiveness in using AI tools.
Table no. 4. “Should have” competencies in AI context
Discomfort Innovativeness Effectiveness in using AI
Independent
and insecurity and optimism tools
variables
Beta t Beta t Beta t
Constant 0.000 7.786 0.000 3.687 0.000 1.916
AI_DIS 0.028 0.491
AI_INN 0.130* 1.739
AI_INS -0.054 0.977
AE
Discomfort Innovativeness Effectiveness in using AI
Independent
and insecurity and optimism tools
variables
Beta t Beta t Beta t
AI_OPT 0.400*** 6.058
COMP_DIG 0.093 0.751 0.317** 3.003 0.157* 1.877
COMP_MAN -0.172 1.342 -0.026 0.240 0.016 0.184
COMP_PERS 0.027 0.169 0.107 0.792 0.147 1.399
COMP_SOC 0.378** 2.814 0.019 0.169 0.063 0.695
COMP_TECH -0.153 1.141 0.203* 1.770 0.047 0.527
R2 0.081 0.327 0.608
F 3.577 19.608 34.158
Note: regression coefficients (Beta) with significant values are highlighted in gray. * - p < 0.05;
** - p < 0.01; *** - p < 0.001.
The results show a significant and positive relationship between social competencies and the
level of discomfort and insecurity (𝛽 = 0.378; t = 2.814; p < 0.01). This fact indicates that the
implementation of AI in a work environment can change the way employees communicate,
make decisions and work as a team. These changes can create a sense of uncertainty and
discomfort among employees, especially if they have not developed the social competencies
to adapt to these changes.
Furthermore, there is a significant and positive relationship between digital competencies
(𝛽 = 0.317; t = 3.003; p < 0.01) and technological competencies (𝛽 = 0.203; t = 1.770;
p < 0.05) and the level of innovation and optimism. In other words, when employees have
more developed digital and technological competencies, this is associated with a significant
and positive increase in their level of innovation and optimism in the use of AI tools.
Last but not least, there is a positive and significant relationship between optimism and
employee innovation in the effective use of AI tools (𝛽 = 0.400; t = 6.058; p < 0.001;
𝛽 = 0.130; t = 1.739; p < 0.05). There is also a positive and significant effect of digital
competencies (𝛽 = 0.157; t = 1.877; p < 0.05). Therefore, employees with well-developed
digital competencies can have a deeper understanding of how to use these tools and have
more confidence in their abilities, which can lead to greater effectiveness in using AI tools.
Next, to identify and highlight the set of necessary conditions for each type of competence,
two distinct thresholds were used: 50% and 100% and for each, an evaluation scale from 1
to 5 (Table no. 5).
Table no. 5. “Must have” competencies in AI context
Discomfort Innovativeness Effectiveness
Independent
and insecurity and optimism in using AI tools
variables
50% 100% 50% 100% 50% 100%
AI_DIS NN 1.514
AI_INN NN 1.679
AI_INS NN NN
AI_OPT NN 2.525
COMP_DIG NN 3.243 NN 2.419 NN 2.862
COMP_MAN NN 2.766 NN 2.064 NN 2.064
COMP_PERS NN 3.561 2.252 3.548 2.143 3.932
COMP_SOC NN 4.733 2.249 4.058 2.249 4.058
COMP_TECH NN 3.142 NN 3.179 NN 2.862
Note: *NN = not necessary. Necessary conditions with significant effect size (d) are marked in gray.
Vol. 26 • No. 65• February 2024 43

The results show that to feel a degree of discomfort and insecurity of 50% when using AI
tools, there is no need for competencies, while to feel a degree of discomfort of 100%, there
is a need for social competencies of 4.733 on a rating scale from 1 to 5. Additionally, to
achieve a 50% level of innovation and optimism in the use of AI tools, two types of
competencies are both necessary: personal competencies (2.252) and social competencies
(2.249). However, to achieve a 100% level of innovation and optimism in the use of AI tools,
personal competencies (3.548) and social competencies (4.058) were identified as
necessary. Regarding the need for employees to be effective in using AI tools, for the 50%
level, there were two types of necessary competencies, namely personal competencies
(2.143) and social competencies (2.249). However, to achieve a 100% effectiveness level, all
types of competencies were identified as cumulatively required, as well as a level of optimism
on the part of employees regarding the use of AI tools.
5. Discussions
Through this research, the aim was to analyse the relationship between employees'
competencies and effectiveness in the use of AI tools to identify possible directions of
managerial intervention aimed at improving the EF and ENF systems in order to reduce the
gaps between competencies available and needed in the labour market amid the expansion of
AI-based systems.
Therefore, the results of empirical research highlighted the fact that there is a significant
relationship between competence and the effectiveness of the use of AI tools, which supports
hypothesis H1. Specifically, employee competencies, including their digital, technical,
managerial, personal, and social competencies, have been associated with significantly
improved effectiveness in the use of AI tools. Previous research has discussed the importance
of employee competencies in using advanced technologies, including AI tools (Younis and
Adel, 2020; Jaiswal et al., 2021). This research has shown that individual competencies and
competencies have a significant impact on the adoption and effectiveness of technologies
within organisations. This fact is supported by studies by He et al. (2007) and Zaman et al.
(2019) that investigated, for example, the impact of technical competencies on performance
in the use of complex software or the impact of social competencies on collaboration in
technological projects. In addition, in the context of rapid changes in the technological
landscape, employees' competencies not only influence how they can use and benefit from
technologies such as AI but also play a significant role in organisations' ability to quickly
adapt to technological changes. Employees with developed digital, technical, and social
competencies are more likely to adopt and integrate new technologies into their work
processes, thus contributing to increased organisational efficiency and innovation (Jaiswal et
al., 2021).
Furthermore, the empirical results of the study show that optimism and innovation in the use
of AI tools significantly and positively mediate the relationship between competencies and
their effectiveness in the use of these IT tools, thus supporting hypothesis H2. Previous
research, such as that of Verma and Singh (2022), shows that AI improves decision-making
skills among employees, enhancing their creativity and innovation ability. The involvement
of employees in innovative activities increases the level of trust (Cheng et al., 2023) thus
favouring not only the increase in effectiveness in the use of AI tools but also the generation
AE
of positive results at the organisational level in the era of AI technologies (Braganza et al.,
2020). Moreover, technological skills can facilitate communication and collaboration
between departments and organisational teams. The effective use of AI tools can lead to faster
and more efficient information sharing, creating a synergistic work environment, and
fostering collaborative innovation (Olan et al., 2022).
Furthermore, the results of the study indicated that discomfort and insecurity in using AI
tools do not significantly mediate the relationship between employee competencies and
effectiveness, rejecting the last hypothesis (H3) of the investigation. Although there may be
some initial discomfort or insecurity in using AI tools, these feelings have not been found to
significantly impact employee effectiveness in implementing and using these technologies.
Organisations that implement AI can provide resources, support and training programs to
help employees overcome initial uncertainty or discomfort (Zhu et al., 2020).
The synergistic use of PLS-SEM and NCA facilitated a significant expansion of the
discussion of various types of competencies essential and necessary to ensure effective use
of AI tools. Thus, based on the results, we can highlight the following aspects.
Digital competencies play a key role in the use of AI tools, so typically AI models are based
mainly on digital competencies, as suggested by Fernandez Sanz et al. (2017). The results of
the present study show that there is a positive relationship between digital competencies and
the level of innovation and optimism. Furthermore, to achieve a 100% effectiveness level in
the use of AI tools, digital competencies have been identified as necessary. Consequently,
amid the rapid evolution of AI technology, characterised by the constant emergence of new
techniques and tools, digital competencies have become an essential tool for employees,
allowing them to stay abreast of these developments and integrate them into their work
activities. day by day (Van Laar et al., 2017). Several studies (Kispeter, 2018; Anisimova et
al., 2021) show that digital competencies are increasingly important for professional success
and active participation in modern society. The ability to work with AI tools can have a
significant impact on employment and personal development opportunities (European
Commission, 2021).
Managerial skills were evaluated as those at the individual level, so the empirical results of
the study show that there is no significant relationship between this type of skill and the
effective use of AI tools. However, following the NCA analysis, these competencies were
identified as necessary to achieve a level of 100% effectiveness in the use of AI tools.
Although individual managerial competencies may seem less relevant to the technical use of
AI tools, they can play a key role in coordinating and motivating teams working with this
technology (Sousa and Rocha, 2019). Research by Cortellazzo et al., (2019) showed that
leadership and communication competencies can influence how a team approaches and uses
technology.
Personal competencies in the use of AI refer to the set of competencies, knowledge and
personal traits needed to interact effectively with AI technologies and to take advantage of
their full potential. These competencies are not technical in nature, but focus more on
behavioural, cognitive and relational aspects (Arun and Carma, 2021). The study results
revealed that to reach a level of 50% in terms of innovation and optimism in the use of AI
tools, personal competencies must reach an average value of 2.252 on a rating scale of 1 to
5. Furthermore, it was found that to achieve 100% innovation and optimism in the use of AI
Vol. 26 • No. 65• February 2024 45

tools, personal competencies were identified as absolutely necessary. This can be supported
by previous research (Chowdhury et al., 2022; Verma and Singh, 2022), which has shown
that behavioural and psychological factors such as trust in technology and adaptive
competencies can significantly influence technology performance and adoption. Last but not
least, both to reach a 50% effectiveness level and to reach a maximum level of 100% in the
use of AI tools, personal competencies are required. Consequently, personal competencies
such as critical thinking, adaptability and ethics can contribute to a more efficient and
responsible use of AI tools, a fact supported by other specialist studies (Younis and Adel,
2020). These competencies can influence how people understand, use and manage AI
technology in different contexts.
Social competence in the use of AI refers to the personal competencies and traits that enable
people to interact and work effectively in different social and collaborative settings. The
results of the present study showed that there is a significant and positive relationship
between this type of competence and the level of discomfort and insecurity felt by employees
when using AI tools. This fact indicates that the implementation of AI in a work environment
can change the way employees communicate, make decisions, and work as a team. It was
also found that to feel a 100% degree of discomfort when using AI tools, a social competence
of 4.733 was recorded as necessary on a rating scale of 1 to 5. This can be justified by the
fact that the use of AI technologies can lead to a decrease in human interactions, including
communication and collaboration with colleagues, in an environment where AI is present.
However, social competencies were identified as necessary both to reach a level of 50% and
100 in the case of innovation and optimism in the context of AI and to use these IT tools.
Technological competence plays an important role in the use of technologies, as
demonstrated by numerous studies (Younnis and Adel, 2020; Jaiswal et al., 2021). The results
of this research revealed a positive relationship between technological competencies and the
level of innovation and optimism felt by employees in the context of the use of AI tools.
Additionally, technological competencies were identified as absolutely necessary to achieve
a 100% level of effectiveness in the use of these tools. Therefore, as employees gain a deeper
understanding of how AI works, they develop a more pronounced sense of optimism about
its capabilities (Chowdhury et al., 2022). With strong technological competencies, employees
can contribute to the development and implementation of innovative solutions that use AI to
tackle complex problems or make significant improvements in various fields (Verma and
Singh, 2022).
Conclusions
This research focused on examining the role that different types of competencies play in the
effective use of AI technologies in a context characterised by the rapid evolution of the work
environment in the digital age. The research successfully achieved its purpose by revealing
significant relationships between employees’ competencies and their effectiveness in using
AI tools. The obtained results provide a clearer understanding of how digital, technical,
managerial, personal and social skills influence employees’ performance with AI tools,
identifying potential interventions in both FE and NFE.
AE
Regarding the research hypotheses, H1 and H2 were confirmed, while H3 was rejected.
Furthermore, the NCA analysis highlighted the necessary and sufficient conditions for each
type of competency required by employees to use AI tools.
Theoretical contributions. This paper introduces novel theoretical contributions to the
specialised literature in the field of AI implications. The integration of two advanced analysis
techniques, namely PLS-SEM and NCA, adds significant value to understanding the intricate
relationships between employees’ competence and the effectiveness of using AI tools.
Through the combined application of these methods, the paper identifies two types of
competence—categorised as “should have” and “must have”—which emerge as essential in
the context of utilising these technological tools. Consequently, this work offers substantial
implications for the existing specialised literature on the adaptation of organisations and
employees to AI, as well as for FE and NFE.
Managerial contributions. The research results delineate potential courses of action for public
authorities and the business environment to adapt the Lifelong Learning (LLL) system to the
demands and challenges posed by AI. Based on best practices, public strategies and policies
should prioritise the implementation of strategic options to improve the Education and
Training Framework (EF) system. These options include tailor-made AI curriculum
development, the establishment of AI competence centers, the promotion of public-private
partnerships, and the development of occupational standards. Simultaneously, the research
findings, particularly in understanding the relationship between employee competence and
effectiveness in using AI tools, offer valuable insights for enhancing human resources
policies within the Employment and National Framework (ENF) system. This involves
creating AI-based competence models, establishing support centers, implementing specific
training and coaching programs and introducing change management initiatives to mitigate
uncertainty and discomfort among employees.
Limits and future research directions. As AI continues to play an increasingly prominent role
in society, future research could focus on two crucial issues: the ethical considerations
surrounding AI use and its broader impact on organisations. Furthermore, to achieve more
generalisable results, future studies may involve a larger number of respondents, thereby
addressing the limitations of our current research.
References
Anisimova, T., Ganeeva, A. and Sharafeeva, L., 2021. Development of Digital Skills and
Engineering Thinking in Students as Part of the Digital Summer Project. International
Journal of Engineering Pedagogy, [e-journal] 11(2), article no. 69.
https://doi.org/10.3991/ijep.v11i2.17215.
Bandura, A., 1994. Self-efficacy. In: V.S. Ramachaudran ed., 1994. Encyclopedia of human
behavior. New York: Academic Press, pp. 71-81.
Braganza, A., Chen, W., Canhoto, A. and Sap, S., 2021. Productive employment and decent
work: The impact of AI adoption on psychological contracts, job engagement and
employee trust. Journal of Business Research, [e-journal] 131, pp. 485-494.
https://doi.org/10.1016/j.jbusres.2020.08.018.
Vol. 26 • No. 65• February 2024 47

Brougham, D. and Haar, J., 2020. Technological disruption and employment: The influence
on job insecurity and turnover intentions: A multi-country study. Technological
Forecasting and Social Change, [e-journal] 161, article no. 120276. https://doi.org/
10.1016/j.techfore.2020.120276.
Chatterjee, S., Chaudhuri, R., Vrontis, D., Thrassou, A. and Ghosh, S.K., 2021. Adoption of
artificial intelligence-integrated CRM systems in agile organizations in India.
Technological Forecasting and Social Change, [e-journal] 168, article no. 120783.
https://doi.org/10.1016/j.techfore.2021.120783.
Chen, J. and Chang, Y.-W., 2023. How smart technology empowers consumers in smart retail
stores? The perspective of technology readiness and situational factors. Electronic
Markets, [e-journal] 33(1), article no. 1. https://doi.org/10.1007/s12525-023-00635-6.
Chen, N., Li, Z. and Tang, B., 2022. Can digital skill protect against job displacement risk
caused by artificial intelligence? Empirical evidence from 701 detailed occupations.
PLOS ONE, [e-journal] 17(11), article no. e0277280. https://doi.org/10.1371/
journal.pone.0277280.
Cheng, B., Lin, H. and Kong, Y., 2023. Challenge or hindrance? How and when
organizational artificial intelligence adoption influences employee job crafting. Journal
of Business Research, [e-journal] 164, article no. 113987. https://doi.org/10.1016/
j.jbusres.2023.113987.
Chowdhury, S., Budhwar, P., Dey, P.K., Joel-Edgar, S. and Abadie, A., 2022. AI-employee
collaboration and business performance: Integrating knowledge-based view, socio-
technical systems and organisational socialisation framework. Journal of Business
Research, [e-journal] 144, pp. 31-49. https://doi.org/10.1016/j.jbusres.2022.01.069.
Chowdhury, S., Dey, P., Joel-Edgar, S., Bhattacharya, S., Rodriguez-Espindola, O., Abadie,
A. and Truong, L., 2023. Unlocking the value of artificial intelligence in human resource
management through AI capability framework. Human Resource Management Review,
[e-journal] 33(1), article no. 100899. https://doi.org/10.1016/j.hrmr.2022.100899.
Colombo, E., Mercorio, F. and Mezzanzanica, M., 2019. AI meets labor market: Exploring
the link between automation and skills. Information Economics and Policy, [e-journal]
47, pp. 27-37. https://doi.org/10.1016/j.infoecopol.2019.05.003.
Cortellazzo, L., Bruni, E. and Zampieri, R., 2019. The Role of Leadership in a Digitalized
World: A Review. Frontiers in Psychology, [e-journal] 10, article no. 1938.
https://doi.org/10.3389/fpsyg.2019.01938.
Da Silva, L.B.P., Soltovski, R., Pontes, J., Treinta, F.T., Leitão, P., Mosconi, E., De Resende,
L.M.M. and Yoshino, R.T., 2022. Human resources management 4.0: Literature review
and trends. Computers & Industrial Engineering, [e-journal] 168, article no. 108111.
https://doi.org/10.1016/j.cie.2022.108111.
Eriksson, T., Bigi, A. and Bonera, M., 2020. Think with me, or think for me? On the future
role of artificial intelligence in marketing strategy formulation. The TQM Journal,
[e-journal] 32(4), pp. 795-814. https://doi.org/10.1108/TQM-12-2019-0303.
European Commision, 2018. The Impact of Artificial Intelligence on Learning, Teaching,
and Education. [online] Available at: <https://publications.jrc.ec.europa.eu/repository/
handle/JRC113226> [Accessed 5 September 2023].
AE
European Commision, 2021. Fostering a European approach to Artificial Intelligence.

[online] Available at: <https://digital-strategy.ec.europa.eu/en/library/communication-
fostering-european-approach-artificial-intelligence> [Accessed 5 September 2023].
Fareri, S., Fantoni, G., Chiarello, F., Coli, E. and Binda, A., 2020. Estimating Industry 4.0
impact on job profiles and skills using text mining. Computers in Industry, [e-journal]
118, article no. 103222. https://doi.org/10.1016/j.compind.2020.103222.
Fernández-Sanz, L., Gómez-Pérez, J. and Castillo-Martínez, A., 2017. e-Skills Match:
A framework for mapping and integrating the main skills, knowledge and competence
standards and models for ICT occupations. Computer Standards & Interfaces, [e-journal]
51, pp. 30-42. https://doi.org/10.1016/j.csi.2016.11.004.
Frey, C.B. and Osborne, M.A., 2017. The future of employment: How susceptible are jobs
to computerisation? Technological Forecasting and Social Change, [e-journal] 114,
pp. 254-280. https://doi.org/10.1016/j.techfore.2016.08.019.
Girasoli, A.J. and Hannafin, R.D., 2008. Using asynchronous AV communication tools to
increase academic self-efficacy. Computers & Education, [e-journal] 51(4),
pp. 1676-1682. https://doi.org/10.1016/j.compedu.2008.04.005.
Gupta, S., Ghardallou, W., Pandey, D.K. and Sahu, G.P., 2022. Artificial intelligence
adoption in the insurance industry: Evidence using the technology–organization–
environment framework. Research in International Business and Finance, [e-journal] 63,
article no. 101757. https://doi.org/10.1016/j.ribaf.2022.101757.
Hair, J.F., Howard, M.C. and Nitzl, C., 2020. Assessing measurement model quality in
PLS-SEM using confirmatory composite analysis. Journal of Business Research,
[e-journal] 109, pp. 101-110. https://doi.org/10.1016/j.jbusres.2019.11.069.
Hair, J.F., Hult, G.T.M., Ringle, C.M. and Sarstedt, M., 2022. A Primer on Partial Least
Squares Structural Equation Modeling (PLS-SEM). 3rd ed. Thousand Oaks, CA: Sage.
Hair, J.F., Risher, J.J., Sarstedt, M. and Ringle, C.M., 2019. When to use and how to report
the results of PLS-SEM. European Business Review, [e-journal] 31(1), pp. 2-24.
https://doi.org/10.1108/EBR-11-2018-0203.
He, J., Butler, B.S. and King, W.R., 2007. Team Cognition: Development and Evolution in
Software Project Teams. Journal of Management Information Systems, [e-journal] 24(2),
pp. 261-292. https://doi.org/10.2753/MIS0742-1222240210.
Henseler, J., Ringle, C.M. and Sarstedt, M., 2015. A new criterion for assessing discriminant
validity in variance-based structural equation modeling. Journal of the Academy of
Marketing Science, [e-journal] 43(1), pp. 115-135. https://doi.org/10.1007/s11747-014-
0403-8.
Holmström, J., 2022. From AI to digital transformation: The AI readiness framework.
Business Horizons, [e-journal] 65(3), pp. 329-339. https://doi.org/10.1016/
j.bushor.2021.03.006.
International Labour Organisation, 2023. Generative AI and Jobs: A global analysis of
potential effects on job quantity and quality. [online] Available at: <https://www.ilo.org/
global/publications/working-papers/WCMS_890761/lang--en/index.htm> [Accessed 5
September 2023].
Vol. 26 • No. 65• February 2024 49

Issa, H., Jabbouri, R. and Palmer, M., 2022. An artificial intelligence (AI)-readiness and
adoption framework for AgriTech firms. Technological Forecasting and Social Change,
[e-journal] 182, article no. 121874. https://doi.org/10.1016/j.techfore.2022.121874.
Jaiswal, A., Arun, C.J. and Varma, A., 2022. Rebooting employees: upskilling for artificial
intelligence in multinational corporations. The International Journal of Human Resource
Management, [e-journal] 33(6), pp. 1179-1208. https://doi.org/10.1080/09585192.
2021.1891114.
Kabashkin, I., Misnevs, B. and Puptsau, A., 2023. Transformation of the University in the
Age of Artificial Intelligence: Models and Competences. Transport and
Telecommunication Journal, [e-journal] 24(3), pp. 209-216. https://doi.org/10.2478/ttj-
2023-0017.
Kispeter, E., 2018. What digital skills do adults need to succeed in the workplace now and in
the next 10 years? The Warwick Research. [pdf] Available at:
<https://wrap.warwick.ac.uk/144740/7/WRAP-what-digital-skills-do-adults-need-
succeed-workplace-now-Kipester-2020.pdf> [Accessed 12 September 2023].
Kowal, B., Włodarz, D., Brzychczy, E. and Klepka, A., 2022. Analysis of Employees’
Competencies in the Context of Industry 4.0. Energies, [e-journal] 15(19), article no.
7142. https://doi.org/10.3390/en15197142.
Leyer, M. and Schneider, S., 2021. Decision augmentation and automation with artificial
intelligence: Threat or opportunity for managers? Business Horizons, [e-journal] 64(5),
pp. 711-724. https://doi.org/10.1016/j.bushor.2021.02.026.
Li, C., Zhang, Y., Niu, X., Chen, F. and Zhou, H., 2023. Does Artificial Intelligence Promote
or Inhibit On-the-Job Learning? Human Reactions to AI at Work. Systems, [e-journal]
11(3), article no. 114. https://doi.org/10.3390/systems11030114.
Liaw, W., Kueper, J.K., Lin, S., Bazemore, A. and Kakadiaris, I., 2022. Competencies for
the Use of Artificial Intelligence in Primary Care. The Annals of Family Medicine,
[e-journal] 20(6), pp. 559-563. https://doi.org/10.1370/afm.2887.
Lingmont, D.N.J. and Alexiou, A., 2020. The contingent effect of job automating technology
awareness on perceived job insecurity: Exploring the moderating role of organizational
culture. Technological Forecasting and Social Change, [e-journal] 161, article no.
120302. https://doi.org/10.1016/j.techfore.2020.120302.
Logg, J.M., Minson, J.A. and Moore, D.A., 2019. Algorithm appreciation: People prefer
algorithmic to human judgment. Organizational Behavior and Human Decision
Processes, [e-journal] 151, pp. 90-103. https://doi.org/10.1016/j.obhdp.2018.12.005.
Marinaș, C.V., Goia, S.I., Gora, A.A., Igreț, R.Ș., Roman, M. and Ștefan, S.C., 2021.
A Bibliometric Analysis of the Relationship Internship–Skills–Employ-Ability. In: S.n.,
The 15th International Management Conference, Managing People and Organizations in
a Global Crisis. N.d., Bucharest, Romania. S.l.: s.n.
Martínez-Plumed, F., Gómez, E. and Hernández-Orallo, J., 2021. Futures of artificial
intelligence through technology readiness levels. Telematics and Informatics, [e-journal]
58, article no. 101525. https://doi.org/10.1016/j.tele.2020.101525.
Olan, F., Ogiemwonyi Arakpogun, E., Suklan, J., Nakpodia, F., Damij, N. and Jayawickrama,
U., 2022. Artificial intelligence and knowledge sharing: Contributing factors to
AE
organizational performance. Journal of Business Research, [e-journal] 145, pp. 605-615.

Organisation for Economic Co-operation and Development, 2023. The impact of AI on the
workplace: Main findings from the OECD AI surveys of employers and workers. [online]
Available at: <https://www.oecd.org/publications/the-impact-of-ai-on-the- workplace-
main-findings-from-the-oecd-ai-surveys-of-employers-and-workers-ea0a0fe1-en.htm>
[Accessed 5 September 2023].
Organization for Economic Co-operation and Development, 2022. Recommendation of the
Council on Artificial Intelligence. [online] Available at: <https://legalinstruments.oecd.org/
en/instruments/oecd-legal-0449> [Accessed 5 September 2023].
Pan, Y. and Froese, F.J., 2023. An interdisciplinary review of AI and HRM: Challenges and
future directions. Human Resource Management Review, [e-journal] 33(1), article no.
100924. https://doi.org/10.1016/j.hrmr.2022.100924.
Parasuraman, A. and Colby, C.L., 2015. An Updated and Streamlined Technology Readiness
Index: TRI 2.0. Journal of Service Research, [e-journal] 18(1), pp. 59-74.
https://doi.org/10.1177/1094670514539730.
Parasuraman, A., 2000. Technology Readiness Index (Tri): A Multiple-Item Scale to
Measure Readiness to Embrace New Technologies. Journal of Service Research,
[e-journal] 2(4), pp. 307-320. https://doi.org/10.1177/109467050024001.
Qureshi, M.I., Khan, N., Raza, H., Imran, A. and Ismail, F., 2021. Digital Technologies in
Education 4.0. Does it Enhance the Effectiveness of Learning? A Systematic Literature
Review. International Journal of Interactive Mobile Technologies, [e-journal] 15(04),
article no. 31. https://doi.org/10.3991/ijim.v15i04.20291.
Richter, N.F., Schubring, S., Hauff, S., Ringle, C.M. and Sarstedt, M., 2020. When predictors
of outcomes are necessary: guidelines for the combined use of PLS-SEM and NCA.
Industrial Management & Data Systems, [e-journal] 120(12), pp. 2243-2267.
https://doi.org/10.1108/IMDS-11-2019-0638.
Ringle, C.M., Wende, S. and Becker, J.M., 2022. SmartPLS 4. Oststeinbek: SmartPLS
GmbH. [online] Available at: <http://www.smartpls.com> [Accessed 30 August 2023].
Sousa, M.J. and Rocha, Á., 2019. Skills for disruptive digital business. Journal of Business
Research, [e-journal] 94, pp. 257-263. https://doi.org/10.1016/j.jbusres.2017.12.051.
The Lisbon Council, 2022. What Artificial Intelligence Can Do for Workers and What Skilled
Workers Can Do for the World. [online] Available at: <https://lisboncouncil.net/ wp-
content/uploads/2022/11/LISBON-COUNCIL-RESEARCH_Skills_Innovation_and_
AI.pdf> [Accessed 5 September 2023].
Tilibașa, M.A., Boncilică, A.N., Popa, I., Ștefan, S.C. and Tărăban, I., 2023. Implications of
digital risks on teachers’ motivation and intention to use digital tools: a PLS-POS
perspective in Romanian preuniversity education system. Kybernetes, [e-journal] 52(13),
pp. 45-60. https://doi.org/10.1108/K-06-2023-1116.
United Nations Educational, Scientific and Cultural Organization, 2019. Beijing consensus
on artificial intelligence and education. [online] Available at: <https://unesdoc.
unesco.org/ark:/48223/pf0000368303> [Accessed 5 September 2023].
Vol. 26 • No. 65• February 2024 51

United Nations Educational, Scientific and Cultural Organization, 2021. AI and education.
Guidance for policy-makers. [online] Available at: <https://unesdoc.unesco.org/
ark:/48223/ pf0000376709> [Accessed 5 September 2023].
Van Laar, E., Van Deursen, A.J.A.M., Van Dijk, J.A.G.M. and De Haan, J., 2017. The
relation between 21st-century skills and digital skills: A systematic literature review.
Computers in Human Behavior, [e-journal] 72, pp. 577-588. https://doi.org/10.1016/
j.chb.2017.03.010.
Varma, A., Dawkins, C. and Chaudhuri, K., 2023. Artificial intelligence and people
management: A critical assessment through the ethical lens. Human Resource
Management Review, [e-journal] 33(1), article no. 100923. https://doi.org/
10.1016/j.hrmr.2022.100923.
Verma, S. and Singh, V., 2022. Impact of artificial intelligence-enabled job characteristics
and perceived substitution crisis on innovative work behavior of employees from high-
tech firms. Computers in Human Behavior, [e-journal] 131, article no. 107215.
https://doi.org/10.1016/j.chb.2022.107215.
Weber, P., 2023. Unrealistic Optimism Regarding Artificial Intelligence Opportunities in
Human Resource Management: International Journal of Knowledge Management,
[e-journal] 19(1), pp. 1-19. https://doi.org/10.4018/IJKM.317217.
World Bank Group, 2019. The changing nature of work. [online] Available at: <https://www.
worldbank.org/en/publication/wdr2019> [Accessed 5 September 2023].
Younis, R.A.A. and Adel, H.M., 2020. Artificial intelligence strategy, creativity-oriented
HRM and knowledge-sharing quality: Empirical analysis of individual and organisational
performance of AI-powered businesses. In: s.n., The Annual International Conference of
The British Academy of Management (BAM) 2020: Innovating for a Sustainable Future.
2-4 September 2020, London, United Kingdom. S.l.: s.n.
Zaman, U., Jabbar, Z., Nawaz, S. and Abbas, M., 2019. Understanding the soft side of
software projects: An empirical study on the interactive effects of social skills and
political skills on complexity – performance relationship. International Journal of Project
Management, [e-journal] 37(3), pp. 444-460. https://doi.org/10.1016/j.ijproman.
2019.01.015.
Zhang, Z., 2023. The HRM Model Based on Competency Model in the Context of New Age
Intelligence. Wireless Communications and Mobile Computing, [e-journal] 2023,
pp. 1-13. https://doi.org/10.1155/2023/6030183.
Zhu, Y.-Q., Corbett, J.U. and Chiu, Y.-T., 2021. Understanding employees’ responses to
artificial intelligence. Organizational Dynamics, [e-journal] 50(2), article no. 100786.
https://doi.org/10.1016/j.orgdyn.2020.100786.
AE
CHALLENGES OF ARTIFICIAL INTELLIGENCE ON THE LEARNING PROCESS

IN HIGHER EDUCATION
Nicoleta Sîrghi1* , Mirela-Catrinel Voicu2 , Grațiela Georgiana Noja3
and Oana-Ramona Socoliuc (Guriță)4
1)2)3)
West University of Timișoara, Timișoara, Romania.
4)
Alexandru Ioan Cuza University of Iași, Iasi, Romania.

Sîrghi, N., Voicu, M.C., Noja, G.G. and Socoliuc Received: 21 September 2023
(Guriță), O.R., 2023. Challenges of Artificial Revised: 13 November 2023
Intelligence on the Learning Process in Higher Accepted: 17 December 2023
Education. Amfiteatru Economic, 26(65), pp. 53-70.
Abstract
Artificial intelligence (AI) has expanded to unprecedented proportions in recent decades,
penetrating vast areas, including education. The debates on the usefulness of incorporating
AI into university education, with its subsequent opportunities and challenges, have
captured the attention of the current research agenda. The fruitful exploitation of the AI
advantages at the level of Romanian higher education is highly dependent on the specific
set of knowledge, competences and abilities, even the capacity of the system to adapt to
such a dynamic environment. Consequently, the objective of our research is to identify the
skills necessary for the specific digital learning environment of Romanian higher education
to encourage students as beneficiaries of the educational act to adopt AI technologies. Thus,
our methodology consists of structural equation modelling applied to an original data set
collected on the basis of a questionnaire addressed to undergraduate students from higher
economic education. Results emphasise that the intention to adopt applications using AI
among students directly depend on constructs such as perceived usefulness, attitude
towards these technologies, perceived hedonic value, expected performance, or degree of
compatibility, while the interactivity of the applications also has an important but indirect
influence.
Keywords: artificial intelligence, higher education, learning outcomes, digital skills,
structural equation modelling.
JEL Classification: D8, O33, I23, C15, J24
*
Corresponding author, Nicoleta Sîrghi – e-mail: [email protected]
Vol. 26 • No. 65 • February 2024 53

AE Challenges of Artificial Intelligence on the Learning Process in Higher Education
Introduction
Artificial intelligence (AI) is perceived as a research subfield belonging to computer science
whose nodal concern is the simulation, extension, and enhancement of human intelligence
(Shi and Zheng, 2006). From 1956 until the present, AI has followed an extraordinary
dynamic, with an unprecedented expansion in vast areas like industry, technology, economy,
research and development, implicitly education, where the target has been from the very
beginning higher education (European Commission, 2022). Currently, AI is facing an
unprecedented dynamic from the perspective of constantly evolving and improving in terms
of technology, with the United States, the European Union and China and being the global
leaders. Reasoning in terms of specific capabilities, AI fits into the pattern of a vast area of
machine learning based on intelligent behaviour built on automatic learning algorithms.
On the one hand, AI supports people by helping them extend their capabilities when dealing
with stressful environments and incomplete information, can provide real-time feedback, and
is useful in making predictions and recommendations with increased accuracy (Binu and
Rajakumar, 2021). On the other hand, the fabulous boost of AI also draws attention to the
other perspective, namely the challenges it entails referring to ethical concerns (Gînguță et
al., 2023), security, safety, or even fairness issues (Varshney, 2016).
In recent years, the use of AI in higher education has increased significantly, along with the
development of AI tools designed for both students and educators (Chu et al., 2022). These
AI tools provide valuable feedback to students on the next steps they need to take for a
personalised learning experience, based on their personal profile, learning outcomes,
progress, or context (Verdú et al., 2017). Other applications make it easier for students to
learn independently through interactive games (Dever et al., 2020) or even evaluate student
performance through computer-assisted evaluation in a fully transparent setting
(Baykasoğlu et al., 2018).
Taking into account this complex framework of analysis, the present research aims to
identify the challenges seen from the perspective of additional skills needed in the Romanian
digital learning environment that are necessary to encourage undergraduate students to use
AI technologies, thus improving the process of knowledge assimilation, with beneficial
spillover effects upon their insertion into the labour market, tailored to professional profiles.
Obviously, the ultimate goal is to have a positive influence on the performance of the
learning system, and for this, we have taken into account that this fact depends, in turn, on
the digital competence of the students and their propensity to accept AI technology in the
learning process. Therefore, after a comprehensive review of the literature, we have
identified the benefits and main challenges that accompany or constrain the absorption of
AI-based technologies in educational services, with a clear focus on higher education.
The paper aims to explore a less researched perspective, namely the skills that encourage
the adoption of AI among undergraduate students in Romania. From the literature review, it
can be observed that our country has a high potential for the adoption of AI technologies
through undergraduate curricula, being among the supporters of digital reform in higher
education. Our research endeavour was based, in the first stage, on the design of a
questionnaire addressed to undergraduate students in higher economic education to
determine their positioning towards AI technologies. After identifying essential
characteristics that would encourage the adoption of AI in education, a structural equation
model (SEM) was designed and implemented. The model assesses the interaction between
AI features and the intention to adopt AI, respectively, the attitude towards a series of
applications among Romanian undergraduate students.
AE
The results show that students' learning worries are associated with the consequences of
using AI, as evidenced by the positive influence of all parameters. It provides valuable
insights both for academics who want to use AI technologies to enhance the attractiveness
of their disciplines, as well as for policymakers who could shape educational reforms to
support digitalisation and applied learning, thus supporting the objectives of the EU 2030
Digital Decade strategy. A notable element of novelty of the research lies in the innovative
way in which students' intention to adopt AI technologies is captured and analysed, through
constructs such as their perceived usefulness of AI technologies, their perceived ability of
AI to evoke emotions or enjoyable experiences, their attitude towards AI applications,
expected performance or compatibility. Therefore, the interactivity of applications using AI
becomes a nodal factor in inducing the intention to use AI. The more interactive these
technologies become, the more attracted the learner will be, considering the pleasant
experiences, students absorbing the application, and finding it useful in the learning act.
The remainder of the paper is organised as follows: the first part presents a review of the
literature on artificial intelligence (AI) and its impact on the learning process in the higher
education system. The second section describes the variables used, the research hypotheses,
and the research methodology applied through structural equation modelling. The third
section is devoted to results and discussions. The final section summarises conclusions, but
also limitations, and future research directions.
1. Higher education and the challenges of artificial intelligence: a systematic review of

the literature
Artificial intelligence (AI) embeds multiple definitions in the relevant scientific literature.
Thus, AI is defined as an assembly of systems that denote intelligent behaviour involving
analysis of the environment and the ability to act, with some autonomy, to meet
predetermined goals (Boucher, 2020). Artificial intelligence belongs to computer science
and is attributed entirely to machines, computers (Pan, 2016). AI is designed to solve
specific tasks in all fields of activity undertaken by humans (Makridakis, 2017). In other
words, AI is the automated version of human intelligence (Fenwick and Molnar, 2022).
As part of the development of contemporary society, the use of AI in the context of
digitisation has become an intrinsic component of individuals' lives and education. In this
perspective, the higher education sector is frequently associated with highly dynamic
environments, interdependent with information systems or digitalisation, making it suitable
for the adoption of artificial intelligence technology for educational acts (Escotet, 2023).
Although AI is promoted as a tool that can be used to improve the lives of individuals to its
full potential, its implementation in higher education is strongly encouraged by the
advancement of information and communication technologies. Technological progress has
opened up a wide range of opportunities for both students and teachers, as the delivery of
educational materials online and the rapid distribution of information with a fast and secure
connection between the educational service provider and the beneficiary fuels increased
accessibility to tertiary education. Moreover, on-line learning technologies allow for a
greater degree of customisation for academic materials and the use of blended learning
models, giving students the freedom to study at their own pace while having flexible access
to courses. While in theory, the use of AI is easy, once transposed into practice, it can
produce several effects, which differ from one field of study to another. For this reason, the
Vol. 26 • No. 65 • February 2024 55

metamorphosis of the teaching-learning act involves a continuous effort from both sides,
teachers, and students (Escotet, 2023). Today, the pressure exerted by the use of digital
technology is determining an increased pace of change in universities, opening them up to
the world of digital skills. Thus, their competitiveness is often reflected in the ability to
update their disciplines, making them more attractive to students and available to all
through distance learning platforms. The traditional pattern of student presence in the
classroom must be replaced with a modern one in which technology captures the attention.
New forms of learning combine digital instruments with AI technologies such as
Grammarly, ChatGPT, Brainly, Mendeley, DeepL Write, or others that stimulate
problematisation, hands-on activities through interactive experiences such as virtual
classrooms, virtual apps or experiments, and access to various forms of online educational
resources, simulations, tutorials, or practical team projects (Pellas, Mystakidis and
Kazanidis, 2021).
However, the call for interactivity comes with some challenges. Thus, studies show that
reading, one of the essential instruments for human intellectual development, is gradually
being replaced by sounds or images, which makes the large-scale implementation of AI
pave the way toward perpetuating behaviour that can sabotage human intelligence in the
long run (Pan, 2016). This is not the only risk of widespread adoption of AI in education.
Moreover, we can mention issues such as invasion of personal space and security, as most
applications using AI request and collect users' personal data, sometimes even without their
consent. Furthermore, excessive automation can lead to individual isolation and decrease
the level of social interaction, severely diminishing the role of the teacher in the teaching-
learning act. Another aspect is the learning process limited to a specific context or set of
data, which could sabotage the human need for lifelong learning, implicitly ethical issues
(Gînguță et al., 2023), and moral ones (Saghiri et al., 2022). Based on these shortcomings,
which required a well-defined framework for deployment, the European Parliament and the
European Commission initiated in 2021 the first proposal to regulate AI at the EU level,
giving it precise guidelines for use based on a legal framework that would turn AI into a
support for European citizens while limiting its risks.
Education and training are strategic directions for the prosperity of all nations. Therefore, as
a complement to the Act, the European Commission (European Commission, 2021, p. 9)
has developed an AI Monitoring Index (2021) to include the most relevant dimensions for
policy makers, to target those areas that need further investment. University education is
part of this index, specifically the Societal Aspects pillar, and is assessed on three levels: (1)
AI in EU university programmes, which measures the intensity of use of AI in university
curricula, seen as an indicator of the AI skills of current students (future employees). From
this point of view, at the level of the 2020-2021 academic year, Romania had a modest
performance of only 5% of AI programmes in total undergraduate programmes (European
Commission, 2022, p.45); (2) places in AI-content universities, which shows the number of
places available in AI-content university programmes and provides an overview of the AI
skills acquired by the potential AI workforce trained in university education. In this respect,
Romania ranks third in Europe, after Germany and Poland, with more than 55.000
undergraduate places in universities; (3) AI intensity in places offered by universities, an
indicator showing the proportion of places available in university programmes with a high
AI content out of the total number of places. Even at this dimension, Romania occupies a
leading position, second after Estonia, with a share of about 14% in the total number of
undergraduate places (European Commission, 2022, p.47).
AE
Based on these considerations, higher education in Romania (undergraduate level) presents

a high potential for the implementation of technologies using AI, being on the same
wavelength as the EU digitisation strategy. Romania's openness towards AI offers the
necessary prerequisites for reforming the education system to support the digital skills and
training of the younger generation. This is in line with the objectives of the EU's Recovery
and Resilience Mechanism by ensuring that the workforce has access to retraining and
further qualification that is valid across the EU, based on the recognition of specialised
higher education programmes in AI in all member states (European Commission, 2021).
Given this potential concerning the openness toward interactivity and AI in the Romanian
university education system, which was dominated until recently by the traditional
teaching-learning formula, our study completes the overall framework with feedback
provided by the beneficiaries of the educational act. It reveals the defining constructs that
would lead students to adopt and use, with confidence, AI technologies in their university
training. These characteristics become all the more important, as the openness of
undergraduate programmes in Romania towards AI denotes boldness and creativity, first of
all, on the part of higher education institutions that are committed to a permanent exercise
of curriculum development and adaptation. Then, on the part of teachers, who must
redefine themselves in the effort to model competences by selecting the proper AI
applications able to add value to the educational act, in conditions of increased
attractiveness for their students.
2. Methodology
The main research objective of this study is to identify the attributes needed by the specific
digital learning environment for Romanian higher education that will allow students to
adopt AI technologies. Therefore, the research activity was designed using a cascade
approach, being fully planned and executed in stages. To achieve this objective, we rely on
newly compiled data and advanced econometric procedures that incorporate simultaneous
equation reasoning/structural equation modelling (SEM). Therefore, to gather the necessary
data, in the initial phase of the research, an online questionnaire was administered using
Google Forms to undergraduate students within the Faculty of Economics and Business
Administration, West University of Timisoara, between June and July 2023.
The data used in the SEM application were obtained by administering the online
questionnaire with questions based on 5-point Likert scale responses. Data processing and
analysis were performed using SmartPLS v4 econometric software. Thus, we obtained a
database with 267 valid responses, out of a total of 293 respondents – students in higher
education in economics, undergraduate level, the structure of the respondents being the
following: 51% female and 49% male. The allocation of respondents was first made
according to the year of study for the undergraduate level, thus: first-year students – 26%,
second-year students – 32%, and third year – 42%. Regarding the distribution according to
the birthplace counties, there were 140 students from Timiș county (52.43%), 29 students
from Arad (10.86%); 27 students from Hunedoara, (10.11%); 23 students from Caraș-
Severin (8.62%); 19 students from Gorj (7.12%); 17 students from Mehedinți (6.37%);
12 students from other counties (like Bihor, Vrancea, Maramureș, Botoșani, and Sălaj),
which represent 4.91%. The AI applications that were mostly used by students are
ChatGPT (all respondents); Bard or Microsoft Bing (62%); under 24% of the respondents
Vol. 26 • No. 65 • February 2024 57

use other AI applications like Grammarly, My AI on Snapchat, Wolfram Alpha, Duolingo,

Bloomai, Merlin AI, Tutor AI, Anthropic or Cohere.
In shaping the constructs that lead respondents to adopt AI applications, the following ten
attributes were taken into consideration in this research according to which a number of
12 hypotheses were established, which are illustrated in Table no.1
Table no. 1. The establishment of working hypotheses
No. Attributes Hypotheses
1-2 AI adoption intention refers to individuals' feelings, H1. Attitude towards AI
positive or negative, related to the achievement of the applications positively influences
target behaviour, respectively, AI adoption (Venkatesh the intention of AI adoption.
et al., 2003; Chatterjee and Bhattacharjee, 2020);
Attitude towards AI applications defines the positive or
negative emotions of a person related to a particular
activity or affective response triggered by an application
(Chauhan and Jaiswal, 2016; Chatterjee and
Bhattacharjee, 2020; Almaraz-López, Almaraz-
Menéndez and López-Esteban, 2023).
3 Perceived usefulness refers to the conscious benefits H2. Perceived usefulness has a
students would gain from using AI (Dubey and Sahu, positive impact on the intention to
2021; Malik et al., 2021). adopt AI.
4 The perceived hedonic value illustrates the personal and H3. The perceived hedonic value
subjective satisfaction gained from using a service has a positive effect on the
(Alam et al., 2023). intention to adopt AI.
5 Compatibility captures how appropriate AI is to the H4.Compatibility has a positive
student's values and habits (Rogers, 1962; Venkatesh et impact on the intention to adopt
al., 2003). AI.
6 Performance expectancy represents a construct that H5. Performance expectancy has
reflects a person's attitude toward the use of a new a positive influence on AI
system and its noticeable contribution to his increased adoption intention.
professional achievements (Venkatesh et al., 2003).
7 Immersion is an important construct as it allows the H6. Immersion positively
learner to be totally immersed in applications using AI, influences the attitude towards AI
practising real skills in a realistic setting but consumed applications.
in a virtual space (Luan et al., 2020; Hannan and Liu,
2023; Rodway and Schepman, 2023).
8 System quality refers directly to the openness of the H7. The quality of the system has
education system to e-learning (Maisha and Shetu, positive inferences on immersion.
2023).
9 Interactivity is the ability of AI technology to allow the H8. Interactivity positively
user to participate actively in the communication process influences immersion.
(Pillai et al., 2023). H9. Interactivity has a notable
inference on perceived usefulness.
H10. Interactivity significantly
shapes the perceived hedonic
value.
10 Personalisation is the construct that indicates the H11. Personalisation influences
usefulness of an application that is intended to enhance perceived usefulness.
the user's perceived value. Its role is to improve the H12. Personalisation influences
user's experience, helping him achieve greater the perceived hedonic value.
productivity (Asif and Krogstie, 2013).
AE
To test the research hypotheses, we designed a research model containing the constructs
listed above, namely: intention to adopt AI; attitude towards AI applications; perceived
usefulness; perceived hedonic value; compatibility; performance expectancy; immersion;
system quality; interactivity; and personalisation (Figure 1 and Figure 2).
Figure no. 1. The structural conceptual model

Note: IA (AI adoption intention); AT (Attitude towards AI applications); UP (Perceived usefulness);
VHP (Perceived hedonic value); CO (Compatibility); PA (Performance expectancy);
IM (Immersion); CS (System quality); INT (Interactivity); PER (Personalisation).
The research methodology consists of structural equation modelling (SEM) and analyses the
interaction between the features of the learning system and the willingness of the students to
accept AI technology in order to transfer part of their tasks into a digitisation-based automated
programme. This methodology belongs to multivariate statistics (Kaplan, 2001) and offers the
possibility to model complex relationships between observed or manifest variables, in our
case students' propensity to integrate AI technologies in the act of learning and latent
variables, respectively, that are also called constructs represented by the features of the
learning system, providing in-depth information of the correlations between them.
3. Results and discussion

To estimate the research model (Figure 1), partial least squares equation modelling (PLS-
SEM) (Hair et al., 2021) was applied using SmartPLS v4 statistical software. Thus, we
were able to determine the links between the variables and their corresponding latent
constructs in the estimation model and the relationships between the constructs in the
structural model. We also examined whether there is predictive relevance of endogenous
latent variables (Henseler, 2018). From the perspective of model validity and reliability, the
results of Table 2 confirm the internal consistency of the model presented in Figure 1.
Vol. 26 • No. 65 • February 2024 59

As presented in Table 2, the item loadings range from 0.701 to 0.934. These values are
greater than 0.70, meaning that they make a substantial contribution to the constructs
assigned to them (Hair et al., 2021). The recommended values for rho_c and rho_a are
greater than 0.70, with the Cronbach's Alpha coefficient associated with internal reliability
(Taber, 2018). Consequently, for each variable, Cronbach's Alpha >0.838; composite
reliability rho_a > 0.847; rho_c > 0.892. The values of the average variance extracted
(AVE) indicator show, on average, how much of the variance of the associated items can be
explained by the construct. The literature recommends values greater than 0.50. In our
research, the results synthesised in Table 2 entail an average variance extracted (AVE)
higher than 0.628.
Table no. 2. Validation of the constructs and scales used in the questionnaire
Adapted after α/rho_a/
Construct Code Item Loading
the authors rho_cAVE
ADI1 I will continue to use AI apps as Chatterjee and 0.701 0.862/0.874/
often as I do now. Bhattacharjee 0.902/0.649
(2020)
ADI2 Next time I have to work on a Chatterjee and 0.739
project with a colleague, I would Bhattacharjee
like to use AI applications. (2020)
AI adoption ADI3 I intend to use AI applications in Chatterjee and 0.854
intention the future when preparing my Bhattacharjee
projects. (2020)
ADI4 I would like to use AI applicationsChatterjee and 0.869
in the future when learning. Bhattacharjee
(2020)
ADI5 I would recommend to my friends Chatterjee and 0.864
to use AI applications when Bhattacharjee
learning. (2020)
ATT1 I love talking to friends, family, or Chauhan and 0.892 0.91/0.913/
colleagues about AI applications. Jaiswal, 2016 0.937/0.787
ATT2 I recommend AI applications to Chauhan and 0.897
friends, family or colleagues. Jaiswal, 2016
ATT3 I share information through socialAlmaraz-López, 0.869
Attitude towards networks that I have learnt using Almaraz-
AI AI applications. Menéndez and
applications López-Esteban,
2023
ATT4 I share content from AI Almaraz-López, 0.890
applications with colleagues. Almaraz-
Menéndez and
López-Esteban,
2023
PU1 AI applications Dubey and 0.897 0.901/0.902/
...improve my ability to learn. Sahu, 2021 0.931/0.772
PU2 ...I save time when I study. Malik et al., 0.842
Perceived
2021
usefulness
PU3 ... improve the way I search for Malik et al., 0.905
information. 2021
PU4 ...allow me to get information Malik et al., 0.870
faster. 2021
AE

PHV1 AI applications are Alam et al., 0.874 0.927/0.929/
...fun. 2023 0.948/0.821
PHV2 ... interesting. Alam et al., 0.923
Perceived
2023
hedonic value
PHV3 ... delightful. Alam et al., 0.934
2023
PHV4 ... exciting. Alam et al., 0.891
2023
CO1 AI applications fit well with my Venkatesh et al., 0.851 0.839/0.848/
learning style. 2003 0.904/0.758
CO2 With AI applications, I can learn Venkatesh et al., 0.928
Compatibility
the way I like. 2003
CO3 Using AI applications, I can find Venkatesh et al., 0.830
new ways to learn in an enjoyable 2003
way.
PE1 Using AI applications when Chatterjee and 0.921 0.87/0.874/
learning helps me to Bhattacharjee 0.920/0.793
... understand concepts faster. (2020)
Performance PE2 ... achieve higher grades. Chatterjee and 0.886
expectancy Bhattacharjee
(2020)
PE3 I believe that AI applications Chatterjee and 0.864
make learning more efficient. Bhattacharjee
(2020)
IMM1 While using AI learning Hannan and Liu, 0.880 0.904/0.908/
applications, I was 2023 0.929/0.725
...absorbed by what I was doing.
IMM2 ... immersed in the task that I wasHannan and Liu, 0.887
performing. 2023
IMM3 When they use AI applications, I Luan et al., 2020 0.883
Immersion can discover and explore many
concepts.
IMM4 I would like to spend more time Luan et al., 2020 0.817
with AI applications when I am
learning.
IMM5 Working with AI applications is Rodway and 0.784
fascinating. Schepman,
2023
SQ1 AI applications respond promptly Maisha and 0.876 0.887/0.888/
to my requests and deliver reliable Shetu, 2023 0.922/0.747
results.
SQ2 AI applications Maisha and 0.861
....perform their functions quickly Shetu, 2023
System
and efficiently.
quality
SQ3 ... are reliable (they are always up Maisha and 0.845
and running, they run error-free Shetu, 2023
and do what they are supposed to
do).
SQ4 ... provides accurate information. Maisha and 0.874
Shetu, 2023
Vol. 26 • No. 65 • February 2024 61


INT1 ... have interaction functions that Pillai et al., 2023 0.702 0.852/0.872/
help me make decisions when 0.893/0.628
solving problems.
INT2 I can interact with AI applicationsPillai et al., 2023 0.733
to get information tailored to my
specific needs.
Interactivity INT3 By interacting with AI Pillai et al., 2023 0.844
applications, I can get a deep
understanding of the concepts I
am learning.
INT4 AI applications have remarkable Pillai et al., 2023 0.845
interaction features.
INT5 I can interact with AI applicationsPillai et al., 2023 0.837
to get new information.
PER1 AI applications Asif and 0.818 0.857/0.862/
...fit my needs well when I am Krogstie, 2013 0.897/0.637
learning.
PER2 ... gives me information that is Asif and 0.794
useful to understand the concepts Krogstie, 2013
Personalisation I am studying.
PER3 ...understand my requirements. Asif and 0.832
Krogstie, 2013
PER4 ... understand what I want to learn. Asif and 0.825
Krogstie, 2013
PER5 ... allow me to learn at my own Asif and 0.716
pace. Krogstie, 2013
Rönkkö and Cho (2020) showed that, through discriminant validity, items correlate more
with each other than they correlate with other items from other constructs. In our research,
we used the Fornell-Larcker criterion to determine the discriminant validity, as entailed in
Table no. 3 below.
Table no. 3. Discriminant validity analysis (Fornell and Larcker criterion)

ATT SQ CO IMM ADI INT PE PER PU PHV
ATT 0.887
SQ 0.652 0.864
CO 0.713 0.749 0.871
IMM 0.744 0.73 0.781 0.851
ADI 0.725 0.77 0.77 0.741 0.806
INT 0.598 0.789 0.736 0.685 0.74 0.792
PE 0.62 0.684 0.651 0.568 0.757 0.711 0.891
PER 0.629 0.757 0.742 0.66 0.727 0.752 0.643 0.798
PU 0.567 0.775 0.686 0.656 0.802 0.762 0.747 0.719 0.879
PHV 0.669 0.743 0.758 0.757 0.802 0.764 0.697 0.654 0.736 0.906
AE
According to the data in Table no. 3, following the Fornell-Larcker criterion, it can be seen
that the square root of the average variance extracted is higher than the absolute value of
the correlation coefficients, thus confirming the discriminant validity of each construct in
the model. In assessing the structural model, the perspective offered by Streukens and
Leroi-Werelds (2016, p.626) that “from a statistical explanatory modelling point of view,
hypothesis testing is a critical element in developing a relevant and rigorous theory. In a
PLS-SEM context, hypothesis testing relies on bootstrapping”. The recommendations refer
to the use of P-values and bootstrap confidence intervals. If P ≤ 0.05, the hypothesis is
accepted; otherwise, it is rejected.
In our research, the bootstrapping procedure was implemented as a statistical resampling
method, often reported in the scientific literature because of its advantage to generate
pseudo-observations that will ensure a good convergence of the estimators towards real
values. We performed the robustness check for the accuracy of the results obtained after
processing our research model and concluded that all hypotheses are supported, as shown in
Table 4.
Table no. 4. Validity of the assumptions
Coefficients T
Direct effects P Value Hypotheses
of Effect statistic
Attitude towards AI applications -> 0.202 3.395 0.001 H1- supported
Adoption intention
Perceived usefulness -> Adoption 0.311 5.013 0 H2- supported
intention
Perceived hedonic value -> Adoption 0.224 3.775 0 H3- supported
intention
Compatibility -> Adoption intention 0.146 2.607 0.009 H4- supported
Performance expectancy -> Adoption 0.149 2.482 0.013 H5- supported
intention
Immersion -> Attitude towards AI 0.744 19.815 0 H6- supported
applications
Sistem Quality -> Immersion 0.502 6.147 0 H7- supported
Interactivity -> Immersion 0.289 3.215 0.001 H8- supported
Interactivity -> Perceived usefulness 0.509 7.724 0 H9- supported
Interactivity -> Perceived hedonic 0.336 5.732 0 H10- supported
value
Personalisation -> Perceived utility 0.625 10.968 0 H11- supported
Personalisation -> Perceived hedonic 0.185 3.334 0.001 H12- supported
value
In the evaluation of the structural model, the effects coefficients were used. According to
Nakagawa, Johnson and Schielzeth (2017), the decomposition of total effects coefficients
into direct and indirect effects takes place. In our case, the effect coefficients are presented
in Figure 2 of the structural model and Table 4, respectively.
Vol. 26 • No. 65 • February 2024 63

Figure no. 2. The structural model

Note: IA (AI adoption intention); AT (Attitude towards AI applications); UP (Perceived usefulness);
VHP (Perceived hedonic value); CO (Compatibility); PA (Performance expectancy); IM
(Immersion); CS (System quality); INT (Interactivity); PER (Personalisation)
Table 5 illustrates the total effects (consisting of direct and indirect effects) of the variables
in the structural model listed in Figure 2.
Table no. 5. Total effects
Coefficients T P
Total effects
of Effect statistic Value
Interactivitt -> Adoption intention 0.341 6.745 0.000
Perceived usefulness -> Adoption intention 0.311 5.013 0.000
Perceived hedonic value -> Adoption intention 0.224 3.775 0.000
Attitude towards AI applications -> Adoption intention 0.202 3.395 0.001
Immersion -> Adoption intention 0.150 3.300 0.001
Performance expected -> Adoption intention 0.149 2.482 0.013
Compatibility -> Adoption intention 0.146 2.607 0.009
Personalisation -> Adoption intention 0.146 4.046 0.000
Sistem Quality -> Adoption intention 0.075 2.761 0.006
Immersion -> Attitude towards AI applications 0.744 19.815 0.000
Sistem Quality -> Attitude towards AI applications 0.373 5.582 0.000
Interactivity -> Attitude towards AI applications 0.215 3.150 0.002
Sistem Quality -> Immersion 0.502 6.147 0.000
Interactivity -> Immersion 0.289 3.215 0.001
Interactivity -> Perceived usefulness 0.509 7.724 0.000
Personalisation -> Perceived usefulness 0.336 5.732 0.000
Interactivity -> Perceived hedonic value 0.625 10.968 0.000
Personalisation -> Perceived hedonic value 0.185 3.334 0.001
AE
The standardised root mean square residual (SRMR) reported in Table 5 was calculated in
SmartPLS. According to Henseler et al. (2014), values less than 0.10 are considered to
show a good model fit. In our research, we obtained an SRMR value of 0.069, which
indicates a good fit of the model. Stone-Geisser values (Q2) are 0.691 (Adoption intention),
0.623 (Perceived usefulness) and 0.59 (Perceived hedonic value), 0.554 (Immersion), and
0.426 (Attitude towards AI applications). They also describe the important predictive
relevance of the PLS path model.
The purpose of our research was to identify the additional skills needed in the Romanian
digital learning environment specific to higher education that would lead undergraduate
students to adopt AI technologies, thus improving their knowledge assimilation process.
Romania has significant potential from this point of view; the biggest challenge is to
identify the optimal approach in the selection of AI applications that can be easily
assimilated by students while bringing a real added value to the whole educational act. This
is a necessary transition to the digital era, and the fact that Romanian universities have
shown openness to insert AI into the matrix of the educational act since the academic year
2020-2021 offers not only proof of courage, but implicitly an undeniable proof that the
necessary prerequisites to deepen the process already exist (European Commission, 2022).
The results of our analysis show that the intention to adopt AI-containing applications is
directly supported by constructs embedding challenges, such as: perceived usefulness
(0.311), perceived hedonic value (0.224), attitude towards AI applications (0.202),
performance expectancy (0.149) and compatibility (0.146). In other words, students are
open to these new technologies only if they are convinced that it is appropriate for their
own needs and increase their satisfaction with the act of learning, a perspective also shared
by Malik et al., 2021; Alam et al., 2023; Almaraz-López, Almaraz-Menéndez and López-
Esteban, 2023. Considering the total effects, the intention to adopt AI is influenced by
interactivity (0.341), perceived usefulness (0.311), perceived hedonic value (0.224), attitude
towards AI applications (0.202), immersion (0.150), performance expectancy (0.149),
compatibility (0.146), and personalisation (0.146). Consequently, the holistic perspective
shows us that all the attributes inserted in the analysis are relevant in shaping students'
intention to adopt AI. Being familiar with various social networks, students combine the
efficiency and usefulness of AI applications with their personal needs, becoming sensitive
to the ability of AI to test their real-life skills in a virtual space placed on the same
wavelength as their own needs. Also, our analyses reveal that although interactivity does
not directly affect the intention to adopt AI applications, it is nevertheless the factor with
the highest influence, determining directly and significantly perceived hedonic value
(0.625), perceived usefulness (0.509), and immersion (0.289), an aspect also confirmed by
the studies of Pillai et al. (2023) or Dubey and Sahu (2021).
The main empirical findings of the study carried out in the current research emphasise the
importance of incorporating AI applications in teaching activities to encourage students to
use them in their learning process. To achieve this, AI applications should possess both
interactivity and facilitation features. In this way, AI applications can directly contribute to
enhancing learning capabilities, saving time, and obtaining information quickly. Professors
play a crucial role in the teaching-learning process by implementing interactive strategies
that enable students to use AI applications to acquire new skills and knowledge, leading to
better learning outcomes. Furthermore, AI applications can adapt to student learning needs,
Vol. 26 • No. 65 • February 2024 65

while providing useful information to understand various notions and to assimilate

information at their own pace.
The answers provided by our respondents show that the new generation of young people
needs to validate the usefulness of this technology. Even more, the specific benefits they
could get from AI could materialise into an improved ability to learn, saving time in the
learning process, facilitating the search for information, allowing faster retrieval of
information, etc. Considering the pattern of the target group – undergraduate students –
here, interactivity is very important. The more interesting the AI applications, the easier it
will be for the students to adopt them. This feature highlights that students do not feel
comfortable with rigid applications that put pressure on them, but rather prefer AI
technologies that provide them freedom of choice and generate accurate results in terms of
knowledge accumulation. Teachers also have the opportunity to track students' engagement,
their level of immersion, and whether they feel useful performing some tasks in the learning
process using AI applications. This way students can be stimulated in the learning process
and can be checked to see whether the AI applications selected by teachers are compatible
with the students' learning style or not.
Conclusions
The findings of our research show that students are open to using AI applications that give
them satisfaction and are compatible with their own learning needs. The present study
applied to the Romanian university area, even if on a micro-scale (undergraduate students
from one faculty), validates the overall theoretical framework regarding the contribution of
AI technologies to the learning process. Its added value lies in providing milestones with
the status of challenges that need to be validated when encouraging the adoption of AI in
the act of learning by Romanian students. With a rather more traditional educational
culture, Romania needs this new impulse to bring tertiary education in line with Western
standards.
The usefulness of the present research derives from the fact that it highlights the need to
adopt a set of measures, in the Romanian university area, aimed at stimulating and
exploiting the benefits of AI to the advantage of students in the learning process, based on
the strong identified constructs. Currently, AI is present in the Romanian university
educational act, either in the form of plagiarism detection applications such as Turnitin or
Copyscape, which can be used by both teachers and students, or through tools that facilitate
correct expression in foreign languages such as Grammarly or DeepL Write. ChatGPT,
which provides students with complex information on various topics, in real time, or
Mendeley, which supports citation and the organisation of bibliographic references for
various projects, are other common examples. Thus, their widespread applicability still
requires time to adapt and embed into daily behaviour.
After processing the questionnaire, it was found that the respondents have indicated a
number of AI applications they are already using in the educational process, such as:
ChatGPT (100%); Bard or Microsoft Bing (62% of respondents); less than 24% of
respondents use Grammarly, My AI on Snapchat, Wolfram Alpha, Duolingo, Bloomai,
Merlin AI, Tutor AI, Anthropic, or Cohere.
AE
The paper also has some limitations arising from the fact that the data was collected from a
single university centre. To gain additional representativeness in the application of AI in
Romanian higher education, it would be beneficial to extend the application of the
Universitaria Consortium. Also, another element worth of consideration refers to the
insertion in the analysis of new constructs such as: AI support conditions (showing us
whether AI has the necessary support for implementation, or if it has been used before by
the students) and the perceived risk of adopting AI. On the one hand, students who already
know the usefulness of AI, trust it, and are open to discover more via new applications. For
students who have not used AI in the past, their expectations may go far beyond the actual
results of using these technologies. This provides valuable conclusions concerning the need
for a proper awareness-raising campaign on AI technologies among students, in order to
familiarise them with the real benefits of AI for stimulating and supporting learning, before
the actual implementation of the technologies. On the other hand, the construct of perceived
risk of AI adoption may implicitly signal beneficiaries' fears related to an AI-assisted
learning act. Student feedback can be relevant to the academic or higher education
management decision-making process regarding the choice of AI-based technologies or
applications.
As future research directions, we propose not only to insert the above-mentioned constructs
into the analysis, but inherently to extend the perspective concerning the intention to adopt
AI also to teachers who manage the educational process with the support of AI. This
perspective will provide an overall picture with respect to AI - higher education relationship
so that the associated challenges can be addressed for the benefit of all parties involved in
raising the performance of higher education.
References
Alam, S.S., Masukujjaman, M., Ahmad, M. and Jaffor, R., 2023. Acceptance of online
distance learning (ODL) among students: Mediating role of utilitarian and hedonic
value. Education and Information Technologies, 28(7), pp. 8503-8536.
https://doi.org/10.1007/s10639-022-11533-3.
Almaraz-López, C., Almaraz-Menéndez, F. and López-Esteban, C., 2023. Comparative
Study of the Attitudes and Perceptions of University Students in Business
Administration and Management and in Education toward Artificial Intelligence.
Education Sciences, 13(6), article no. 609. https://doi.org/10.3390/educsci13060609.
Asif, M. and Krogstie, J., 2013. The role of personalization in mobile service adoption.
Proceedings of the International Conference on Multimedia and Human Computer
Interaction, Canada [online] Available at: <https://www.researchgate.net/publication/
258847921_Role_of_Personalization_in_Mobile_Services_Adoption#fullTextFileConte
nt> [Accessed 2 September 2023].
Baykasoğlu, A., Özbel, B.K., Dudaklı, N., Subulan, K. and Şenol, M.E., 2018. Process
mining based approach to performance evaluation in computer‐ aided examinations.
Computer Applications in Engineering Education, 26(5), pp. 1841-1861.
https://doi.org/10.1002/cae.21971.
Binu, D. and Rajakumar, B.R., 2021. Artificial Intelligence in Data Mining: Theories and
Applications. Cambridge: Academic Press.
Vol. 26 • No. 65 • February 2024 67

Boucher, P., 2020. Artificial intelligence: how does it work, why does it matter, and what
can we do about it? [pdf] Panel for the Future of Science and Technology, European
Parliamentary Research Service. Available at: <https://www.europarl.europa.eu/
RegData/etudes/STUD/2020/641547/EPRS_STU(2020)641547_EN.pdf> [Accessed 31
August 2023].
Chatterjee, S. and Bhattacharjee, K.K., 2020. Adoption of artificial intelligence in higher
education: a quantitative analysis using structural equation modelling. Education and
Information Technologies, 25(5), pp. 3443-3463. https://doi.org/10.1007/s10639-020-
10159-7.
Chauhan, S. and Jaiswal, M., 2016. Determinants of acceptance of ERP software training in
business schools: Empirical investigation using UTAUT model. The International
Journal of Management Education, 14(3), pp. 248-262. https://doi.org/10.1016/
j.ijme.2016.05.005.
Chu, H.C., Hwang, G.H., Tu, Y.F. and Yang, K.H., 2022. Roles and research trends of
artificial intelligence in higher education: A systematic review of the top 50 most-cited
articles. Australasian Journal of Educational Technology, 38(3), pp. 22-42.
https://doi.org/10.14742/ajet.7526.
Dever, D.A., Azevedo, R., Cloude, E.B. and Wiedbusch, M., 2020. The Impact of
Autonomy and Types of Informational Text Presentations in Game-Based
Environments on Learning: Converging Multi-Channel Processes Data and Learning
Outcomes. International Journal of Artificial Intelligence in Education, 30(4),
pp. 581-615. https://doi.org/10.1007/s40593-020-00215-1.
Dubey, P. and Sahu, K.K., 2021. Students’ perceived benefits, adoption intention and
satisfaction to technology-enhanced learning: examining the relationships. Journal of
Research in Innovative Teaching & Learning, 14(3), pp. 310-328.
https://doi.org/10.1108/JRIT-01-2021-0008.
Escotet, M.Á., 2023. The optimistic future of Artificial Intelligence in higher
education. PROSPECTS. https://doi.org/10.1007/s11125-023-09642-z.
European Commission, 2021. Shaping europe’s digital future – coordinated plan on
artificial intelligence 2021 review. [pdf] Available at: <https://digital-
strategy.ec.europa.eu/en/library/coordinated-plan-artificial-intelligence-2021-review>
[Accessed 30 August 2023].
European Commission, 2022. AI watch index 2021. [online] Publications Office of the
European Union. Available at: <https://data.europa.eu/doi/10.2760/921564> [Accessed
30 August 2023].
Fenwick, A. and Molnar, G., 2022. The importance of humanizing AI: using a behavioral
lens to bridge the gaps between humans and machines. Discover Artificial Intelligence,
2(1), article no. 14. https://doi.org/10.1007/s44163-022-00030-8.
Gînguță, A., Ștefea, P., Noja, G.G. and Munteanu, V.P., 2023. Ethical Impacts, Risks and
Challenges of Artificial Intelligence Technologies in Business Consulting: A New
Modelling Approach Based on Structural Equations. Electronics, 12(6), article no.
1462. https://doi.org/10.3390/electronics12061462.
Hair, J.F., Hult, G.T.M., Ringle, C.M., Sarstedt, M., Danks, N.P. and Ray, S., 2021. Partial
least squares structural equation modeling (PLS-SEM) using R: a workbook. Classroom
Companion Business. Cham, Switzerland: Springer.
AE
Hannan, E. and Liu, S., 2023. AI: new source of competitiveness in higher
education. Competitiveness Review: An International Business Journal, 33(2),
pp. 265-279. https://doi.org/10.1108/CR-03-2021-0045.
Henseler, J., 2018. Partial least squares path modeling: Quo vadis? Quality & Quantity,
52(1), pp. 1-8. https://doi.org/10.1007/s11135-018-0689-6.
Henseler, J., Dijkstra, T.K., Sarstedt, M., Ringle, C.M., Diamantopoulos, A., Straub, D.W.,
Ketchen, D.J., Hair, J.F., Hult, G.T.M. and Calantone, R.J., 2014. Common Beliefs and
Reality About PLS: Comments on Rönkkö and Evermann (2013). Organizational
Research Methods, 17(2), pp. 182-209. https://doi.org/10.1177/1094428114526928.
Kaplan, D., 2001. Structural Equation Modeling. In: International Encyclopedia of the
Social & Behavioral Sciences, pp. 15215-15222. https://doi.org/10.1016/B0-08-043076-
7/00776-2.
Luan, H., Geczy, P., Lai, H., Gobert, J., Yang, S.J.H., Ogata, H., Baltes, J., Guerra, R., Li,
P. and Tsai, C.-C., 2020. Challenges and Future Directions of Big Data and Artificial
Intelligence in Education. Frontiers in Psychology, 11, article no. 580820.
https://doi.org/ 10.3389/fpsyg.2020.580820.
Maisha, K. and Shetu, S.N., 2023. Influencing factors of e-learning adoption amongst
students in a developing country: the post-pandemic scenario in Bangladesh. Future
Business Journal, 9(1), article no. 37. https://doi.org/10.1186/s43093-023-00214-3.
Makridakis, S., 2017. The forthcoming Artificial Intelligence (AI) revolution: Its impact on
society and firms. Futures, 90, pp. 46-60. https://doi.org/10.1016/j.futures.2017.03.006.
Malik, R., Shrama, A., Trivedi, S. and Mishra, R., 2021. Adoption of Chatbots for Learning
among University Students: Role of Perceived Convenience and Enhanced
Performance. International Journal of Emerging Technologies in Learning (iJET),
16(18), article no. 200. https://doi.org/10.3991/ijet.v16i18.24315.
Nakagawa, S., Johnson, P.C.D. and Schielzeth, H., 2017. The coefficient of determination
R 2 and intra-class correlation coefficient from generalized linear mixed-effects models
revisited and expanded. Journal of The Royal Society Interface, 14(134), article no.
20170213. https://doi.org/10.1098/rsif.2017.0213.
Pan, Y., 2016. Heading toward Artificial Intelligence 2.0. Engineering, 2(4), pp. 409-413.
https://doi.org/10.1016/J.ENG.2016.04.018.
Pellas, N., Mystakidis, S. and Kazanidis, I., 2021. Immersive Virtual Reality in K-12 and
Higher Education: A systematic review of the last decade scientific literature. Virtual
Reality, 25(3), pp. 835-861. https://doi.org/10.1007/s10055-020-00489-9.
Pillai, R., Sivathanu, B., Metri, B. and Kaushik, N., 2023. Students’ adoption of AI-based
teacher-bots (T-bots) for learning in higher education. Information Technology &
People. https://doi.org/10.1108/ITP-02-2021-0152.
Rodway, P. and Schepman, A., 2023. The impact of adopting AI educational technologies
on projected course satisfaction in university students. Computers and Education:
Artificial Intelligence, 5, article no. 100150. https://doi.org/10.1016/j.caeai.2023.
100150.
Rogers, E.M., 1962. Diffusion of innovations (1st ed.). New York: Free Press of Glencoe.
Vol. 26 • No. 65 • February 2024 69

Rönkkö, M. and Cho, E., 2022. An Updated Guideline for Assessing Discriminant Validity.
Organizational Research Methods, 25(1), pp. 6-14. https://doi.org/10.1177/
1094428120968614.
Saghiri, A.M., Vahidipour, S.M., Jabbarpour, M.R., Sookhak, M. and Forestiero, A., 2022.
A Survey of Artificial Intelligence Challenges: Analyzing the Definitions,
Relationships, and Evolutions. Applied Sciences, 12(8), article no. 4054.
https://doi.org/10.3390/ app12084054.
Shi, Z.-Z. and Zheng, N.-N., 2006. Progress and Challenge of Artificial
Intelligence. Journal of Computer Science and Technology, 21(5), pp. 810-822.
https://doi.org/10.1007/s11390-006-0810-5.
Streukens, S. and Leroi-Werelds, S., 2016. Bootstrapping and PLS-SEM: A step-by-step
guide to get more out of your bootstrap results. European Management Journal, 34(6),
pp. 618-632. https://doi.org/10.1016/j.emj.2016.06.003.
Taber, K.S., 2018. The Use of Cronbach’s Alpha When Developing and Reporting
Research Instruments in Science Education. Research in Science Education, 48(6),
pp. 1273-1296. https://doi.org/10.1007/s11165-016-9602-2.
Varshney, K.R., 2016. Engineering safety in machine learning. In: 2016 Information
Theory and Applications Workshop (ITA). La Jolla, CA, USA: IEEE. pp. 1-5.
https://doi.org/10.1109/ITA.2016.7888195.
Venkatesh, V., Morris, M.G., Davis, G.B. and Davis, F.D., 2003. User Acceptance of
Information Technology: Toward a Unified View. MIS Quarterly, 27(3), article no. 425.
https://doi.org/10.2307/30036540.
Verdú, E., Regueras, L.M., Gal, E., De Castro, J.P., Verdú, M.J. and Kohen-Vacs, D., 2017.
Integration of an intelligent tutoring system in a course of computer network design.
Educational Technology Research and Development, 65(3), pp. 653-677.
https://doi.org/10.1007/s11423-016-9503-0.
Challenges for Competency-Based Education in the Context of Development
AE
EXPLORING UNIVERSITY STUDENTS’ PERCEPTIONS

OF GENERATIVE ARTIFICIAL INTELLIGENCE IN EDUCATION
Mădălina Lavinia Țală1 , Cătălina Nicoleta Müller2 ,
Irina Albăstroiu Năstase3 , Olimpia State4 and Georgică Gheorghe5
1)3)4)5)
2)
Hochschule Fresenius, Berlin, Germany

Țală, M.L., Müller, C.N., Albăstroiu Năstase, I., State, Received: 30 September 2023
O. and Gheorghe, G., 2024. Exploring University Revised: 9 November 2023
Students' Perceptions of Generative Artificial Accepted: 12 December 2023
Intelligence in Education. Amfiteatru Economic,
26(65), pp. 71-88.
Abstract
Artificial intelligence, the latest chapter of the technological revolution, has a tremendous
potential to change every area of our lives. This article has focused on a specific form of
artificial intelligence, namely generative intelligence, which facilitates the generation of
content in all its forms (text, image, video, audio, programming codes, etc.). Thus,
generative artificial intelligence has a crucial role in education, allowing for the
personalisation of educational content and facilitating the learning process.
In the beginning, the paper has highlighted conceptual delimitations regarding artificial
intelligence and its applications in education, along with advantages and limitations,
highlighting that the adoption of generative artificial intelligence solutions, such as
ChatGPT, in higher education in economics has been relatively underexplored in the
literature. In order to cover these gaps identified in the literature, have been presented, in
the second part of the paper, the methodology and results of an exploratory research,
conducted on a sample of 364 undergraduate and master's students at the Faculty of
Business and Tourism within the Bucharest University of Economic Studies. The research
has provided insight into the perception of Business Administration students regarding
these applications. The results indicated a high level of awareness and interest in content
generation models and highlighted that users with favourable perceptions regarding the
quality of content generated by such applications tend to believe that their integration into
academic endeavours can foster creativity and enhance employment prospects.
Keywords: generative artificial intelligence, ChatGPT, perceptions, competencies, higher
education in economics.
JEL Classification: M19, O33.
 Corresponding author, Irina Albăstroiu Năstase – e-mail: [email protected]

Attribution License, which permits unrestricted use, distribution, and reproduction in any
medium, provided the original work is properly cited. © 2024 The Author(s).
Vol. 26 • No. 65 • February 2024 71

AE Exploring University Students’ Perceptions
of Generative Artificial Intelligence in Education
Introduction
Generative artificial intelligence (GAI) has become an extremely common concept
nowadays. It is based on artificial intelligence (AI) and implies the ability of a machine to
learn from experience, adapt to new information inputs and perform tasks like humans
(Duan, Edwards and Dwivedi, 2019). AI is a multidisciplinary amalgam, based on
computer science and logic, designed to solve easy and restrictive tasks (Tîrnăcop, 2023).
Although it has 70 years of existence behind it, only the last twenty years are considered
exceptional, during which AI frequently exceeds various milestones of knowledge (Zang
and Lu, 2021).
Generative AI is spinning off from AI, its use generating new content such as text, video,
image, software, product design, etc. Generative AI spreads massively in 2016 through
WaveNet, continues with GPT-2 (2019) and GPT 3 (2020), and becomes extremely popular
at the end of 2022 through ChatGPT (Sætra, 2023). GPT 4 appears also shortly after (Maar,
2023). We can only conclude that we are talking about a recent phenomenon, and moreover
one that propagates rapidly in professional and private life, that is estimated to have an impact
similar to that of the steam engine, electricity or internet in their time (Gartner, 2023).
As in other areas, education systems are also somehow shaken by the emergence and use of
generative AI - not only by students, but also by teachers and researchers. Being at the
beginning, many questions arise about the limits of using these technologies, from issues
related to the correctness of the generated content, to ethical aspects, such as appropriating
materials and presenting them as personal creations. On the other hand, we should
recognise that society is already impregnated with AI, and in order to be competitive, it is
necessary to acquire knowledge and skills to understand and use it (Chiu et al., 2022).
Thus, generative AI in education must be seen as an opportunity rather than a threat
(Duckworth and Ungar, 2023).
A search of Web of Science materials on AI and education between 1975 and 2023
returned more than 57,400 titles, 80% of which have been written in the last 10 years
(Clarivate – Web of Science, 2023), with a noticeable year-on-year increase in interest.
Chen, L., Chen, P. and Lin (2020) analyse a number of articles from 2010 to 2019 and
conclude that the use of similar platforms and tools by teachers has improved their
efficiency and effectiveness, leading to a more qualitative training process. Students also
benefited from materials tailored to their needs and capabilities.
Despite being less than a year old, ChatGPT is already quite popular among authors.
Designed to engage in conversational interactions with users, providing extremely fast
responses, it can synthesise information, suggest structure and article titles, generate
literature lists among many other things (Salvagno, Taccone and Gerli, 2023). Sohail et al.
(2023) analysed 109 articles published in Scopus related to ChatGPT and showed that 27%
of the articles relate to the ability to write scientific material. Other authors have been
concerned about ChatGPT's relationship with teachers (Lim et al., 2023; Naumova, 2023,)
or with students (Gaševic, Siemens and Sadiq, 2023). An interesting study comes from
Greitemeyer and Kastenmüller (2023) who researched the relationship between personality
traits and intent to defraud at the academic level.
We appreciate that an important moment is the extremely recent appearance of the first
Guide for generative AI in Education and Research, under the direction of UNESCO
AE
(UNESCO, 2023). Its purpose is to ensure that the use of generative AI helps teachers,
learners and researchers to achieve better results, in a shorter time, without "usurping"
human intelligence. The guide is a useful tool that validates generative AI's role in present,
but especially in future education, highlighting the need for a new social contract for
education, in which we need to redefine the relationship with technology.
The purpose of this paper, through its research, is to analyse how the use of generative AI
is perceived by students, in the context of the newness of this technology. Thus, it aims to
identify the connection between the use of AI and generative AI, make a radiography of the
use of AI tools for generating text, such as ChatGPT, and the perception that students have
about these tools. Also, we place in the light how respondents position themselves in
relation to the status of learner, and that of future employee: better employability
perspectives, creativity, and ethics. Given the very short time elapsed since the advent of
ChatGPT and the completion of questionnaires, we appreciate that we bring to the reader's
attention a very useful work, with results that will be referential in the future. Equally, the
article shows students' preferences in relation to how skills related to knowledge and use of
generative AI can be formed, preferences that can be taken into account by teachers and
university management when developing the content of disciplines and curricula.
The paper is organised into three main sections. The first section delves into the existing
literature regarding the presence, benefits, challenges, and constraints of AI in education,
as well as the research on the topic of students' perceptions of generative AI. The second
section provides an overview of the research methodology, outlining the steps undertaken
and leading up to the data analysis phase. The final and most extensive section showcases
the research findings and, where applicable, establishes connections with prior studies.
Alongside the introductory section, the paper also includes a conclusion section that
summarises the research's significance, its limitations, and potential avenues for future
research.

1.1. The use of artificial intelligence systems in education
Educational institutions will need to continuously adapt their curriculum due to the rapid
development of AI. Generative AI has the potential to change education in several ways.
Because AI development is interdisciplinary, a rigorous curriculum that combines technical
(programming, data science) and non-technical (critical thinking, ethics) skills is required.
Moreover, Baker, Smith and Anissa (2019) discuss in their report three potential
approaches to using educational AI tools: student-centeredness, teacher-centeredness, and
the AI in education system (AIEd). In this context, personal tutors, intelligent support for
collaborative learning, and intelligent virtual reality are three categories of AI software
products for education that are already available, according to Luckin et al. (2016).
UNESCO AI for Education (UNESCO, 2020) (a summary of the current state, challenges,
and potential applications of AI in education), OECD Learning Compass 2030 (OECD,
2023) (a paradigm for reinventing education in the age of AI), EdTech Hub (2023) (a forum
for cooperation, research and innovation in education technology), the Artificial
Intelligence (AI) for K-12 Initiative program (AI4K12, 2023) (recommendations and tools
for incorporating AI into K-12 curricula) are all excellent sources of information on AI in
education (Technological Innovation, 2023).
Vol. 26 • No. 65 • February 2024 73

The introduction and use of AI in higher education has created new opportunities and
challenges. The challenges, extensively discussed in the UNESCO Guide to generative AI
in Education and Research (UNESCO, 2023), involve not only training instructors for AI,
but also training AI to understand education. The guide also includes a new curriculum for
the digital age.
According to Park and Kwon (2023), students in the AI era will interact with technology in
fundamentally different ways than the previous generation. To prepare them to live in this
future, the emphasis is on educating them using a constructivist learning approach, as well
as design and creative thinking (Ali et. al., 2019). Additionally, students should be required
to take AI literacy courses in K12 education. (Park and Kwon, 2023),
On the other spectrum, it is hopeful to see that teachers are quite interested in AI and are
aware of the need for AI education (Park and Kwon, 2023). Essays, translations, and
creative writing are already examples of AI-generated texts for use in education (Bailey,
2023).
1.2. The advantages, opportunities, but also the challenges and limits of AI
Development of AI technologies brings a plethora of opportunities to competency-based
education. Competency-based education is an approach to education that emphasises the
development of skills rather than just the teaching of knowledge. Aspects such as critical
thinking, problem solving and collaboration skills are pursued (Sanusi et al., 2022).
For example, depending on the impairments, technology can provide appropriate resources
for specific needs; AI systems can simultaneously provide students with immediate
feedback, helping them understand their errors and directing them to the right solution;
collecting and evaluating performance data can help teachers improve their lesson plans
and curricula (Greene-Harper, 2022). Also, AI can automate administrative tasks such as
scheduling and grading, giving teachers more time to focus on teaching (The Knowledge
Review, 2023). However, automated grading using AI-generated text patterns has the
potential to negatively affect students' final grades and future career prospects (Akgun and
Greenhow, 2021).
The use of AI-generated writing in academic settings has raised questions about the
likelihood of cheating and the ethical ramifications (Bailey, 2023), as not all educational
institutions have the resources, such as state-of-the-art computers and specialised software, to
teach AI effectively. Furthermore, students with different backgrounds and skill levels require
customised approaches to AI education, making it difficult to provide a one-size-fits-all
curriculum. Competency-based education requires innovative assessment methods that assess
practical skills, and these can be difficult to design and implement. Also, there may be a gap
between what industry needs and what education provides in terms of AI skills. Add to this a
potential bias and questionable fairness in AI systems, the risk of losing jobs to AI, and a
decrease in human connection, a crucial component of learning, and we can have a wider
picture of the challenges related to AI (Greene-Harper, 2023; Vallis et al., 2023).
Research findings by Rajabi et al. (2023) on student perceptions of using ChatGPT
highlights the need for ChatGPT users to be aware of the tool's limitations, such as the
possibility of incorrect or biased responses. They also emphasise the need to double-check
chatbot responses and the need for a balance between classwork and homework to prevent
potential abuse of ChatGPT, while maintaining a positive learning environment.
AE
Recognising that AI is a tool to enhance learning and not a replacement for human teachers,
it is essential to strike a balance between the advantages and disadvantages of AI in the
classroom. For example, ChatGPT can be a useful tool to enhance learning, but it is
important to keep in mind its limitations and the value of human connection in the
classroom.
1.3. Adoption of generative AI systems in higher education and students’ perceptions
of them
Students’ perceptions of generative AI, a topic of interest in education, have been the
subject of several studies already. One research brings forth the comparison of the attitudes
toward AI shown by students training for a career in education, respectively business
administration and management, and highlights that 82.51% of those studying economics
and business management and 85.82% of those studying education had favourable attitudes
towards AI (Almaraz-López, Almaraz-Menéndez and López-Esteban, 2023). Another study
recommended the inclusion of generative AI in entrepreneurship curricula, so that students
could investigate the possibilities to use it to develop new goods and services (Bell, R. and
Bell, H., 2023).
In another study carried out by Ibrahim et al.'s (2023) on the perceptions and performance
of text-based conversational AI in 32 universities, it was found that student views of this
new technology were mixed, with some expressing positive views and others expressing
negative views. Based on this, we can infer that not all learners will be interested in AI -
some may be really worried and nervous, while others are interested, and others just don't
care. As a result, teachers must take into account a variety of points of view when
implementing AI in the classroom (Guzman, 2023). University students' impressions of
generative AI technologies such as ChatGPT were examined in another study at the
University of Hong Kong (Chan and Hu, 2023). The research found that students saw
advantages such as individualised learning, improved engagement, and increased creativity,
and also noticed certain difficulties, such as lack of control and ethical issues.
Other studies (Grassini, 2023; Shaji, Hovan and Gabrio, 2023) on the impact of ChatGPT
on student learning and preparation for the future workforce have produced conflicting
findings. ChatGPT could improve learning by expanding on already known information
and facilitating research, but relying solely on ChatGPT, on the other hand, can inhibit
critical thinking, practice, and communication with instructors, which could delay learning
development and have a negative effect on work readiness.
Shoufan (2023) asked senior computer engineering students to evaluate ChatGPT and
revealed that they find the tool fascinating, inspiring, and useful for both study and works;
the students were also enjoying how easy it was to use and the fact that it provided well-
organised answers and clear justifications, as if they were given by a human. However,
many students believed that ChatGPT answers were not always correct, and most
appreciated that having a solid knowledge is necessary for employment, as ChatGPT does
not replace human intelligence. The students believed that ChatGPT needs improvement
and expressed their hope that this will happen soon.
According to research carried out by Sánchez-Ruiz et al. (2023), the students’ opinions
about the harmful effects of ChatGPT on education, academic integrity, employment and
other aspects of life were divided. The students showed they were able to quickly adapt to
use ChatGPT, demonstrating strong confidence in his responses and overall use in the
Vol. 26 • No. 65 • February 2024 75

learning process, along with a positive evaluation. However, questions have been raised
about the potential impact on the ability of future engineers to develop critical lateral skills.
Additionally, Iuga (2023) carried out an interesting SWOT analysis for Romania in relation
to generative AI, highlighting the prospects for Romanian enterprises to take advantage of
this technology to improve operational procedures and build new capabilities. While this is
not specific research that directly examines the perceptions of economics students
regarding generative artificial intelligence in education in Romania, this study can offer
valuable insights and a relevant context for exploring related subjects.
As indicated throughout this article, aspects related to the use of artificial intelligence
systems in education, as well as the advantages, benefits, challenges, and limitations of
these solutions, have been addressed in the specialised literature. However, it is important
to note that there is a lack of research on the degree of usage and on perceptions of aspiring
economists regarding these issues, especially since we have not identified other studies
targeting students and master's students from Romania, neither in the field of economics, in
general, nor in the field of Business Administration studies. To address these gaps in the
specialised literature, this article aims to present the results of a broader research conducted
among undergraduate and master's students at the Faculty of Business and Tourism within
the Bucharest University of Economic Studies.
The research aimed to determine the perceptions of generative AI models among
undergraduate and graduate students. The primary objectives of our research were as
follows:
• O1: Identify the impact of digital skill levels on the adoption of generative AI models
(for text, images, videos, etc.);
• O2: Determine the connection between respondents' gender and the use of generative
AI;
• O3: Examine the relationship between using generative AI models in general and using
generative AI for text for academic purposes;
• O4: Explore the correlation between perceptions of the quality of AI-generated texts
and those related to creativity and employment prospects in the context of AI.
Related to these objectives, we have formulated the following hypotheses:
 H1: The acquired digital competences favour the adoption of generative AI;
 H2: Gender influences the extent to which generative AI is used;
 H3: Users of generative AI models rely on AI solutions that generate text (such as Chat
GPT) for academic purposes;
 H4: Users who evaluate positively the quality of AI-generated texts believe that using
AI for academic purposes improves creativity and employment prospects.
This research was based on a standard methodology that is applied in the questionnaire-
based surveys. The choice of such a method is justified by the fact that it is commonly used
AE
in social sciences to investigate user/ consumer behaviour and test related theories. (Hinkle,
Wiersma and Jurs, 2005; Groves et al., 2009).
The method chosen for data collection was the self-administered questionnaire, which was
developed on the Google Forms platform and was distributed as a link by e-mail to the
students of the Faculty of Business and Tourism (undergraduate and graduate) and through
the forums and groups dedicated to them. The questionnaire consisted of both open and
closed questions, multichotomous and with semantic scale. The opening section of the
questionnaire briefly explained its purpose to ensure a common frame of reference for all
potential respondents. We note that this questionnaire underwent testing prior to its
finalisation and distribution to the research community.
The sample used in the survey was 364 persons. We specify that initially there were 371
respondents, but 7 questionnaires were not validated, being subsequently excluded. Given
the limited sample size, this online questionnaire-based survey is an exploratory research,
which is used to discover relationships, interpretations, and characteristics of subjects that
suggest new theories and define new problems (Swanson and Holton, 2005). Data
collection was carried out between May and June 2023. We used the IBM SPSS Statistics
version 26 for descriptive analysis and for hypothesis testing.

In order to assess how students perceive generative AI in the context of education, the
questionnaire focused on the following areas: awareness of generative AI models,
frequency of use of AI-generated text models, quality of AI-generated text, impact on
grades, creativity & employability & ethical implications. Moreover, one of the questions
targeted the interest students have in pursuing formal education on AI tools, which can
inform the university management in its future endeavours.
The starting question aimed to understand the context in which students find themselves
with respect to generative AI – have they heard about such models, have they used them,
and how interested they are in this topic (Table no. 1).
Table no. 1. The degree of use of generative AI models
Answer Frequency Percent
Yes, I have heard about them, but haven’t used any. I’m interested to try 83 22.8%
them.
Yes, I have heard about them, but I don't want to use them. 51 14.0%
Yes, I have used AI models for generating content a few times. 191 52.5%
Yes, I use them regularly. 20 5.5%
No, I have never heard about AI models for generating content. 19 5.2%
Total 364 100%
It is interesting to notice that a significant portion of the university students surveyed have
some level of awareness of AI models for generating content - 95% heard of them and have
a strong opinion on using them or not using them. The majority (52.5%) have used AI
models for generating content at least a few times. Additionally, a notable percentage
(22.8%) have heard about these models and express interest in trying them. However, it is
important to note that there is also a segment of 14% who have heard about these models
but do not want to use them, possibly due to concerns or other reasons. Additionally, a
Vol. 26 • No. 65 • February 2024 77

smaller percentage (5.5%) use AI-generated content regularly, while a similar percentage
(5.2%) have never heard about AI models for generating content. In comparison, in the
study conducted by Byles, Lea, and Howe in May 2023 at the University of Northampton,
62% of the respondents reported not using AI models, while in the study by Chan and Hu,
published in July 2023, 33% of the respondents stated that they have not used AI models
such as ChatGPT. Based on this trend over time, we can infer an increased rate of adoption
of these tools.
When giving details about the reasons behind their choice (use or not use), the students
mention diverse motives that range from “curiosity” to “incorrect information”. Let us first
focus on the 90% of the students that are interested in using generative AI models:
 33% express a curiosity-driven interest that suggests that they are intrigued by the
technology and explored or will soon explore its capabilities.
 A similar percentage (32%) consider using AI-generated content models for greater
efficiency. They comment that the tools represent ways to save time and effort in content
creation and learning.
 Around 14% of students see AI-generated content as a source of inspiration, which
highlights the potential of these tools to spark creative ideas and foster innovation within
the academic context. Moreover, this suggests that students don’t plan to use the content as
it is, but they will bring their own ideas.
 A troubling percentage of 10% of students use AI-generated content models to find
information (e.g., they mentioned “some statistical data”, “restaurants in Bucharest” or
“historical and geographical facts about different countries”) and, moreover, consider
reliable the content generated. Students say that they used AI when the exact information
was not available on Google or when they wanted to generate exact information. This could
indicate a huge potential risk of misinformation and mistrust. Moreover, we may witness a
growing dependence on AI-generated content, which may have both positive and negative
consequences. On the positive side, it can fill knowledge gaps when traditional sources are
insufficient, while on the negative side we will see the rising risk of spreading false
information and of over-reliance, potentially leading to a lack of critical thinking and
verification of information.
In the beginning of the study 10% of the students declared that they are not interested in
using generative AI. Their reasons vary between expecting the content generated by AI to
be incorrect, forecasting negative effects on their cognitive abilities, preference for using
their own mind, seeing it as a case of plagiarism or not perceiving any added value.
Speaking of all generative AI models in commercial use, the one that stands out is
ChatGPT - it is mentioned by 76% of the students that have used generative-AI and by 53%
of those who have limited experience or no experience with AI tools, but are interested.
This finding is concurrent with that of the study carried out in Germany in 2023 on 6.311
students from 395 colleges (Garrel, Mayer and Mühlfeld, 2023). When asked how they use
AI tools in their studies and for what purposes, 63.2% of the students stated that they had
already used or are currently using AI-based tools for their studies and 48.9% of them
already used ChatGPT. The higher percentage of familiarity we have noticed in our study
can be explained by the time passed in between the two studies and the positive results the
people experience that further catered to widespread discussions and use of the tools.
AE
When looking at the frequency of use for generative-AI text models, most of the students
declared they use them for academic purposes “rarely” (42%) or “sometimes” (37%) or
expect to use them “rarely” (37%) or “sometimes” (46%), which suggest that students
believe they will continue to rely on the traditional methods for academic content creation.
When looking at students’ perception about the quality of AI-generated text, it is interesting
to notice that more of those with little or no experience expect the quality to be lower than
their own written text – 30%. From those who speak from experience on using AI-
generated text, only 15% perceive the quality to be lower than their own writing. This
suggests that lack of exposure to AI-generated content might lead to a more skeptical or
cautious outlook, whereas a hands-on experience with AI-generated text can positively
influence perceptions of quality, possibly because students have seen the potential benefits
and improvements in their work. This finding is consistent with that of the study carried out
between 2018 and 2019 in Turkey, when AI was not yet that prominent on the firmament of
breakthrough technologies (Keleş and Suleyman, 2021). The test used showed that, at that
point in time, negative perceptions of all sample groups about artificial intelligence were
more significant than positive perceptions.
Overall, the content generated with the help of AI is expected to be incorrect or irrelevant
“often” by approximatively 13% of the respondents that used or are interested in using the
tools, and “sometimes” by 44%. This translates into the fact that too many of the surveyed
students (43%) have a relatively positive view of AI-generated content: they seem to
believe that AI-generated content is rarely or never incorrect or irrelevant. This finding
should sound the alarm, as the high level of trust students seem to have and the perception
as a reliable source of content is not supported by how the models work. AI-generated text
models are famous for their “hallucinations” or “confabulations” (if we don’t want to
anthropomorphise AI). Natively, there is nothing in a GPT (generative pre-trained
transformers) model's raw data set that separates fact from fiction (Edwards, 2023). If used
as a brainstorming tool, its logical leaps and confabulations might lead to creative
breakthroughs, but when used as a factual reference, the model could cause real harm.
On this note in the end of the questionnaire we aimed to assess the interest of students in
being better prepared for using AI (Table no. 2).
Table no. 2. Students’ interest in learning about generative AI
and using AI tools in a formal setting
Answer Frequency Percent of N = 364
I am interested in taking a stand alone course on AI tools 105 29%
(applications, how to use them safely and ethically). (A)
I am interested in using AI tools in different disciplines and 172 47%
finding out how they can be applied. (B)
I am interested in attending extracurricular events that focus on 94 26%
AI tools. (C)
I am not interested in learning or using any AI models. 77 21%
Other. 3 1%
Total 451
78% of the respondents are interested to learn more about generative AI tools and their
applications in a formal context, a percentage similar to that recorded by Bistas et al.
(2021). The highest percentage, 47%, indicated an interest in using AI tools across various
Vol. 26 • No. 65 • February 2024 79

disciplines and exploring their potential applications; within them more than half chose
only this option, while the rest opted for a combination of options – A+B+C, A+B or A+C.
To our surprise, a relatively high percentage (21% of all respondents) stated that they are
not interested in learning about or using AI models, which means that the rows of those that
declared themselves uninterested in this technology from the very beginning were further
thickened by some of those that used generative AI and re-assessed their future interest in
the technology. Their reasons lie in their belief that they don’t need to or that using such
models will have a negative impact on their cognitive abilities. This finding further
underscores the importance of providing a balanced and informed education about
generative AI. Addressing misconceptions and highlighting the potential benefits of AI in
various contexts can help bridge the gap and engage a broader spectrum of students in AI-
related learning opportunities.
This particular finding is supported by the fact that 26% of the respondents interested in AI
models also state that “is unethical to use AI-generated text in academic assignments”. This
group may be concerned with issues related to plagiarism, originality, and the integrity of
their work. Their stance reflects a cautious approach to the use of AI in academic writing.
On the other spectrum, a minority of students (7%) believe that AI should be used without
any restrictions. This group is likely more permissive when it comes to the use of AI in
academic assignments and may not see the need for stringent ethical guidelines. It is also
worth noting that many of the students surveyed (52%) declare that AI can be used in
academic settings in certain conditions: proper citation and significant modifications of the
text (30%); proper citation (15%); significant modifications (7%). These results indicate an
understanding of the importance of giving credit to AI-generated content, aligning with
academic integrity standards, while also valuing the transformative role of human input in
the content creation process.
To achieve our research objectives, we used SPSS v. 26 and conducted a dependency test
to determine whether there is a relationship between the variables analysed.
For the first objective (O1), we considered the independent variable “acquired digital
competences” and the dependent variable “use of AI-generated content models”. We
applied the Chi-square test and calculated the Cramer’s V value to determine the strength of
the relationship, if any. The results show that there is a relationship between the two
variables, which is confirmed by the value of 0.000 for asymptotic significance (Table no.
3). It is a direct correlation, which means that an increase in the acquired digital
competences leads to an increase in the use of generative AI. We can say that the
relationship has a medium intensity, as evidenced by the Cramer’s V value of 0.188 (Kim,
2017). The hypothesis was tested in 364 cases. Based on the results, we can say that
hypothesis H1 – “The acquired digital competences favour the adoption of generative AI
for content” is confirmed.
Table no. 3. Correlation between general digital skills
and use of generative AI models for content
Asymptotic
Elements Value df = min (r – 1, c – 1)
Significance
Chi-Square (χ2) 38.590 0.000 3
Cramer’s V 0.188 0.000 -
N of Valid Cases 364 - -
AE
Regarding the second objective (O2), another topic of interest was whether gender plays a
role in the use of content-generating AI models. According to other research (Siregar,
Hasmayni and Lubis, 2023), gender plays a role in the use of AI for content generation, in
the sense that men tend to use more AI models. At the same time, other studies
(Niethammer, 2020) have emphasised that only 22% of individuals working in the AI
industry are female.
In our study out of the 364 respondents, 61.8% were female, 36.8% were male, 0.8% were
non-binary and 0.5% chose not to respond. Based on the results of testing the strength of
the association between “gender” and “use of generative AI for text”, which yielded a value
of asymptotic significance greater than 0.05, we can conclude that hypothesis H2 –
“Gender influences the extent to which generative AI is used” is refuted. According to our
study and contrary to other findings, gender does not influence the use of generative AI for
text (Table no 4).
Table no. 4. Correlation between gender and usage of generative AI models for text
Asymptotic
Elements Value df = min (r – 1, c – 1)
Significance
Chi-Square (χ2) 10.023 0.614 3
Cramer’s V 0.096 0.614 -
For the third objective (O3), we ran a linear regression to find for a relationship between
the use of generative AI (independent variable) and the use of AI solutions that generate
text for academic purposes (dependent variable). Looking at the significant value of our
model of 0.000, which is less than 0.05, we can confirm that our model is statistically
significant (Table no. 5). The significance of the model can be stated as follows:
F (1.209) = 20698, p = .000 (1)
In the model summary, the percentage of variance is observed, which in our case is 8.6%.
The value can be interpreted as fallow: 8.6% of those who use AI solutions use them to
generate text for academic purposes. The result confirms hypothesis H3 formulated in the
research methodology – “Users of generative AI models rely on AI solutions that generate
text (such as Chat GPT) in education”, and this, according to Yilmaz and Karaoglan
Yilmaz (2023), leads to a significant improvement in users' skills.
Table no. 5. Correlation between the usage of generative AI and the usage
of AI solutions that generate text for academic purposes
Elements F Sig. df
ANOVA 20.698 0.000 1
The fourth objective of the research (O4) was to investigate the correlation between
perceptions of the quality of AI-generated texts and those related to creativity and
employment prospects in the context of AI. In order to validate our hypothesis, that “users
who evaluate positively the quality of AI-generated texts believe that using AI for academic
purposes improves creativity and employment prospects”, we needed to validate both tests
simultaneously.
Vol. 26 • No. 65 • February 2024 81

The results of the first test on the relationship between the positive perception of the quality
of AI-generated text and the impact on creativity show a positive correlation of medium
strength between the variables (Table no. 6). For the analysis, we considered 206 responses
(those who positively evaluated the quality of AI-generated texts). Based on test results, we
can estimate that one in five respondents believe that AI has a positive impact on their
creativity. Similar conclusions are obtained by Marrone, Taddeo and Hill (2022), as well as
by Tapalova and Zhiyenbayeva (2022).
Table no. 6. Correlation between positive assessment of quality of AI-generated texts

and perception on creativity and employment prospects
Asymptotic
Elements for creativity Value df = min (r – 1, c – 1)
Significance
Chi-Square (χ2) - creativity 31.770 0.002 3
Cramer’s V - creativity 0.227 0.002 -
Chi-Square (χ2) – employment 23.703 0.022 3
prospects
Cramer’s V - employment 0.196 0.022 -
prospects
The same test was conducted to assess the relationship between the positive perception of
the quality of AI-generated text and the impact on employment prospects. The results
indicate that there is a positive influence of the medium intensity between these two
variables.
Based on the results of the two tests, we can confirm hypothesis H4 – “Users who evaluate
positively the quality of AI-generated texts believe that using AI for academic purposes
improves creativity and employment prospects”. There is a direct dependency between the
variables, which means that those who have a positive impression of AI are more likely to
believe that using AI for academic purposes improves their creativity, but also their
employment prospects. Significant differences between students who have rarely or never
used these technologies and other participants are reported in the study conducted by Chan
and Hu (2023).
Conclusions
Generative artificial intelligence can be a valuable resource for both students and
professors, enhancing the learning process and fostering competency-based education, in
the context of the digital revolution. Generative AI can create learning resources tailored to
each student's level and needs, providing assistance in academic writing, and encouraging
the development of creativity and critical thinking skills in the educational process. It is the
role of the professors to guide students in discovering the benefits of these solutions and
harnessing the opportunities they offer. However, the adoption of artificial intelligence
systems requires knowledge, competencies, and skills that the education system must
provide. Clearly, this gradual process involves the development of strategies, updating
curricula, introduction of new subjects, and the initial training of educators, among other
steps. Likewise, it is essential to carefully manage authenticity and ethics in the use of
AE
artificial intelligence in an academic context, in order to mitigate potential risks and

challenges associated with this continuously evolving technology.
In this context, we can state that the results of our research are valuable not only to
decision-makers in the field of education, in general (given the ongoing development of the
national strategy for artificial intelligence, which obviously includes this domain), but
especially to those in the field of higher education in economics in Romania, as well as the
key players within the academic community, namely students and professors. This is
because our study provides a comprehensive perspective on the perceptions and attitudes of
students and master's students in the economic field regarding the use of generative
artificial intelligence.
The results of our research indicate a high level of awareness among respondents regarding
AI models for generating generation, with the majority having heard of them, and a
significant number already using them. Additionally, respondents who have not yet utilised
these models, express interest in trying them, citing reasons such as curiosity, desire to
explore technology capabilities, and the potential for time and effort savings in content
creation and learning, aligning with the findings of other studies (Keleş and Aydın, 2021;
Baidoo-Anu and Owusu Ansah, 2023; Garrel, Mayer and Mühlfeld, 2023).
At the same time, it is essential not to lose sight of the respondents that are aware of
generative AI models but have no interest in utilising them, as well as those who are not
interested in learning about AI models in a formal context. Furthermore, some of those who
have used or intend to use AI models, perceive that the AI-generated content is rarely or
never incorrect or irrelevant, raising concerns about the trust placed in AI models. These
findings are not singular, with other studies also highlighting the existence of similar issues
(Ofosu-Ampong et al., 2023; Shoufan, 2023). However, these observations underscore the
need for education and awareness regarding AI and emphasise the importance of a balanced
adoption of generative artificial intelligence in education, requiring a careful approach to
address misconceptions and promote responsible AI usage in an educational context.
Regarding the correlations identified in our analysis, we can conclude that students with
more advanced digital skills are more inclined to use AI for content generation. In this
particular sample, gender did not emerge as a decisive factor in adopting AI-generated
content models, contradicting previous research findings (Siregar, Hasmayni and Lubis,
2023). Furthermore, our results highlight the interconnectedness of AI usage, as users of
generative AI models are also inclined to use AI solutions for text generation in academic
activities, demonstrating the synergy between AI applications. It is worth noting that users
who hold favourable perceptions of AI-generated text quality tend to believe that
integrating AI into academic endeavours can foster creativity and enhance employment
prospects, reinforcing the potential benefits of AI in educational contexts.
There were certain limitations during the preparation of this article, such as the lack of
studies focusing on the use of artificial intelligence in higher education in economics in our
country. This is the main reason for referencing studies carried out in other countries and
related to other fields of study. Clearly, our research also has limitations related to the
sample used, which cannot be statistically representative when considering higher
education in economics in Romania. Future research could be conducted among aspiring
economists from various universities in our country and across different specialisations,
with a focus on aspects and variables that our research addressed less (such as:
Vol. 26 • No. 65 • February 2024 83

investigating the ethical implications of AI integration in higher education, tracking the

long-term academic and career outcomes of students who have experienced/used AI to
determine if there are significant advantages or disadvantages, and exploring how AI-
driven tools and platforms can boost student engagement and motivation in economics
courses). Additionally, consideration could be given to comparing the perceptions of
students and master's students. Obviously, the perspective of teachers and trainers is of
interest and should be investigated, and, as AI systems are adopted and used in different
educational contexts, longitudinal research will become necessary to capture changes over
time in the perceptions of both learners and professors.
References
Akgun, S. and Greenhow, C., 2021. Artificial Intelligence in Education: Addressing Ethical
Challenges in K-12 Settings. AI and Ethics, 2(3), pp. 431-440. https://doi.org/
10.1007/s43681-021-00096-7.
Ali, S., Payne, B.H., Williams, R., Park, H.W. and Breazeal, C., 2019. Constructionism,
Ethics, and Creativity: Developing Primary and Middle School Artificial Intelligence
Education. In: International Workshop on Education in Artificial Intelligence K-12 –
EDUAI’19. Macao, China, 10-16 August 2019. Palo Alto: Mit Media Lab. pp. 1-4.
Almaraz-López, C., Almaraz-Menéndez, F. and López-Esteban, C., 2023. Comparative
Study of the Attitudes and Perceptions of University Students in Business
Administration and Management and in Education toward Artificial Intelligence.
Education Sciences, 13(6), article no. 609. https://doi.org/10.3390/educsci13060609.
Baidoo-Anu, D. and Owusu, A.L., 2023. Education in the Era of Generative Artificial
Intelligence (AI): Understanding the Potential Benefits of ChatGPT in Promoting
Teaching and Learning. Journal of AI, 7(1), pp. 52-62. http://dx.doi.org/10.2139/
ssrn.4337484.
Bailey, J., 2023. AI in Education. Education Next, 23(4), pp. 28-35. [online] Available at:
<https://www.educationnext.org/a-i-in-education-leap-into-new-era-machine-
intelligence-carries-risks-challenges-promises/> [Accessed 20 September 2023].
Baker, T., Smith, L. and Anissa, N., 2019. Educ-AI-tion Rebooted? Exploring the Future of
Artificial Intelligence in Schools and Colleges. London: Nesta. [online] Available at:
<https://media.nesta.org.uk/documents/Future_of_AI_and_education_v5_WEB.pdf>
Bell, R. and Bell, H., 2023. Entrepreneurship Education in the Era of Generative Artificial
Intelligence. Entrepreneurship Education, 6(3), pp. 1-16. https://doi.org/10.1007/
s41959-023-00099-x.
Bisdas, S., Topriceanu, C.-C., Zakrzewska, Z., Irimia, A.-V., Shakallis, L., Subhash, J.,
Casapu, M.-M., Leon-Rojas, J., Pinto dos Santos, D., Andrews, D.M., Zeicu, C.,
Bouhuwaish, A.M., Lestari, A.N., Abu-Ismail, L., Sadiq, A.S., Khamees, A.,
Mohammed, K.M.G., Williams, E., Omran, A.I., Abu Ismail, Y.D. and Ebrahim, E.H.,
2021. Artificial Intelligence in Medicine: A Multinational Multi-Center Survey on the
Medical and Dental Students' Perception. Frontiers in Public Health, 9, article no.
795284. https://doi.org/10.3389/fpubh.2021.795284.
AE
Byles, R., Lea, K. and Howe, R., 2023. Report: Exploring Student Perspectives on
Generative Artificial Intelligence Tools in Higher Education: A Survey-Based Study.
[online] Available at: <https://bpb-eu-w2.wpmucdn.com/mypad.northampton.ac.uk/
dist/0/18842/files/2023/10/AI-survey-report_9-Oct23.pdf> [Accessed 01 December
2023].
Chan, C.K.Y. and Hu, W., 2023. Students' Voices on Generative AI: Perceptions, Benefits,
and Challenges in Higher Education. International Journal of Educational Technology
in Higher Education, 20(43), pp. 1-18. https://doi.org/10.1186/s41239-023-00411-8.
Chen, L., Chen, P. and Lin, Z., 2020. Artificial Intelligence in Education: A Review. IEEE
Access, 8, pp. 75264-75278. https://doi.org/10.1109/ACCESS.2020.2988510.
Chiu, T.K.F., Meng, H., Chai, C.S., King, I., Wong, S. and Yeung, Y., 2022. Creation and
Evaluation of a Pre-Tertiary Artificial Intelligence (AI) Curriculum. IEEE Transactions
on Education, 65(1), pp. 30-39. https://doi.org/10.1109/TE.2021.3085878.
Clarivate - Web of Science, 2023. Web of Science Articles on AI and Education between
1975 and 2023. [online] Clarivate – Web of Science. Available through:
<https://0410q9a1i-y-https-www-webofscience-com.z.e-nformation.ro/wos/woscc/
summary/849a6d7a-6021-4550-8a78-c78dd5824ecb-a622be21/relevance/> [Accessed
21 September 2023].
Duan, Y., Edwards, J.S. and Dwivedi, Y.K., 2019. Artificial Intelligence for Decision
Making in the Era of Big Data – Evolution, Challenges and Research Agenda.
International Journal of Information Management, 48, pp. 63-71.
https://doi.org/10.1016/j.ijinfomgt.2019.01.021.
Duckworth, A. and Ungar, L., 2023. Don’t Ban Chatbots in Classrooms - Use Them to
Change How We Teach. [online] Los Angeles Times, 12 January 2023. Available at:
<https://www.latimes.com/opinion/story/2023-01-19/chatgpt-ai-education-testing-
teaching-changes> [Accessed 23 September 2023].
EdTechHub, 2023. Tools, Resources and Courses to Advance. [online] Available at:
<https://edtechhub.org/edtech-tools/> [Accessed 23 September 2023].
Edwards, B., 2023. Why ChatGPT and Bing Chat Are So Good at Making Things Up.
[online] Available at: <https://arstechnica.com/information-technology/2023/04/why-ai-
chatbots-are-the-ultimate-bs-machines-and-how-people-hope-to-fix-
them/?utm_source=tldrai> [Accessed 23 September 2023].
Garrel, J., Mayer, J. and Mühlfeld, M., 2023. Artificial Intelligence During Studies – A
Quantitative Survey of Students on the Use of ChatGPT & Co. [online] Available at:
<https://tudatalib.ulb.tu-darmstadt.de/handle/tudatalib/3938> [Accessed 22 September
2023].
Gartner, 2023. Gartner Experts Answer the Top Generative AI Questions for Your
Enterprise. [online] Available at: <https://www.gartner.com/en/topics/generative-ai>
Gašević, D., Siemens, G. and Sadiq, S., 2023. Empowering Learners for the Age of
Artificial Intelligence. Computers and Education: Artificial Intelligence, 4, article no.
100130. https://doi.org/10.1016/j.caeai.2023.100130.
Grassini, S., 2023. Shaping the Future of Education: Exploring the Potential and
Consequences of AI and ChatGPT in Educational Settings. Education Sciences, 13(7),
article no. 692. https://doi.org/10.3390/educsci13070692.
Vol. 26 • No. 65 • February 2024 85

Greene-Harper, R.T., 2022. The Pros and Cons of Using AI in Learning: Is ChatGPT
Helping or Hindering Learning Outcomes? eLearning Industry. [online] Available at:
<https://elearningindustry.com/pros-and-cons-of-using-ai-in-learning-chatgpt-helping-
or-hindering-learning-outcomes> [Accessed 23 September 2023].
Greitemeyer, T. and Kastenmüller, A., 2023. HEXACO, the Dark Triad, and Chat GPT:
Who is Willing to Commit Academic Cheating? Heliyon, 9(9), article no. 19909.
https://doi.org/10.1016/j.heliyon.2023.e19909.
Groves, R.M., Fowler, J.F., Couper, M., Lepkowski, M.J., Singer, E. and Tourangeau, R.,
2009. Survey Methodology. Hoboken: John Wiley & Sons.
Guzman, A.L., 2023. Don’t Assume Students are Eager AI Adopters. Inside Higher Ed.
[online] Available at: <https://www.insidehighered.com/opinion/views/2023/04/27/
dont-assume-students-are-eager-ai-adopters> [Accessed 23 September 2023].
Hinkle, D.E., Wiersma, W. and Jurs, S.G., 2003. Applied Statistics for the Behavioral
Sciences. Boston: Houghton Mifflin.
Ibrahim, H., et al., 2023. Perception, Performance, and Detectability of Conversational
Artificial Intelligence Across 32 University Courses. Scientific Reports, 13(1), article
no. 12187.
Iuga, M., 2023. Romania and Generative AI – a Follow Up. [Linkedin] 9 June 2023.
Available at: <https://www.linkedin.com/pulse/romania-generative-ai-follow-up-marin-
iuga> [Accessed 23 September 2023].
Keleş, P.U. and Suleyman, A. 2021. University Students’ Perceptions About Artificial
Intelligence. Shanlax International Journal of Education, 9(1), pp. 212-220.
https://doi.org/10.34293/education.v9iS1-May.4014.
Kim, H.-Y., 2017. Statistical Notes for Clinical Researchers: Chi-Squared Test and Fisher’s
Exact Test. Restorative Dentistry & Endodontics, 42(2), pp. 152-155.
https://doi.org/10.5395/rde.2017.42.2.152.
Lim, W.M., Gunasekara, A., Pallant, J.L., Pallant, J.I. and Pechenkina, E., 2023. Generative
AI and the Future of Education: Ragnarök or Reformation? A Paradoxical Perspective
from Management Educators. International Journal of Management Education, 21(2),
article no. 100790. https://doi.org/10.1016/j.ijme.2023.100790.
Luckin, R., Holmes, W., Griffiths, M. and Forcier, L.B., 2016. Intelligence Unleashed - an
Argument for AI in Education. London: CL Knowledge Lab. [online] Available at:
<https://discovery.ucl.ac.uk/id/eprint/1475756/> [Accessed 23 September 2023].
Marr, B., 2023. A Simple Guide to the History of Generative AI. Available at:
<https://bernardmarr.com/a-simple-guide-to-the-history-of-generative-ai/> [Accessed
21 September 2023].
Marrone, R., Taddeo, V. and Hill, G., 2022. Creativity and Artificial Intelligence – A
Student Perspective. Journal of Intelligence, 10(3), article no. 65.
https://doi.org/10.3390/ jintelligence10030065.
Naumova, E.N., 2023. A Mistake-Find Exercise: A Teacher’s Tool to Engage with
Information Innovations, ChatGPT, and Their Analogs. Journal of Public Health
Policy, 44, pp. 173-178. https://doi.org/10.1057/s41271-023-00400-1.
AE
Niethammer, C., 2020. AI Bias Could Put Women’s Lives at Risk – A Challenge For
Regulators. [online] Forbes. Available at: <https://www.forbes.com/sites/
carmenniethammer/2020/03/02/ai-bias-could-put-womens-lives-at-riska-challenge-for-
regulators/> [Accessed 6 December 2023].
OECD, 2023. The OECD Learning Compass 2030. [online] Available at:
<https://www.oecd.org/education/2030-project/teaching-and-learning/learning/>
Ofosu-Ampong, K., Acheampong, B., Kevor, M.O. and Amankwah-Sarfo, F., 2023.
Acceptance of Artificial Intelligence (ChatGPT) in Education:Trust, Innovativeness and
Psychological Need of Students. Information and Knowledge Management, 13(4),
pp. 37-47.
Park, W. and Kwon, H., 2023. Implementing Artificial Intelligence Education for Middle
School Technology Education in Republic of Korea. International Journal of Technology
and Design Education, pp. 1-27. https://doi.org/10.1007/s10798-023-09812-2.
Rajabi, P., Taghipour, P., Cukierman, D. and Doleck., T, 2023. Exploring ChatGPT’s
Impact on Post-Secondary Education: A Qualitative Study. In: Western Canadian
Conference on Computing Education (WCCCE ’23). Vancouver, Canada, 4-5 May
2023. New York: ACM. Ch.9, pp. 1-6. https://doi.org/10.1145/3593342.3593360.
Sætra, H.S., 2023. Generative AI: Here to Stay, But for Good? Technology in Society, 75,
article no. 102372. https://doi.org/10.1016/j.techsoc.2023.102372.
Salvagno, M., Taccone, F.S. and Gerli, A.G., 2023 Can Artificial Intelligence Help for
Scientific Writing? Critical Care, 27, article no. 75. https://doi.org/10.1186/s13054-
023-04380-2.
Sánchez-Ruiz, L.M., Moll-López, S., Nuñez-Pérez, A., Moraño-Fernández, J.A. and Vega-
Fleitas, E., 2023. ChatGPT Challenges Blended Learning Methodologies in Engineering
Education: A Case Study in Mathematics. Applied Sciences, 13(10), article no. 6039.
https://doi.org/10.3390/app13106039.
Sanusi, I.T., Olaleye, S.A., Agbo, F.J. and Chiu, T.K., 2022. The Role of Learners’
Competencies in Artificial Intelligence Education. Computers and Education: Artificial
Shaji, A.G., Hovan, A.S.G. and Gabrio, M., 2023. ChatGPT and the Future of Work: A
Comprehensive Analysis of AI's Impact on Jobs and Employment. Partners Universal
International Innovation Journal, 1(3), pp. 154-186. https://doi.org/10.5281/
zenodo.8076921.
Shoufan, A., 2023. Exploring Students’ Perceptions of CHATGPT: Thematic Analysis and
Follow-Up Survey. IEEE Access, 11, pp. 38805-38818. https://doi.org/10.1109/
ACCESS.2023.3268224.
Siregar, F.H., Hasmayni, B. and Lubis, A.H., 2023. The Analysis of Chat GPT Usage
Impact on Learning Motivation among Scout Students. International Journal of
Research and Review, 10(7), pp. 632-638. https://doi.org/10.52403/ijrr.20230774.
Sohail, S.S., Madsenb, D.Ø., Himeurc, Y. and Ashrafd, M., 2023. Using ChatGPT to
Navigate Ambivalent and Contradictory Research Findings on Artificial Intelligence.
Frontiers in Artificial Intelligence, 6, article no. 1195797. https://doi.org/10.3389/
frai.2023.1195797.
Vol. 26 • No. 65 • February 2024 87

Swanson, R.A. and Holton, E., 2005. Research in Organizations: Foundations and
Methods of Inquiring. San Francisco: Berrett-Koehler.
Tapalova, O. and Zhiyenbayeva, N., 2022. Artificial Intelligence in Education: AIEd for
Personalised Learning Pathways. Electronic Journal of e-Learning, 20(5), pp. 639-653.
Technological Innovation, 2023. How can AI Help Address the Global Challenges and
Opportunities in Education? [Linkedin] 13 September 2023. Available at:
<https://www.linkedin.com/advice/0/how-can-ai-help-address-global-challenges>
The Artificial Intelligence (AI) for K-12 Initiative (AI4K12), 2023. List of Resources.
[online] Available at: <https://ai4k12.org/resources/list-of-resources/> [Accessed 23
September 2023].
The Knowledge Review, 2023. Using Artificial Intelligence in Education: Pros and Cons.
[online] Available at: <https://theknowledgereview.com/using-artificial-intelligence-in-
education-pros-and-cons/> [Accessed 23 September 2023].
Tîrnăcop, A.B., 2023. Leadership in the Digital Era: Exploring the AI-EI Nexus. Cactus
Tourism Journal, 5(1), pp. 45-52. https://doi.org/10.24818/cts%2F5%2F2023%2F1.05.
UNESCO, 2020. AI in Education: Change at the Speed of Learning. [online] Available at:
<https://unesdoc.unesco.org/ark:/48223/pf0000374947> [Accessed 23 September
2023].
UNESCO, 2023. Guidance for Generative AI in Education and Research. [online]
Available at: <https://www.unesco.org/en/articles/guidance-generative-ai-education-
and-research> [Accessed 23 September 2023].
Vallis, C., Wilson, S., Gozman, D. and Buchanan, J., 2023. Student Perceptions of AI-
Generated Avatars in Teaching Business Ethics: We Might not be Impressed.
Postdigital Science and Education, pp. 1-19. https://doi.org/10.1007/s42438-023-
00407-7.
Yilmaz, R. and Karaoglan Yilmaz, F.G., 2023. The Effect of Generative Artificial
Intelligence (AI)-Based Tool Use on Students’ Computational Thinking Skills,
Programming Self-Efficacy and Motivation. Computers and Education: Artificial
Zhang, C. and Lu, Y., 2021. Study on Artificial Intelligence: the State of the Art and Future
Prospects. Journal of Industrial Information Integration, 23, article no. 100224,
https://doi.org/10.1016/j.jii.2021.100224.
AE
THE SOCIAL IMPACT OF USING ARTIFICIAL INTELLIGENCE

IN EDUCATION
Andreea Simona Săseanu1 , Rodica-Manuela Gogonea2
and Simona Ioana Ghiţă3
1)2)
3)
Bucharest University of Economic Studies, Romania and Institute
of National Economy, Bucharest, Romania

Săseanu, A.S., Gogonea, R.M. and Ghiţă, S.I., 2024. Received: 28 September 2023
The Social Impact of Using Artificial Intelligence in Revised: 11 November 2023
Education. Amfiteatru Economic, 26(65), pp. 89-105. Accepted: 15 December 2023
Abstract
Artificial intelligence currently represents one of the most talked about topics, considering
the need for sustainable economic growth at a global level. When it comes to education,
artificial intelligence is aimed at enhancing systems, ways of learning, as well as at the results
of learning, on the one hand, and training the youth so as to accordingly satisfy the
requirements of their future jobs, on the other hand. In this context, research on higher
education in Romania was conducted, which analysed the students’ opinion on the social
impact of using artificial intelligence in education. As a consequence of performing an
opinion poll, answers were collected online from students from prestigious Romanian
universities. The data registered for the mentioned objective was processed by applying three
statistical and econometric logistic regression models. The results of the first binary logistic
model show the respondents’ opinions on the need and importance of enhancing the learning
experience by using artificial intelligence in education, considering their gender and level of
education. Also, with respect to the two characteristics considered the most significant to the
objective of the paper, the following two multinominal logistic models have been developed.
The results highlight the way in which the use of artificial intelligence in education
influences, on the one hand, the graduates’ prospect for a job and, on the other hand, the
society as a whole.
Keywords: artificial intelligence, education, logistic model, multinominal model, learning
experience, job prospects, society
JEL Classification: C20, D83, I23, J21, L86

Corresponding author, Andreea Simona Săseanu – e-mail: [email protected]
Vol. 26 • No. 65 • February 2024 89

AE The Social Impact of Using Artificial Intelligence in Education
Introduction
The process of sustainable economic development visibly occurs in recent years. This
includes a component that has been increasingly addressed after 2010, artificial intelligence
(AI). In this context, a worldwide manifestation was observed of a global race aimed at
developing, acquiring and funding technologies for artificial intelligence. This leads to
substantial sustainable growth of advanced and emerging economies. Artificial intelligence
(AI) has rapidly developed a significant potential that has visibly occurred in various
disciplines and in many sectors of the economy (Farchy and Denis, 2020).
The use of artificial intelligence in education was aimed at the process of enhancing the
learning environments and levels, as well as the results of learning. Simultaneously, its use
is also aimed at saving the working time of both students and professors. Concurrently, it
should be emphasised that artificial intelligence has had an impact on parents’ capacity of
involvement in children’s education (Mou, 2019).
The involvement of artificial intelligence in education is stimulated by its adequate potential
through the tools and platforms created in favour of both teachers and students. Furthermore,
the artificial intelligence technologies significantly improve the students’ way of learning on
the one hand, and lead to an increased access and greater efficiency of education, on the other
hand. The automated learning process through the use of artificial intelligence technologies
has the ability to customise the learning content. This context allows teachers to properly
evaluate students, thus having the possibility to meet their needs and elevate their
performance level. Additionally, the capacity of artificial intelligence to improve online
training should be noted. As a result, professors get help with finalising the automation
process for routine tasks, modernising the grading process, and offering immediate feedback
to students. Therefore, students get help with better understanding of concepts, on the one
hand, at their own pace, and, on the other hand, at a higher rate of individualisation (Mou,
2019).
The originality and novelty of this study consist in the selection of questions and variables
included in the questionnaire of the online survey and in the results obtained. The
respondents’ profile is highlighted through two variables (gender and level of education) that
constitute basic characteristics used in emphasising the impact of AI in education.
Simultaneously, in the specialty literature, similar studies, that encompass young students’
opinion on the role and impact of artificial intelligence tools used in education - are totally
limited.
The added value of this study also consists in the use of logistic regression models that are
aimed at the interdependencies and impact of AI on the learning experience, on graduates’
employment prospects and on the way it transforms society.
This paper is structured in four sections, introduction excluded. After the introduction, there
is an emphasis on the specialists’ and other interested parties’ points of view. The next section
includes a detailed presentation of the methodology of the research performed, emphasising
on the aspects related to applying the logistic regression models. Performed on the basis of
the research results, the results and discussions section is followed by the conclusions, which
show the interdependency and impact of using AI in education has on the labour market and
on society as a whole. With its content, this paper adds to the studies performed so far
concerning the usefulness of AI in education, thus forming the premises for elaborating and
applying development strategies in the field.
AE
1. Review of scientific literature

Specialists consider that artificial intelligence (AI) should not be defined in a simplistic
manner. Therefore, it represents a set of technological instruments that confer human-like
abilities to machines, from senses, to the capacity to understand, learn, and even create. AI
influences not only people’s lives, their health and education, but also the economy and
society as a whole, in an overwhelming way and rhythm (Goralski and Tan, 2020; Moreno-
Guerrero et al., 2020; Pelau, Ene and Pop, 2021). AI tools interact with impressive volumes
of data, offering people a better and safer life, helping them to enhance their businesses and
to contribute to the sustained development of the world’s economies. Does AI offer real, new
opportunities, or rather does it generate new threats for mankind? According to the 2020
Eurobarometer regarding the attitude of EU citizens to digitalisation, the views on the
benefits and use of digital technologies are rather contradictory, EU citizens having both
positive and negative views, with significant differences amongst groups. For example,
women, elders, less educated people and those who live in childless households self-evaluate
their digital skills at a lower level. There are significant differences between countries as
well, such as in Italy and Hungary, people’s confidence in their own digital skills is lower,
while in Spain, Estonia, or the UK they are polar opposites from this perspective (European
Union – Eurobarometer, 2020).
One of the fields in which AI has a significant presence is the education market, and the key
to this success is the fact that the learning process has been updated due to the introduction
and use of some intelligent training tools (Moreno-Guerrero et al., 2020; Moonpreneur,
2023). However, the role of the teacher continues to be essential (Cope, Kalantzis and
Searsmith, 2020).
Over time, it has been proven that the use of AI tools in the education and learning process
has a contradictory character also, besides the numerous advantages, there are also some less
desired effects. Therefore, among the positive aspects of using AI in education, we mention
the following: increasing accessibility to the learning process and its efficiency, the
possibility to customise learning, the benefit of additional training, increasing the youth’s
motivation and involvement in the learning process, identifying and addressing deficiencies,
weaknesses related to knowledge and skills; the possibility to organise and easily access
information; easier access to education for persons with special needs; using interactive
learning methods and real-time feedback; swift update of the information content; learning
through simulations of real-life situations; bridging the gap when it comes to the students’
socio-economic situations, geographic location, age or gender (Sullins, Craig and Hu, 2015;
Luckin et al., 2016; Hinojo-Lucena et al., 2019; Cukurova, Kent and Luckin, 2019; Ufarte
Ruiz and Manfredi Sánchez, 2019; Wang and Wang, 2019; Knox, 2020; Xiao and Yi, 2020;
Chatterjee and Bhattacharjee, 2020; Van Der Niet and Bleakley, 2020; Felea et al., 2021;
Moonpreneur, 2023). Not lastly, an important advantage is the reduction of costs as well as
the boost in academic performance (Ahmad et al., 2022; Moonpreneur, 2023; Kamalov,
Santandreu Calonge and Gurrib, 2023). Conversely, specialists in the field also mention some
negative aspects, such as: the loss of jobs for professors whose tasks were replaced by AI;
high costs; lack of human contact, of emotional bonds; lack of clear regulations and
instructions for using AI tools in the learning process; the possibility of inequalities,
discriminations in the learning-evaluation process; insufficient measures to guarantee private
data; the dependency on modern technology to access AI tools in the learning process;
managing the limitations of digital systems (data volume, number of users, response time,
Vol. 26 • No. 65 • February 2024 91

etc.) but also moral, ethical concerns (Fahimirad, 2018; Moonpreneur, 2023; Ahmad et al.,
2023). Furthermore, Lai et al. (2023) mention another less desired aspect of using AI learning
systems: the fact that the long term effects of this process on the mental and physical state of
users of smart learning systems are not sufficiently known. They emphasised the negative
correlation between the use of AI in education, on the one hand, and the youth’s social
adaptability and family support, on the other.
The use of AI has effects and consequences on the labour market. It is widely accepted that
a rise in the use of AI will lead to a remodelling of the workforce (Mann Levesque, 2018;
Dumitru and Halpern, 2023). The structure and type of jobs are in a continuous dynamic,
accelerated by the spectacular development of AI tools, which leads to some jobs becoming
obsolete, jobs requiring tasks taken over by machines, in parallel with the occurrence of new
ones. Consequently, the workforce will have to adapt to new requirements in the labour
market, and education systems and AI applications in education will take on the role of
performing this transition more easily and with new tools (Mann Levesque, 2018; Dumitru
and Halpern, 2023). AI is increasing its importance in the workplace (Dellermann et al., 2018;
Sowa, Przegalinska and Ciechanowski, 2021; Füller et al., 2022), changing its role from
executor to leader, with the possibility of reducing employees' promotion chances (Yam et
al., 2022; Tsai et al., 2022). AI tools applied in education (including lifelong training) can
ensure long-term competitiveness for future employees on the labour market, employees
being motivated to train in order to acquire the skills to work with AI (Ciarli et al., 2021).
The use of these tools comes, however, with some risks. Mann Levesque (2018) mentions
the risk of deepening the gaps, discrepancies, and inequalities that already exist between the
educational results of youth depending on race, ethnicity, or income. Frequently, access to
quality education comes easier for youth from families with a better financial standing or
from a better residential setting. AI can deepen the mismatch between the skills and
knowledge of the workforce with the competencies required by new technologies (Acemoglu
and Restrepo, 2019), with people often perceiving AI as a threat to their jobs, generating fear
of unemployment and job insecurity (Brougham and Haar, 2020; Dodel and Mesch, 2020).
Other studies on the effects of the use of AI on the labour market aim at reducing employment
and wages (Acemoglu and Restrepo, 2020), increasing labour productivity, reducing
bargaining power, and changing the length of the work schedule (Huang and Sharif, 2017;
Koch, Manuylov and Smolka, 2021), as well as the influence of AI on workplace learning,
more prominent for older employees, women and those with a lower level of education, for
those with less work experience and autonomy (Li et al., 2023).
The specialty literature presented has emphasised the motivation at the basis of formulating
and performing the research in this article, as premises for conceiving and applying
development strategies in education.
The general scope of this research is represented by the analysis of Romanian students’ views
on the social impact of using artificial intelligence in education. Its accomplishment was
aimed at three main objectives:
 analysing the impact of using artificial intelligence in education on the improvement of
the learning experience;
AE
 the way in which the use of artificial intelligence in education influences the graduates’
prospect for a job;
 the way in which the use of artificial intelligence in education influences society as a
whole.
The data included in this analysis was collected through an opinion poll performed on the
basis of a questionnaire filled in online on the isondaje.ro platform. The 606 students
(Bachelor’s and Masters’ degree) who have answered the questionnaire questions come from
prestigious universities in Romania (Bucharest University of Economic Studies, University
Politehnica of Bucharest, University of Bucharest, “Ștefan cel Mare” University of Suceava,
“Alexandru Ioan Cuza” University of Iași, “Constantin Brâncuşi” University of Târgu Jiu,
University of Craiova, University of Oradea, Babeș-Bolyai University of Cluj, the Western
University of Timișoara). Sample size was determined on the basis of the probability function
argument (z), variance (s2) and margin of error (∆𝑥̅), thus: (𝑧2∙ 𝑠2) ⁄ ∆𝑥̅2. The data collection
period lasted one semester, in the period February-June 2023.
The test contained five questions, two of which targeted the characteristics of the respondents
(gender and education level), and three resulted from the objectives of the study. These three
questions were: “Do you think the use of artificial intelligence in education has improved
your learning experience?”; “How do you think the use of artificial intelligence in education
will affect employment prospects?”; “How do you think the use of artificial intelligence in
education will affect society as a whole?” and the answers were presented in table no. 1.
In order to define respondents’ profiles, two factorial variables were used in their processing,
which constitute their main characteristics: gender (GEN) and level of education (EDLV) as
well as three outcome variables which are aimed at the students’ opinions regarding:
improvement of the learning experience through the use of artificial intelligence in education
(AIILE) and the way in which the use of artificial intelligence in education influences on the
one hand the graduates’ prospect for a job (EMPT) and, on the other hand, the society as a
whole (AIEAS). Table no. 1 presents the identifiers and significance of the variables that
shall be subsequently used on the models applied.
As a consequence of studying the results of the main studies in the specialty literature and
considering the variables presented, the following hypotheses were formulated:
• Hypothesis 1: There is interdependency between the improvement of the learning
experience and the use of artificial intelligence in education.
• Hypothesis 2: There is interdependency between the use of artificial intelligence in
education and graduates’ prospect for a job.
• Hypothesis 3: There is interdependency between the use of artificial intelligence in
education and the development of society.
Additionally, in the case of the three hypotheses there was the aim to study whether there are
significant differences amongst students’ opinions depending on gender and stage of their
education process, basically creating a profile for them.
Vol. 26 • No. 65 • February 2024 93

Table no. 1. Types, coding, and significances of the variables included in the analysis
Variables Type Coding and meaning of the answers

using artificial intelligence in education will
YES
NOMINAL improve learning experience
AIILE
BINARY using artificial intelligence in education will not
NOT
improve learning experience
the use of artificial intelligence in education will
SIGOPP
create significant opportunities on employment
Output variables

EMPT ORDINAL MEDIMP
have a medium positive impact on employment
NOIMP
not have a significant impact on employment
MESF
create a more educated and skilled workforce
AIEAS NOMINAL IAEO increase inequality in access to education and
employment opportunities
JLED
lead to job losses and economic disruption
NOMINAL FEMALE Female respondents
GEN
Input variables
BINARY MALE Male respondents

BY1 First year of undergraduate studies
BY2 Second year of undergraduate studies
EDLV NOMINAL BY3 Third year of undergraduate studies
MY1 First year of university master's studies
MY2 Second year of university master's studies
The achievement of the objectives presented was performed as a consequence of a statistical

and econometric processing based on the logistic regression model.
For the analysis of the results and the identification of the quantitative and qualitative
connections between the respondents’ opinions, besides statistical analyses, we identified and
used logistic regression models, with n factorial variables, in the form:
n n
ln(odds)  ln( k )   ln(ORi )  a   Bi  xi
i 1 i 1 (1)
p
where odds  represents the opportunity for the event to take place, whose likelihood
1 p
odds k
ratio is p, and ORk  is Odds Ratio.
oddsbaza
Under these conditions, the odds valuei to opt for a certain statement expressing a certain
standpoint of respondents, depending on the factorial variables is:
n n
ln(k )  ln(ORi ) a   Bi x
oddsi  e i 1 e i 1 (2)
On this basis (equations no. (1) and (2)), in order to analyse students’ opinions on the social
impact of using artificial intelligence in education, we developed three regression models,
one of which being a binary logistic model and the other two being multinominal logistic
AE
models. The binary logistic model aims to inform students’ opinions on the enhancement of
the learning experience by using artificial intelligence in education (AIILE), and
multinominal models are aimed at students’ opinions on the way in which the use of artificial
intelligence in education influences graduates’ prospects for employment (EMPT), on the
one hand, and their opinions on the way in which artificial intelligence in education will
affect society as a whole, on the other hand (AIEAS). Each of these contains two factorial
variables: GEN variable (respondents’ gender) and EDLV variable (respondents’ level of
education).
The fact that data collection was performed by using an application specialised in surveys
(isondaje.ro) has allowed for the processing and analysis of data with the help of the SPSS
software.

The data processing and analysis began with the two characteristics defining respondents’
profile: gender and level of education. Therefore, the sample structure on the basis of
respondents’ gender is as follows: 49.01% male students and 51.99% female students, similar
to the school population structure at university level, according to data supplied by the
National Institute of Statistics (National Institute of Statistics, 2023). Concurrently, the
sample structure when it comes to the education level (Bachelor’s or Master’s degree) is as
follows: 21.4% of the total number of respondents are in their first year of Bachelor’s studies,
13.5% in their second year of Bachelor’s studies, 21.9% in their third year of Bachelor’s
studies, 20.8% in their first year of Master’s studies, 22.4% in their second year of Master’s
studies. Next, three questions were extracted from the questionnaire, which were significant
for the achievement of the research objectives. The first question has binary answers (Yes,
No): “Do you consider that the use of artificial intelligence in education has improved your
learning experience?”. The answers were structured in 60.8% students who consider that the
use of artificial intelligence in education has improved the learning experience and 39.2%
students who gave a negative answer. The second question extracted from the questionnaire,
which is included in the processing: “How do you think the use of artificial intelligence in
education will impact employment prospects?” has three possible answers: “It will create
significant opportunities” (SIGOPP), “It will have a medium positive impact” (MEDIMP),
“It will not have a significant impact” (NOIMP).
The structure of students’ opinions on the impact of using artificial intelligence in education
on employment prospects shows that most students (59%) believe that artificial intelligence
will have a medium positive impact on employment prospects. With a 35% difference from
the maximum percentage, we find the respondents who said that artificial intelligence will
create significant opportunities for employment prospects. Only 16% of the total respondents
consider that artificial intelligence used in education will not have a significant impact on
employment prospects. The third question extracted from the questionnaire is: “How do you
think the use of artificial intelligence in education will affect society as a whole?” and it also
has three possible answers: “It will create a more educated and skilled workforce” (MESF),
“It will increase inequality when it comes to the access to education and to the employment
opportunities” (IAEO), “It will lead to the loss of jobs and economic disturbances” (JLED).
The students’ answers to this topic are not very optimistic, although they reflect reality
(Figure no. 1). Most students (37%) consider that the use of artificial intelligence in education
Vol. 26 • No. 65 • February 2024 95

leads to the loss of jobs and to economic disturbances affecting society, as a whole. 34% of
the respondents consider that artificial intelligence in education will increase inequality when
it comes to the access to education and to the employment opportunities. Only 29% of the
total number of respondents consider that the use of artificial intelligence in education will
create a more educated and skilled workforce. With slight variations, it is notable that
respondents’ views are quite uniformly distributed across the three possible answers.
Figure no. 1. Structure of students’ opinions on the impact

of using artificial intelligence in education on society
Considering the variables and hypotheses presented, three logistic regression models have
been applied.
The first model resulting from data analysis and processing evaluates the students’ answers
on the improvement of the learning experience through the use of artificial intelligence in
education (AIILE). The model includes two explanatory categorical variables: the binary
variable GEN and the multinominal variable EDLV, the bases being the variables GEN_MALE
(male students’ opinions) and EDLV_MY2 (the opinions of the second year of Master’s
students). The model characteristics are presented in Table no. 2.
Table no. 2. Characteristics of the binominal binary logistic model according
to students’ opinions on the improvement of the learning experience through the use
of artificial intelligence in education
Hosmer & Lemeshow
Chi-square =2.265 df=4 Sig.=0.587
Test
Variables B Wald df Sig. Exp(B) AIILE
GEN(FEMALE) .108 4.445 1 0.035 1.115 oddsMALE oddsFEMALE
EDLV 9.537 4 0.049 NOT YES NOT YES
EDLV(BY1) .452 7.035 1 0.008 1.571 0.285 3.505 0.318 3.145
EDLV(BY2) .346 3.025 1 0.082 1.414 0.257 3.895 0.286 3.495
EDLV(BY3) .322 2.857 1 0.091 1.380 0.251 3.990 0.279 3.580
EDLV(MY1) .383 5.502 1 0.019 1.467 0.266 3.753 0.297 3.367
Constant -1.706 8.423 1 0.004 0.182 0.182 5.506 0.202 4.940
Note: Reference Category: Last (GEN_MALE and ELDV_MY2)
AE
The values of the Hosmer & Lemeshow test (the Goodness of fit) lead to the acceptance of
the null hypothesis: the chosen model fits the data well. Considering the values of the Wald
statistics and the level of significance Sig., corresponding to the EDLV(BY2) and
EDLV(BY3) variables, we can state that the parameters of the explanatory variables are
statistically significant (for a significance level of 5%, respectively 10%), according to the
hypotheses developed and presented in the methodology section. By analysing the results
obtained (Table no. 2), we see the greatest value of odds is registered in the case of male
students in their second year of Master’s studies.
𝑜𝑑𝑑𝑠𝑌𝐸𝑆_𝑀𝐴𝐿𝐸_𝑀𝑌2 = 1/𝑜𝑑𝑑𝑠𝑁𝑂𝑇𝑀𝐴𝐿𝐸 = 1/𝑒 (−1.706) = 5.506

𝑀𝑌2 (3)
It is then closely followed by the female students in the same year of studies:
𝑜𝑑𝑑𝑠𝑌𝐸𝑆_𝐹𝐸𝑀𝐴𝐿𝐸_𝑀𝑌2 = 1/𝑜𝑑𝑑𝑠𝑁𝑂𝑇𝑀𝐴𝐿𝐸 = 1/𝑒 (−1.706+0.108) = 4.940
𝑀𝑌2 (4)
The results obtained, including from equations (3) and (4), emphasise the fact that the most
convinced students of the positive role of artificial intelligence in education in enhancing the
learning experience are students in their last year of Master’s degree studies, as well as the
ones in the last year of their Bachelor’s degree studies, therefore, students in the final years
of each cycle of studies. In their case, the odds values highlight the fact that the chances of
agreeing with the statement that the use of artificial intelligence in education has improved
the learning experience are 3.99 times greater in the case of male students and 3.58 times
greater in the case of female students, compared to the chances of negating this statement. In
contrast, the ratio among the students’ genders related to their opinions about improving the
education experience with the use of artificial intelligence, leads to the conclusion that,
despite the slight difference, male students are more confident than female students
(oddsAIILE_MALEYES/oddsIILE_FEMALE_YES=1.115). On the other hand, both in the case of
Bachelor’s degree students, as well as in Master’s degree students, the chances of stating that
the use of artificial intelligence in education will improve the learning experience, compared
to rejecting this statement, increase proportionally with the students’ promotion to the next
year of study. Therefore, for example, for female students, from an odds value of 3.145 for
first-year students, to 3.495 for second-year students, up to 3.580 for third-year Bachelor’s
degree students, which represents a 1.138 times increase. In the case of Master’s degree
students, the odds value increase is even more significant, from 3.367 for first year students,
to 4.940 for second year students, which represents a 1.467 times increase. The results
presented are similar to those obtained by Cukurova, Kent and Luckin (2019), Xiao and Yi
(2020), or Chatterjee and Bhattacharjee (2020).
The second model analyses students’ opinions on the way in which the use of artificial
intelligence in education influences the graduates’ prospect for a job (EMPT). In this model,
amongst the possible answers, the third statement was chosen as a basis: the use of artificial
intelligence in education will not have a significant impact on the employment prospects
(NOIMP). The characteristics of the logistic multinominal model AIEAS are presented in
table no. 3.
Considering the values of the Wald statistics and the level of significance (Sig.), except for
SIGOPP_EDLV_MY1 and MEDIMP_EDLV_MY1 coefficients for which the confidence
level is at least 90% (Sig.<0.1), in the case of the others the confidence level is 95%
Vol. 26 • No. 65 • February 2024 97

(Sig.<0.05), as well as LHR values, the conclusion is that, on the one hand, the introduction
of explanatory variables brings more information, and on the other hand, considering the
confidence levels presented, the model coefficients are statistically significant.
Table no. 3. Characteristics of the multinominal logistic model for the analysis
of students’ opinions on the influence of artificial intelligence in education
on graduates’ prospects for employment
Likelihood Ratio Test (LHR) Chi-square=19.267 df=10 Sig.=0.037
MALE FEMALE
EMPPa B Wald df Sig. Exp(B) log log
odd odd
(odd) (odd)
Intercept 0.741 4.598 1 .032
Female -.188 4.019 1 .045 1.334
GEN
Male 0b 0
SIGOPP
BY1 -.293 5.169 1 .023 .746 0.448 1.565 0.260 1.297

BY2 -.462 6.465 1 .011 .630 0.279 1.322 0.091 1.096
EDLV
BY3 .263 5.024 1 .025 1.301 1.004 2.729 0.816 2.261

MY1 -.289 3.291 1 .070 .169 0.452 1.572 0.264 1.303
MY2 0 0 0.741 2.098 0.553 1.739
Intercept 0.978 3.910 1 .048
Female .153 4.350 1 .037 1.451
GEN
Male 0b 0
MEDIMP
BY1 -.944 3.380 1 .066 .389 0.034 1.034 0.186 1.205

BY2 -1.038 4.135 1 .042 .354 -0.060 0.942 0.093 1.097
EDLV
BY3 -0.815 4.956 1 .026 .443 0.163 1.177 0.316 1.371

MY1 -0.779 3.260 1 .071 .459 0.199 1.220 0.352 1.421
MY2 0 0 0.978 2.659 1.131 3.097
Note: a. The reference category is: NOIMP; b. This parameter is set to zero because it is redundant.
The analysis of students’ opinions on the way in which the use of artificial intelligence in
education will influence graduates’ prospects for employment emphasises a series of
particularities from the point of view of the level and year of study, as well as from the point
of view of their gender. Therefore, the greatest value of odds with respect to ensuring
significant employment opportunities (SIGOPP) is registered in the case of male third-year
Bachelor’s students (5).
𝑜𝑑𝑑𝑠𝑆𝐼𝐺𝑂𝑃𝑃_𝐵𝑌3_𝑀𝐴𝐿𝐸 = 𝑒 (0.741+0.0+0.263) = 2.729 (5)
it is then closely followed by the female students in the same year of studies (6), for which:
𝑜𝑑𝑑𝑠𝑆𝐼𝐺𝑂𝑃𝑃_𝐵𝑌3_𝐹𝐸𝑀𝐴𝐿𝐸 = 𝑒 (0.741−0.188+0.263) = 2.261 (6)
This means that, for third year Bachelor’s degree students, the chances of considering that
the use of artificial intelligence in education offers significant opportunities for employment
are 2.729 times greater than the pessimistic point of view (NOIMP), of male students,
respectively 2.261 times greater for female students. Comparing the two values, it results that
male students are more optimistic (oddsSIGOPP_MALE/oddsSIGOPP_FEMALE = 1.207). Additionally,
significant values of odds were registered in the case of second-year Master’s degree
AE
students, respectively 2.098 for male students and 1.731 for female students. The second-
year Bachelor’s degree students exhibit the least expectations concerning significant
opportunities for employment, for which the odds values are 1.322, in the case of male
students and 1.096, for female students. When it comes to the opinion according to which the
use of artificial intelligence in education will have a medium positive impact on the students’
employment prospects (MEDIMP), the greatest values of odds are registered in the case of
second-year Master’s degree students, respectively 2.659 for male students and 3.097 for
female students and the lowest values of odds belong to second-year Bachelor’s degree
students, respectively 0.942 for male students and 1.097 for female students. Unlike those
who consider that the use of artificial intelligence in education offers significant employment
prospects, in the case of those with medium opinion, the ratios between the genders are
reversed, here oddsMEDIMP_MALE/oddsMEDIMP_FEMALE=0.859. Overall, considering both the
maximum values of odds, as well as the ratios between the students’ opinions depending on
their education cycle, it results that third year Bachelor’s students are the most optimistic
about the impact of using artificial intelligence in education on providing significant
opportunities for employment, which is emphasised by the values of ratios between the
oddsSIGOPP_MALE_BY3/oddsSIGOPP_MALE_MY2=1.301, respectively between the
oddsSIGOPP_MALE_BY3/oddsMEDIMP_MALE_MY2=1.027, as well as between the
oddsSIGOPP_MALE_BY3/oddsMEDIMP_MALE_BY3=2.318. With respect to the influence of students’
gender on their opinion regarding the impact of artificial intelligence in education on the
employment prospects, it resulted that female Bachelor’s degree students are less optimistic,
the majority of odds values pertaining to the opinion about the medium impact on
employment prospects (MEDIMP) being greater than the results of odds aimed at the
existence of significant opportunities for employment (SIGOPP). When it comes to Master’s
degree students, the greatest value of odds pertains to female second-year Master’s degree
students (3.097), which emphasises that their dominant opinions suggest a medium positive
impact on employment prospects. These conclusions are also highlighted by the ratios of
odds values between genders, which emphasise the more reserved opinions of female
students. The results of the second model are similar to those obtained by Mann Levesque
(2018) or Ciarli et al. (2021).
The third model estimates student’s opinions regarding the way in which the use of artificial
intelligence in education will influence society, as a whole (AIEAS). In the AIEAS model,
from the possible answers, the first statement was chosen as the basis: the use of artificial
intelligence in education will create a more educated and skilled workforce (MESF). The
characteristics of the AIEAS logistic multinominal model are presented in Table no. 4.
Testing the statistical significances of the coefficient values for the AIEAS model was
performed with the Wald test. Taking into account the values of the significance level (Sig.),
the coefficients of the variables IAEO_EDLV_MY1, JLED_Intercept and JLED_FEMALE
are statistically significant for the confidence level of at least 90% (Sig.<0.1), and the other
coefficients are statistically significant for the 95% confidence level (Sig.<0.05).
Vol. 26 • No. 65 • February 2024 99

Table no. 4. Characteristics of the AIEAS logistic multinominal model for the analysis
of student’s opinions regarding the way in which the use of artificial
intelligence in education will influence society, as a whole
Likelihood Ratio Test (LHR) Chi-square=18,646 df=10 Sig.=0,045
MALE FEMALE
AIEASa B Wald df Sig. Exp(B)
log(odd) odd log(odd) odd
Intercept .227 8.284 1 .004
Female -.131 4.176 1 .041 0.87
GEN
b
Male 0 0
BY1 -.307 4.653 1 .031 0.735 -0.080 0.923 -0.211 0.810
IAEO
BY2 -.464 7.033 1 .008 0.629 -0.236 0.790 -0.367 0.693

EDLV
BY3 -.693 4.135 1 .042 0.500 -0.466 0.627 -0.597 0.550

MY1 -.356 3.148 1 .076 0.700 -0.129 0.879 -0.260 0.771
MY2 0 0 1.000 0.227 1.255 0.096 1.101
Intercept .306 2.986 1 .084
Female .266 3.537 1 .060 1.30
GEN
b
Male 0 0
BY1 -.522 4.891 1 .027 0.59

JLED
-0.215 0.806 0.051 1.052

BY2 -.252 4.445 1 .035 0.77 0.055 1.056 0.320 1.378
EDLV
BY3 .067 3.910 1 .048 1.07 0.374 1.453 0.639 1.895

MY1 -.199 5.803 1 .016 0.81 0.107 1.113 0.373 1.452
MY2 0 0 0.306 1.359 0.572 1.772
Note: a. The reference category is: MESF. b. This parameter is set to zero because it is redundant
A first conclusion resulting from the analysis of students’ opinions regarding the way in
which the use of artificial intelligence in education will affect society as a whole is that the
highest value of odds is registered in the case of female students in their third year of
Bachelor’s studies, who consider that the effect with the highest chances of occurrence is job
loss:
𝑜𝑑𝑑𝑠𝐽𝐿𝐷𝐸_𝐹𝐸𝑀𝐴𝐿𝐸_𝐵𝑌3 = 𝑒 (0.306+0.266+0.067) = 1.895
(7)
This means that, according to the result no. (7), from their point of view, by using artificial
intelligence in education, the chance of losing jobs is 1.895 times higher than the one of
creating a more educated and more skilled workforce, which is shown through
oddJLED_FEMALE_BY3/oddIAEO_FEMALE_BY3=3.445 times greater, than the chance of increasing the
inequality when it comes to the access to education and to the employment opportunities.
Second year Master’s degree students also have a similar standpoint. In this case as well, the
highest values of odds are also registered in the case of female students (8):
𝑜𝑑𝑑𝑠𝐽𝐿𝐷𝐸_𝐹𝐸𝑀𝐴𝐿𝐸_𝑀𝑌2 = 𝑒 (0.306+0.266+0.0) = 1.772 (8)

AE
Under these conditions, the ratio of odds values between genders regarding possible losses
of jobs caused by the use of artificial intelligence in education is
oddJLED_FEMALE/oddIAEO_MALE=1.304, which means that female population is more fearful of
the job loss than the male one. Furthermore, it should be pointed out that the fear of job loss
is increasing as the students approach the last years of studies, therefore approaching the
moment of entering the labour market. For example, for male Bachelor’s degree students, the
odds values for first-year students increase from 0.806 to 1.056 for second-year students and
up to 1.453 for third-year Bachelor’s degree students. The same situation is registered in the
case of Master’s degree students, the odds value in the case of male first year students is
1.113 (greater value than in the case of first year and second year Bachelor’s degree students)
and in the case of second year Master’s degree students it is 1.359. Regarding student
opinions on the statement that the use of artificial intelligence in education will lead to an
increase in inequality when it comes to the access to education and to the employment
opportunities (IAEO), the ratios with the MESF statement (the use of artificial intelligence
in education will create a more educated and skilled workforce) chosen as a basis are
reversed. Therefore, except for second year Master’s degree students for which the values of
oddIAEO are greater than one (1.255 for male students and 1.101 for female students), in all
other cases the registered values are less than one. The lowest values of oddIAEO were
registered in the case of third-year Bachelor’s degree students, of 0.672, in the case of male
students, and 0.355, for female students. This means that the odds of considering the use of
artificial intelligence in education will create a more educated and skilled workforce (MESF),
compared to the one related to the increase in inequality when it comes to the access to
education and to the employment opportunities (IAEO), is 1.594 for male students
(1/oddIAEO_MALE_BY3=1/0.627) and 1.817 for female students (1/oddIAEO_FEMALE_BY3=1/0.355).
The results of the third model are similar to those obtained by Mann Levesque (2018),
Brougham and Haar (2020), or Ciarli et al. (2021).
Conclusions
The research performed an analysis of the opinions of students from prestigious universities
in Romania on the social impact caused by the use of artificial intelligence in education. An
opinion poll was developed and respondents were required to fill out a questionnaire online,
but, in order to accomplish the mentioned objective, only a few questions deemed significant
in this sense were selected. Considering two significant characteristics of respondents
(gender and level of education) and three variables of interest for expressing the impact of
using artificial intelligence in education (aimed at improving the learning experience,
employment prospects and changes to the society) three logistic regression models were
applied.
The results obtained by applying a binary logistic model, referring to the respondents’
opinions on the improvement of the learning experience through the use of artificial
intelligence in education, highlighted that, as the students are in higher years of study of the
bachelor’s and master’s cycles, compared to those in the first year, the greater the chance of
improving the learning experience through the use of artificial intelligence. At the same time,
the most convinced of the positive effect of artificial intelligence on the improvement of the
learning experience are the male students in the last year of the master’s degree, followed by
those in the third year of the bachelor's degree. Female students are less confident than male
students, regardless of education level. Regarding the students’ opinions on the employment
Vol. 26 • No. 65 • February 2024 101

prospects of the graduates, as a result of the use of artificial intelligence in education, they
were analysed using a multinominal regression model. The conclusion was that female
students in their third year of Bachelor’s studies and second year of Master’s degree studies
consider that the chance of losing jobs is 1.895 times higher than that of creating a more
educated and more skilled workforce. Concurrently, this chance is 3.445 times greater than
that of increasing inequality when it comes to the access to education and to employment
opportunities. The fear of losing a job increases in intensity as students approach the last
years of studies, the odds values increase from less than one in the first year of studies, to
more than one in the last year of studies. The exceptions are second year Master’s degree
students, the odds in their case being much lower than that of creating a more educated and
skilled workforce. From the analysis of odd-ratios it results that, according to respondents’
views, the use of artificial intelligence in education will, first of all, lead to the loss of jobs
and, only secondly, to creating a more educated and skilled workforce.
A general conclusion is that respondents who are in their final years of Bachelor’s and/or
Master’s degree studies are much more aware of the impact of using artificial intelligence in
education. Male students are more optimistic, both with respect to the chance of improving
the learning experience, as well as to the significant and medium effect of using AI in
education on the employment prospects. With regard to female students’ answers, a more
reserved attitude is formed as for the positive impact of using artificial intelligence in
education.
The research results offer an overview of the perception of young students enrolled in higher
education institutions in Romania on the multiple influences the use of AI has on education.
The research results can be starting points in designing the implementation of artificial
intelligence in student-oriented education, in becoming aware of skills in order to elevate
them, and in creating a positive attitude in the context of sustainable development.
The research limits are represented, on the one hand, by the fact that the research was
performed only at the level of university students, and, on the other hand, by the variables
presented. Expanding the research could also include other categories of people from the
field of education (i.e., professors) as well as other levels of education.
As it appears from other studies mentioned (Moonpreneur, 2023), the applicability of the
research results is useful not only at the academic level. Students’ perception on the impact
of artificial intelligence on education is also important for future workplaces in the direction
of their efficiency. At the same time, artificial intelligence is also important for the transition,
training, and adaptation of young graduates to the requirements of future jobs, in the
development and application of specialised work training programs.
Highlighting the usefulness of artificial intelligence in the process of assessing, smart
guidance, developing the level of knowledge at global level, of learning and perfecting at any
level of education, of adapting the demand to the offer of skills, represent many perspectives
of expansion for future research studies, which would analyse other institutions and fields of
activity.
References
Acemoglu, D. and Restrepo, P., 2019. Artificial intelligence, automation, and work. In:
A. Agrawal, J. Gans and A. Goldfarb eds., 2019. The Economics of Artificial Intelligence:
An Agenda. Chicago, IL, USA: University of Chicago Press, pp. 197-236.

AE
Acemoglu, D. and Restrepo, P., 2020. Robots and Jobs: Evidence from US Labor
Markets. Journal of Political Econonmy, [e-journal] 128(6), pp. 2188-2244.
https://doi.org/10.1086/705716.
Ahmad, S.F., Alam, M.M., Rahmat, M.K., Mubarik, M.S. and Hyder, S.I., 2022. Academic
and Administrative Role of Artificial Intelligence in Education. Sustainability, 14(3),
article no. 1101. https://doi.org/10.3390/su14031101.
Ahmad, S.F., Han, H., Alam, M.M., Rehmat, M.K., Irshad, M., Arraño-Muñoz, M. and
Ariza-Montes, A., 2023. Impact of artificial intelligence on human loss in decision
making, laziness and safety in education. Humanities and Social Sciences
Communications, [e-journal] 10, article no. 311. https://doi.org/10.1057/s41599-023-
01787-8.
Brougham, D. and Haar, J., 2020. Technological disruption and employment: The influence
on job insecurity and turnover intentions: A multi-country study. Technological
Forecasting and Social Change, [e-journal] 161, article no. 120276.
education: A quantitative analysis using structural equation modelling. Education and
Information Technologies, [e-journal] 25(5), pp. 3443-3463. https://doi.org/10.1007/
s10639-020-10159-7.
Ciarli, T., Kenney, M., Massini, S. and Piscitello, L., 2021. Digital technologies, innovation,
and skills: Emerging trajectories and challenges. Research Policy, [e-journal] 50(7),
article no. 104289. https://doi.org/10.1016/j.respol.2021.104289.
Cope, B., Kalantzis, M. and Searsmith, D., 2020. Artificial intelligence for education:
Knowledge and its assessment in AI-enabled learning ecologies. Educational Philosophy
and Theory, [e-journal] 53(12), pp. 1229-1245. https://doi.org/10.1080/00131857.
2020.1728732.
Cukurova, M., Kent, C. and Luckin, R., 2019. Artificial intelligence and multimodal data in
the service of human decision-making: A case study in debate tutoring. British Journal of
Educational Technology, [e-journal] 50(6), pp. 3032–3046. https://doi.org/10.1111/
bjet.12829.
Dellermann, D., Lipusch, N., Ebel, P. and Leimeister, J.M., 2018. Design principles for a
hybrid intelligence decision support system for business model validation. Electronic
Markets, [e-journal] 29(3), pp. 423-441. https://doi.org/10.1007/s12525-018-0309-2.
Dodel, M. and Mesch, G.S., 2020. Perceptions about the impact of automation in the
workplace. Informaiton, Commununication & Society, [e-journal] 23(5), pp. 665-680.
https://doi.org/10.1080/1369118X.2020.1716043.
Dumitru, D. and Halpern, D.F., 2023. Critical Thinking: Creating Job-Proof Skills for the
Future of Work. Journal of Intelligence, [e-journal] 11(10), article no. 194.
https://doi.org/10.3390/jintelligence11100194.
European Union – Eurobarometer, 2020. Attitudes towards the Impact of Digitalisation on
Daily Lives. [online]. Available at: <https://europa.eu/eurobarometer/surveys/
detail/2228> [Accessed 12 august 2023].
Fahimirad, M., 2018. A Review on Application of Artificial Intelligence in Teaching and
Learning in Educational Contexts. International Journal of Learning and
Development, [e-journal] 8(4), pp. 106-118. https://doi.org/10.5296/ijld.v8i4.14057.
Vol. 26 • No. 65 • February 2024 103

Farchy, J. and Denis, J., 2020. Artificial intelligence. [online] Available at:
<https://EconPapers.repec.org/RePEc:hal:journl:halshs-04006773> [Accessed 22 July
2023].
Felea, M., Bucur, M., Negruțiu, C., Nițu, M. and Stoica, D.A., 2021. Wearable Technology
Adoption Among Romanian Students: A Structural Model Based on TAM. Amfiteatru
Economic, [e-journal] 23(57), pp. 376-391. https://doi.org/10.24818/EA/2021/57/376.
Füller, J., Hutter, K., Wahl, J., Bilgram, V. and Tekic, Z., 2022. How AI revolutionizes
innovation management – Perceptions and implementation preferences of AI-based
innovators. Technological Forecasting and Social Change, [e-journal] 178,
article no. 121598. https://doi.org/10.1016/j.techfore.2022.121598.
Goralski, M.A. and Tan, T.K., 2020. Artificial intelligence and sustainable development.
International Journal of Management Education, [e-journal] 18(1), article no. 100330.
https://doi.org/10.1016/j.ijme.2019.100330.
Hinojo-Lucena, F.-J., Díaz, I.A., Reche, M.P.C. and Romero-Rodríguez, J.-M., 2019.
Artificial Intelligence in Higher Education: A Bibliometric Study on its Impact in the
Scientific Literature. Education Sciences, [e-journal] 9(1), article no. 51.
https://doi.org/10.3390/ educsci9010051.
Huang, Y. and Sharif, N., 2017. From ‘labour dividend’ to ‘robot dividend’: Technological
change and workers’ power in South China. Agrarian South: Journal of Political
Economy, [e-journal] 6(1), pp. 53-78. https://doi.org/10.1177/2277976017721284.
Kamalov, F., Santandreu Calonge, D. and Gurrib, I., 2023. New Era of Artificial Intelligence
in Education: Towards a Sustainable Multifaceted Revolution. Sustainability, [e-journal]
15(16), article no. 12451. https://doi.org/10.3390/su151612451.
Knox, J., 2020. Artificial intelligence and education in China. Learning Media Technology,
[e-journal] 45(3), pp. 298-311. https://doi.org/10.1080/17439884.2020.1754236.
Koch, M., Manuylov, I. and Smolka, M., 2021. Robots and firms. Economic Journal, 131,
pp. 2553-2584.
Lai, T., Xie, C., Ruan, M., Wang, Z., Lu, H. and Fu, S., 2023. Influence of artificial
intelligence in education on adolescents’ social adaptability: The mediatory role of social
support. PLoS ONE, [e-journal] 18(3), article no. e0283170. https://doi.org/10.1371/
journal.pone.0283170.
Li, C., Zhang, Y., Niu, X., Chen, F. and Zhou, H., 2023. Does Artificial Intelligence Promote
or Inhibit On-the-Job Learning? Human Reactions to AI at Work. Systems,
[e-journal] 11(3 article no. 114. https://doi.org/10.3390/systems11030114.
Luckin, R., Holmes, W., Griffiths, M. and Forcier, L.B., 2016. Intelligence Unleashed: An
Argument for AI in Education. London, UK: Pearson Education.
Mann Levesque, E., 2018. The Role of AI in Education and the Changing US Workforce,
Brookings. [online]. Available at: <https://www.brookings.edu/articles/the-role-of-ai-in-
education-and-the-changing-u-s-workforce/> [Accessed 12 September 2023].
Moonpreneur, 2023. AI’s Negative and Positive Impact on Education Industry. [online].
Available at: <https://moonpreneur.com/blog/ai-negative-and-positive-impact-on-
education-industry> [Accessed 5 September 2023].

AE
Moreno-Guerrero, A.-J., López-Belmonte, J., Marín-Marín, J.-A. and Soler-Costa, R., 2020.
Scientific Development of Educational Artificial Intelligence in Web of Science. Future
Internet, [e-journal]12(8), article no. 124. https://doi.org/10.3390/fi12080124.
Mou, X., 2019. Artificial Intelligence, No 32652. World Bank Publications - Reports, The
World Bank Group. [online] Available at: <https://EconPapers.repec.org/RePEc:
wbk:wboper:32652> [Accessed 22 July 2023].
National Institute of Statistics, 2023. Tempo online database. [online]. Available at:
<http://statistici.insse.ro:8077/tempo-online/#/pages/tables/insse-table> [Accessed 30
June 2023].
Identity and Human Skills. Amfiteatru Economic, [e-journal] 23(56), pp. 33-45.
https://doi.org/10.24818/EA/2021/56/33.
Sowa, K., Przegalinska, A. and Ciechanowski, L., 2021. Cobots in knowledge work-AI
collaboration in managerial professions. Journal of Business Research, [e-journal]125,
pp. 135-142. https://doi.org/10.1016/j.jbusres.2020.11.038.
Sullins, J., Craig, S.D. and Hu, X., 2015. Exploring the effectiveness of a novel feedback
mechanism within an intelligent tutoring system. International Journal of Learning
Technology, [e-journal]10(3), pp. 220-236. https://doi.org/10.1504/IJLT.2015.072358.
Tsai, C.-Y., Marshall, J.D., Choudhury, A., Serban, A., Hou, Y.T.-Y., Jung, M.F., Dionne,
S.D. and Yammarino, F.J., 2022. Human-robot collaboration: A multilevel and integrated
leadership framework. The Leadership Quarterly, [e-journal] 33(1), article no. 101594.
https://doi.org/10.1016/j.leaqua.2021.101594.
Ufarte Ruiz, M. J. and Manfredi Sánchez, J. L., 2019. Algorithms and bots applied to
journalism. The case of Narrativa Inteligencia Artificial: structure, production and
informative quality. Doxa Comunicación, [e-journal] 29, pp. 213-233.
https://doi.org/10.31921/doxacom.n29a11.
Van Der Niet, A. and Bleakley, A., 2020. Where Medical Education Meets Artificial
Intelligence: “Does Technology Care?”. Medical Education, [e-journal] 55(1), article no.
30-36. https://doi.org/10.1111/medu.14131.
Wang, Y.-Y. and Wang, Y.-S., 2019. Development and validation of an artificial intelligence
anxiety scale: An initial application in predicting motivated learning behavior. Interactive
Learning Environments, [e-journal] 30(4), pp. 619-634. https://doi.org/10.1080/
10494820.2019. 1674887.
Xiao, M. and Yi, H., 2020. Building an efficient artificial intelligence model for personalized
training in colleges and universities. Computer Applications in Engineering Education,
[e-journal] 29(2), pp. 350-358. https://doi.org/10.1002/cae.22235.
Yam, K.C., Goh, E.-Y., Fehr, R., Lee, R., Soh, H. and Gray, K., 2022. When your boss is a
robot: Workers are more spiteful to robot supervisors that seem more human. Journal of
Experimental Social Psychology, [e-journal] 102, article no. 104360.
https://doi.org/10.1016/j.jesp.2022.104360.
Vol. 26 • No. 65 • February 2024 105

AE Exploring the Antecedents of Artificial Intelligence Products’ Usage.
The Case of Business Students
EXPLORING THE ANTECEDENTS OF ARTIFICIAL INTELLIGENCE

PRODUCTS’ USAGE. THE CASE OF BUSINESS STUDENTS
Rodney Duffett1 , Rodica Milena Zaharia2* , Tudor Edu3 ,
Raluca Constantinescu4 and Costel Negricea5
1)
Cape Peninsula University of Technology, Cape Town, South Africa.
2)4)
Bucharest University of Economic Studies, Bucharest, Romania.
3)5)
Romanian-American University, Bucharest, Romania.

Duffett, R., Zaharia, R.M., Edu, T., Constantinescu, R. Received: 29 September 2023
and Negricea., C., 2024. Exploring the Antecedents of Revised: 22 November 2023
Artificial Intelligence Products’ Usage. The Case of Accepted: 20 December 2023
Business Students. Amfiteatru Economic, 26(65),
pp. 106-125.
Abstract
This study aims to investigate to what extent facilitating conditions (those means that users
consider necessary to use for a certain technology) and other predictors (perceived risk and
lack of trust in technology, gender, education, income, technology proficiency and
equipment used to access the Internet) influence the use of Artificial Intelligence Products
(AIP) in general and for education. Using the Unified Theory of Acceptance and Use of
Technology (UTAUT), the data collected, through an online questionnaire from a sample of
450 Romanian business students, were examined using principal component analysis
(PCA) and logistic regression. Facilitating conditions indicated a direct effect (positive) on
the dependent variables, and the combination between perceived risk and perceived lack of
trust in technology displayed an opposite effect (negative) on the dependent variables.
Female students showed a greater tendency to use AIPs in general and for education.
Undergraduate students were more inclined to use AIPs in general. Students not using
smartwatches or personal computers are inclined to use AIPs more in general and for
education. This study advances the theory by exploring the actual use of AIPs for
educational purposes, developing the UTAUT model by isolating facilitating conditions
and using descriptive variables as predictors. At the same time, the present research
contributes to enriching the empirical evidence related to UTAUT on the acceptance and
use of technology in Romania. The results of the research allow for the formulation of
practical recommendations for universities as current and potential providers of AIPs in
order to make the educational process more efficient.
*
Corresponding author, Rodica Milena Zaharia – email: [email protected]

Keywords: Artificial Intelligence (AI); Artificial Intelligence Product (AIP); business

students; Unified Theory of Acceptance and Use of Technology (UTAUT); principal
component analysis (PCA); logistic regression, Romania.
JEL Classification: I23, M31, C51, C53
Introduction
Artificial intelligence (AI) has become a comprehensive and ubiquitous phenomenon, with
its widespreadness acknowledged at individual and organisational levels, but also at the
level of the European Union. Defining AI is a daunting task. The European Parliament (EP)
(2020) attempted to provide a broad definition of AI by focussing on the transfer of human
features and skills, such as thinking, accumulating knowledge, making decisions, or
creating. AI consists of AI software (covering instruments delivering search results and
recommendations, or performing voice and image analysis) and object-orientated AI,
referring to objects that collect data, such as home appliances, cars, robots, or drones
(European Parliament, 2020).
The educational process is under high pressure to incorporate digital culture from the early
stages of the training. The incorporation of AI resources into the process of education
(course content, assessment instruments, and smart learning tools), and the usage of
different digital objects (e.g., smart devices) in education (Chassignol et al., 2018) will
influence the future of why, how, and what students will learn (Escotet, 2023). In the near
future, probably everyone will need to have a basic understanding of the technologies that
support AI (Touretzky et al., 2019).
The process of integration of AI into education is increasing and diversifying. Studies
display a variety of investigations related to the role of AI in improving the educational
process, increasing student satisfaction and retention (Rico-Bautista et al., 2021), or
providing invaluable support to learners by creating the infrastructure that makes possible
flexible learning and self-education (Alzahrani, 2023). However, Raffaghelli et al. (2022)
argue that the impact of AI on education should be carefully evaluated to avoid extremely
optimistic or extremely pessimistic conclusions and recommend that student perceptions of
the use of technology should be explored to allow higher education institutions to
understand how to adapt their processes. Chatterjee and Bhattacharjee (2020)
recommended that future studies on AI in higher education include 'actual use' as a
dependent variable, while data should be collected from 'actual adopters' of AI.
This article looks to assess the prerequisites of AI usage in general, and for educational
purposes in the case of business students from Romanian universities, by employing the
Unified Theory of Acceptance and Use of Technology (UTAUT) framework. UTAUT was
intended to have a unitary framework, both theoretically and empirically, to anticipate
human behaviour related to the acceptance and use of technology (Venkatesh et al., 2003).
Thus, UTAUT, in addition to helping to understand the complex phenomenon of
acceptance and use of technology (by identifying the influence factors: performance
expectation, effort expectation, social influence, and facilitation conditions), also offers the
possibility of empirical testing. In education, UTAUT is increasingly used to determine the
most important factors in accepting and using technology, given the fundamental role that
education plays in the behavioural modelling of future leaders, entrepreneurs, and
Vol. 26 • No. 65 • February 2024 107

employees (Chatterjee and Bhattacharjee, 2020; Gansser and Reich, 2021; Raffaghelli et
al., 2022). By identifying the importance of factors of influence in the acceptance and use
of technology, educational institutions, in general, and higher education, in particular, could
develop effective strategies for the use of technology and its adaptation to the behaviour of
users (both students and professors).
Focus on Generation Z is the most appropriate, as this generation is considered to be the
most open to adopting AI, and studies focussing on their usage behaviour are scarce (Kelly,
Kaye and Oviedo-Trespalacios, 2023). The UTAUT is one of the most commonly used
models in research concerning intention, acceptance and actual usage of new technologies,
due to its high explanatory power (Gansser and Reich, 2021; Rico-Bautista et al., 2021).
The article aims to investigate the potential effects of facilitating conditions (the means
users would consider necessary to exist in order to be able to use the technology), perceived
risk and perceived trust on the usage of AI in general and in education in the case of
business students. This research is innovative both theoretically and practically. First of all,
according to the experience and knowledge of the authors, there is no such study that has
applied UTAUT to Romanian business students. The present study extends UTAUT by
including perceived risk and the negative vision of trust in technology/AIP in the form of
perceived lack of trust, as well as by transforming gender and experience with technology
into independent variables and also including three other descriptive variables as
explanatory variables.
From a practical perspective, the study is added to the empirical research in the field.
Business students are one of the biggest categories of students in Romania. They will be
tomorrow's business leaders, entrepreneurs, or employees, all acquainted with the use of
Artificial Intelligence Products (AIPs). The results of the study, although qualitatively
valuable (the findings could not be generalised to all business students in Romania) suggest
that universities should facilitate the adoption of AIPs for educational purposes, through a
stronger integration in all phases of the educational process: teaching, learning, teacher-
student interaction. Thus, university strategies can be adapted to the evolution of AIPs in
order to facilitate education.
The article is structured as follows: A literature review of studies supporting the role of AI
in shaping new capabilities and skills for students and in acknowledging the limits and risks
associated with the use of new technology, as well as a presentation of the UTAUT
framework, is provided to support the research hypotheses. The article continues with
research methodology, where the applied model, sampling, data collection instrument and
process are explained. The data analysis and results follow, continuing with discussion, and
the article ends with conclusions.
1. Literature review and the research hypotheses

1.1. Acceptance and use of AI technology and Artificial Intelligence Products (AIPs) in
higher education: advantages and limits
AI is dubbed an instrument meant to improve the lives and working conditions of people
from all walks of life (Gansser and Reich, 2021). For education, AI provides adaptive
learning content tailored to the specificities of various groups of students, going down even
to the individual level (Chatterjee and Bhattacharjee, 2020). AI is considered to have a

transformational impact on collaborative education, management of teaching and learning,

educational infrastructure (Papaspyridis, 2020), aiding students to scrutinise large quantities
of data, and helping them to make forecasts (Hasan, Shams and Rahman, 2021).
Universities recognise the competitive advantage that comes from the implementation of AI
in the educational process and transform themselves into smart universities, merging classic
educational content with new technology to deliver collaborative education and an adaptive
learning experience (Rico-Bautista et al., 2021). Schmitt (2019) supports the idea that
equipment and information cannot exist separately in the current reality. Investigating AI
technology-related learning behaviour, Wang et al. (2022) termed the applications used to
collect and manage data connected to terminals and storage equipment as AI
techniques/products. Therefore, the current study uses their angle and calls them Artificial
Intelligence products (AIPs).
It is obvious that the implementation of digital technology in the educational process
involves awareness of the advantages, as well as the limits and risks of the proliferation of
AIPs, from the technical limitations and ethical implications of these AIPs (Park et al.,
2019) to data security and privacy. Despite the large body of literature developed in recent
years, there is a lack of a systematic and rigorous framework for assessing risk and
uncertainty of the diffusion of AI in education (Zhang et al., 2022, p.1). The focus is
moving from the question of AI replacing humans to the question of how humans can work
alongside AI, how we, as humans, can use AIPs and cooperate with AI in order to improve
our welfare, and how the risks associated with the AI output are mitigated. These aspects
are still less studied and understood (Markauskaite et al., 2022).
AI education provides a wide variety of advantages in the areas of self-assessment,
formulation of comprehensive questions and answers, conflict management and decision-
making (Roy et al., 2022). In their study, Markauskaite et al. (2022, p.4) identify the
advantages brought by AI outputs, as well as the limitations and risks associated with them.
There are four sources of advantages, but also limitations for the AI era and valid for
students as well (Markauskaite et al., 2022). The first source looks at how to interpret and
understand AI outputs. This refers to the capacity of users to understand the rapid changes
of the technology, how to be prepared to internalise it in ordinary usage, and how to
navigate among the multitude of applications or instruments offered by the new technology
in order to capture the best of AI outputs. It also involves being aware of the risks related to
the use of AI coming both from data-related risks (e.g. biased data or out of domain data)
and from model-related risks (e.g. model misspecification) (Zhang et al., 2022) and
continuing with the interpretation of the AI outputs in a proper way. The second source of
advantages and limitations looks to integrate AI and AIPs outputs into human knowledge
systems. The educational process needs to emphasise that it is the human that has the final
decision about integrating AI outputs into a decision, and it is the human that has to decide
how AI objects are used, and how these results need to be interpreted and applied. As
studies mention, “AI is always understood within the contexts of its use” (Bearman and
Ajjawi, 2023, p.16), and students must be aware that the purpose of AI must be meaningful
for society. The third source of advantages, but especially challenges, looks to assess and
evaluate the ethical implications of AI. The ethics of AI are one of the most important
debated issues and one of the most difficult areas to address. AI can provide much wider
access to knowledge, at significantly lower costs, for wider categories of people, in much
less time (Huang et al., 2023; Pisică et al., 2023). However, the ethical issues that AI
generates are by far the most controversial. In 2021, the Pew Research Centre addressed
Vol. 26 • No. 65 • February 2024 109

this topic and underlined these questions summarising the difficulties pertaining to ethics in
AI: “How can ethical standards be defined and applied for a global, cross-cultural, ever-
evolving, ever-expanding universe of diverse black-box systems in which bad actors and
misinformation thrive?”. In addition, teaching AI ethics is a big concern. It is important not
only to teach about AI ethics, but also how to teach ethics related to AI (Burton et al., 2017;
Zeide, 2019; Bearman and Ajjawi, 2023). Finally, the fourth source of advantages and
limitations refers to how to elevate human cognitive work to creativity and meaning/sense-
making domains. It involves efforts from both students and educational institutions.
Students must be prepared for changes in curricula and in the way curricula are delivered,
including the use of AI objects for educational purposes). Universities must be managed
and resourced for a different learning environment for students and teachers, which
includes technological endowments, knowledge for the use of AI, and specific content and
delivery strategies for an AI-infused educational process) (Nurjanah and Pratama, 2020;
Rivers, Nakamura and Vallance, 2022).
1.2 The Unified Theory of Acceptance and Use of Technology (UTAUT) framework
The Unified Theory of Acceptance and Use of Technology (UTAUT) is one of the most
commonly used models in research concerning the intention and actual usage of new
technologies (Gansser and Reich, 2021). The UTAUT model consists of three direct
antecedents of the intention to use technology, namely, performance expectation, effort
expectation and social influence, and two direct prerequisites of usage behaviour,
specifically intention and facilitating conditions. The model also makes use of four
moderators, particularly experience, voluntariness, gender, and age (Venkatesh et al., 2003).
The UTAUT model has been used in AI research in its original format or with additional or
different variables. Chatterjee and Bhattacharjee (2020), in one of the few studies
investigating AI usage, employed the UTAUT model to determine the behavioural intention
and adoption of AI in higher education by probing students, faculty and administrative
staff. They included perceived risk and attitude, but they dropped social influence.
Alzahrani (2023), aiming to determine the attitudes and behavioural intentions of students
to use AI, adapted the UTAUT model by including perceived risk, awareness and attitude,
and removing social influence. Raffaghelli et al. (2022) studied the acceptance of early
warning systems (AI tools) by applying a longitudinal approach (pre-use versus post-use)
and including trust in the model alongside the four original predictors.
Among the four predictive variables of the UTAUT model, facilitating conditions represent
the only category introduced to explain usage, while the other three predictors are used to
describe intention (Wu et al., 2022). Kelly, Kaye and Oviedo-Trespalacios (2023)
concluded that only a small number of studies focused on studying actual use behaviour. In
addition, Chatterjee and Bhattacharjee (2020) recommended that future studies on AI in
higher education include actual use as a dependent variable, while data should be collected
from actual adopters of AI. Corroborating these perspectives, the present study aims to
explore the actual usage of AI in general and in education by students and, thus, preserves
the facilitating conditions from the UTAUT model and removes the antecedents that
describe intention, namely performance expectancy, effort expectancy, and social influence.
As proposed by Venkatesh et al. (2003), the UTAUT model includes four moderators:
experience, voluntariness, gender, and age. The literature prompts situations in which the
moderators were excluded from the model (Chatterjee and Bhattacharjee, 2020; Raffaghelli

et al., 2022; Wu et al., 2022; Alzahrani et al., 2023). This study aims to collect data from
students who are actual users of AI. For this purpose, age and voluntariness cannot provide
value to the model, but this study aims to develop the UTAUT model by employing gender
and experience as explanatory variables instead of moderators and including another three
descriptive and behavioural variables as independent variables: education, income, and
equipment used for Internet access.
Additionally, perceived privacy risk and trust are intricately connected to personal data
sharing, a feature of AI objects, and are, thus, deemed important prerequisites of consumer
adoption and usage of technology (Hasan, Shams and Rahman, 2021). Gansser and Reich
(2021) argue that perceived risk associated with the use of AI and data processing is an
important prerequisite of usage behaviour. In higher education, perceived risk is closely
connected to the negative feelings users may have about AI (Chatterjee and Bhattacharjee,
2020), and thus a negative relationship is expected between perceived risk and user
behaviour. The existing literature on AI shows evidence of the inclusion of perceived risk
as an independent variable in the UTAUT model (Chatterjee and Bhattacharjee, 2020; Wu
et al., 2022; Alzahrani, 2023). Alzahrani (2023) considers that perceived risk improves the
explanatory power of the UTAUT model, especially in the case of AI. Trust is another
important predictor of AI usage. Leichtmann et al. (2023) suggested that trust in AI builds
over time as users better understand the technology. As users’ understanding of AI
improves, expectations can be developed realistically and thus users can become satisfied
with the experience and trust in AIPs can be acquired (Kamila and Jasrotia, 2023). Thus, a
positive relationship is expected between trust and AIPs user behaviour. Another research
stream investigates lack of trust, providing an opposite view (Du and Xie, 2021; Thiebes,
Lins and Sunyaev, 2021), as a negative connection is documented with AI or technology
behavioural outcome. Rico-Bautista et al. (2021) recommend that trust be included as an
explanatory variable of AI usage behaviour in education.
1.3 Research hypotheses

Venkatesh et al. (2003) explained that facilitating conditions would refer to the means users
would consider necessary for using technology. Facilitating conditions have been analysed
as an explanatory variable for the implementation of AI in education. Strzelecki (2023)
determined that the facilitating conditions had a significant positive influence on the usage
behaviour of ChatGPT by students. Chatterjee and Bhattacharjee (2020) discovered a
significant positive effect of facilitating conditions on the intention of behaviour and the
expectancy of effort in the case of adopting AI in higher education.
As facilitating conditions represent the only predictor of technology usage in the original
version of the UTAUT model (Venkatesh et al., 2003) and the aim of this study is to
explain the usage of AIP in general and in education, facilitating conditions could represent
a strong predictor. Therefore, the first two hypotheses are as follows: H1. Facilitating
conditions have a direct (positive) influence on the usage of AIPs in general by students;
H2. Facilitating conditions have a direct (positive) influence on the use of AIPs for
educational purposes by students.
The risk perceived by users when using AIPs is considered an important antecedent of
behavioural intention or actual usage of the respective technology, with a high perceived
risk being deemed a significant barrier to technology adoption (Gansser and Reich, 2021).
Chatterjee and Bhattacharjee (2020) discovered a significant negative effect of perceived
Vol. 26 • No. 65 • February 2024 111

risk on the attitude of higher education stakeholders towards AI adoption. Alzahrani (2023)
supports the idea that perceived risk should be investigated in the context of AI in
education, as it strengthens the explanatory power of the UTAUT model. Considering the
already established impact of perceived risk in AI in general and in education, and the call
for further exploration of perceived risk in the UTAUT model, the next hypotheses were
developed: H3. Perceived risk has an inverse (negative) effect on the usage of AIPs in
general by students; H4. Perceived risk has an inverse (negative) effect on the usage of
AIPs for educational purposes by students.
Trust in AI can be attained through consistent service delivered by the applications (Kamila
and Jasrotia, 2023), but it has to be built on transparency, trustworthiness and
accountability (Robinson, 2020), as lack of trust is an important obstacle in the adoption of
AI (Du and Xie, 2021; Thiebes, Lins and Sunyaev, 2021). Kelly, Kaye and Oviedo-
Trespalacios (2023) concluded that trust positively predicted the use behaviour of AIPs in
the context of applying the UTAUT model. The trust of students in AI is essential to build
credibility in AI outputs, with studies identifying a direct positive effect of perceived trust
on the intention to adopt AI (Pillai et al., 2023). Considering the identified influence of trust
in AI and the dual approach (trust vs. lack of trust), the present study focusses on exploring
the impact of perceived lack of trust on AIP usage in general and in education by students,
therefore prompting hypotheses: H5. Perceived lack of trust has an inverse (negative) effect
on the use of AIPs in general by students; and H6. Perceived lack of trust has an inverse
(negative) effect on the usage of AIPs for educational purposes by students.
Descriptive variables have been employed in studies on AI and/or using the UTAUT model
as moderators, beginning with Venkatesh et al. (2003), who used in this regard gender, age,
experience and voluntariness, but, especially, to describe the profile of the respondent. This
study aims to expand the role of descriptive variables, particularly gender, education,
income, technology proficiency and equipment used for Internet access, from profile
descriptors or moderators to predictors.
Gender has been employed in previous studies in technology-related education and AI
studies for profiling purposes (Shaya, Madani and Mohebi, 2023) and as a moderator
(Khechine and Lakhal, 2018). However, there is evidence on the impact of gender on
various outcomes related to AI in education. Lin et al. (2021), for example, determined that
women had a lower tendency to learn AI compared to male students. On the other hand,
there is a growing interest in IT education among women considering course selection,
dropout rates, or IT tool usage in education (Dziuban and Moskal, 2001; Wrycza,
Marcinkowski and Gajda, 2017). This study considers that gender could shed light on the
use of AIPs in general and for educational purposes, and thus posits the following
hypotheses: H7. Male students are more inclined to use AIPs in general than female
students; and H8. Male students are more inclined to use AIPs for educational purposes
than female students.
Educational level has been used in research on technology implemented in education and
AI, especially to describe the respondents and the research context. The findings of studies
investigating the effects of educational level on technology and/or AI use for learning
purposes are divergent, as some studies prompt a clear effect (Matzavela and Alepis, 2021;
Rodriguez-Hernandez et al., 2021), while others show no effect at all (Wang et al., 2023).
This study aims to employ educational level as an explanatory variable for the usage of
AIPs in general, and for educational purposes, and therefore postulates two hypotheses: H9.

Postgraduate students are more inclined to use AIPs in general than undergraduate students;
and H10. Postgraduate students are more inclined to use AIPs for educational purposes than
undergraduate students.
Income level has been employed to a lesser extent in studies on technology-related education
and AI, in addition to for the description of the respondents (Shaya, Madani and Mohebi,
2023). Exceptions are the studies of Rodriguez-Hernandez et al. (2021), which used monthly
family income to explain academic achievements, and of Matzavela and Alepis (2021),
which employed parent income to describe the academic performance of students. As there
is a lack of evidence on the impact of income level on the usage of AIPs in general and for
educational purposes, this study aims to investigate its impact and, hence, suggests the
following hypotheses: H11. Students with a higher income are more inclined to use AIPs in
general than students with a lower income; and H12. Students with a higher income are more
inclined to use AIPs for educational purposes than students with a lower income.
Experience with technology has been used in studies exploring the impact of technology in
education to describe the sample and as a moderator. Shaya, Madani and Mohebi (2023)
used the experience with the usage of m-learning apps (beginner, intermediate, and expert)
to describe the user of mobile learning apps. Khechine and Lackhal (2018) employed
experience in using computers, measured in years, as a moderator in an endeavour to assess
the influence of facilitating conditions on the usage behaviour of a webinar system by
students, the analysis rendering a nonsignificant result. This study aims to evaluate the
impact of technology proficiency on the usage of AIPs in general and for educational
purposes, and thus formulates the following two hypotheses: H13. Students that are more
technologically proficient are more inclined to use AIPs in general than students who are
less technologically proficient; and H14. Students that are more technologically proficient
are more inclined to use AIPs for educational purposes than students who are less
technologically proficient.
The equipment used for Internet access has been employed in previous research on the
usage of technology in education and the usage of AIPs to not only profile the respondents,
but also explain behavioural aspects. Alshammari (2021) collected data about devices used
by teaching staff in a study on the use of virtual classrooms, identifying that PCs, laptops,
smartphones and tablets were used by academics. Gansser and Reich (2021) argued that
smartwatches were considered devices that were extensively used for convenience. The
present study focusses on analysing the impact of the equipment used for Internet access on
the usage of AIPs in general and for educational purposes, and therefore prompts two
hypotheses: H15. Students who use particular equipment are more inclined to use AIPs in
general than students who do not use that particular equipment: H15a. smartphone; H15b.
laptop; H15c. tablet; H15d. smartwatch; H15e. personal computers; and H16. Students who
use a particular equipment are more inclined to use AIPs for educational purposes than
students who do not use that particular equipment: H16a. smartphone; H16b. laptop; H16c.
tablet; H16d. smartwatch; H16e. personal computer.
The theoretical model proposed in this study is shown in Figure no. 1.
Vol. 26 • No. 65 • February 2024 113

Figure no. 1. The theoretical model
This study considered the UTAUT model proposed by Venkatesh et al. (2003) and its
extension employed by Arfi et al. (2021), adapting them for business students already using
AI. Thus, three groups of variables were adapted from the two studies, specifically
facilitating conditions of usage, perceived lack of trust and perceived risk, to determine
their impact on the usage of AIPs in general and for educational purposes. Additionally,
this study expanded the UTAUT model by transforming gender and experience (measured
in this study through technology proficiency) from moderators into independent variables
and by including other descriptive variables as independent variables, such as education,
income, and equipment used for Internet access.
The sample was built through a nonprobability sampling method, the online questionnaire
used to collect data (Pillai et al., 2023) being disseminated among Romanian business
students. Data collection was carried out between May and June 2023, focussing on
obtaining data from students of both genders enrolled in undergraduate and post-graduate
academic programmes. 474 responses were collected. After reviewing them for
completeness, 450 questionnaires were retained for analysis, the sample size corresponding
to a 95% confidence interval and a margin of error of 4.62%. The sample size matches or
exceeds the sample sizes used in previous studies on AI (Chatterjee and Bhattacharjee,
2020; Raffaghelli et al., 2022).

The questionnaire started with the definition of AI and included the seven areas of
implementation of the European Parliament (2020). The questionnaire included 9 Likert
scales with 5 levels (Wu et al., 2022) for assessing facilitating conditions (Venkatesh et al.,
2003), perceived risk and perceived lack of trust (Arfi et al., 2021) (Table no. 1), two
dichotomous scales for the dependent variables and categorical scales for the descriptive
variables: gender (Raffaghelli et al., 2022), education and technology proficiency (Rodriguez-
Hernandez et al., 2021; Shaya, Madani and Mohebi, 2023), income (Matzavela and Alepis,
2021) (Table no. 1), using as benchmarks the net values of the minimum and average salary
in Romania at the beginning of 2023, and the equipment used for Internet access
(Alshammari, 2021; Gansser and Reich, 2021). Before uploading the questionnaire to Google
Forms for dissemination, 12 students enrolled in undergraduate and postgraduate academic
programmes, in equal shares (adapted from Zaharia et al., 2022), with four items modified.
In this study, two methods of data analysis were used. First, principal component analysis
(PCA) was used to group the items used into factors (Field, 2009). In a second phase,
logistics regression was used to explain the two investigated dependent variables, the use of
AIPs in general and the use of AIPs for educational purposes, according to the factors
rendered by the performed PCA and the descriptive variables: gender, education, income,
technology proficiency and equipment used for Internet access. Logistic regression is very
useful when it is intended to determine the probability that one case is distributed in one
category over another or other categories, the method presenting two advantages, the first
being that the dependent variable should not comply to a normal distribution, and the
second being that the model can also include categorical predictors (Wuensch and Poteat,
1998; Wuensch, 2014).
3. Data analysis and results

Data related to the descriptive variables are presented in Table no. 1. Some aspects stand
out. 4.00% of the respondents preferred not to respond about gender. 72.4% of the
respondents were registered in postgraduate study programmes. All respondents mentioned
that they access the Internet via the smartphone.
Table no. 1. Descriptive data of the sample
Variable Category Frequency Percentage Variable Category Frequency Percentage
Men 252 56.0 Do you Yes 450 100.0
access the
Internet
Women 180 40.0 using your No 0 0
Gender
smartphone?
I prefer not to
18 4.0 Do you Yes 420 93.3
respond access the
Enrolled in Internet
undergraduate 124 27.6 through No 30 6.7
programme your laptop?
Education
Enrolled in
undergraduate 326 72.4 Do you Yes 258 57.3
programme access the
Internet
Less than or through a
Income equal to 1900 192 42.7 tablet? No 192 42.7
RON
Vol. 26 • No. 65 • February 2024 115

Variable Category Frequency Percentage Variable Category Frequency Percentage

1901-3565 Do you
234 52.0 Yes 216 48.0
RON access the
Internet
3566 RON
24 5.3 through a No 234 52.0
and above smartwatch?
Rather yes 270 60.0 Do you Yes 252 56.0
access the
Technologically Internet
proficient Rather not 180 40.0 through a No 198 44.0
personal
computer?
Assessing the nine scales used in the questionnaire (Table no. 2) in a PCA using a Varimax
rotation based on factor loadings of a minimum of 0.40, Eigenvalues higher than 1 and a
Scree plot (Field, 2009), and a Cronbach Alpha higher than 0.7 (Nunnally, 1978), three
factors were delineated. However, one factor included only one item. When forcing the
analysis with only two factors, the item AIPs provide a secure work environment where
personal data is transmitted securely showed a low communality, thus displaying no
relations with the rest of the items of none of the two factors, being eliminated from the
analysis according to the recommendations of Osborne and Costello (2009). In conclusion,
two factors were retained (Table no. 2). The procedure was repeated by using Principal
Axis Factoring as suggested by Osborne and Costello (2009), as this extraction method
displays the covariance of items, the conclusion being the same regarding the two factors.
The two factors were named F1 (Risk and lack of trust associated with AIPs) and F2
(Facilitating conditions for AIPs).
Table no. 2. Principal Component Analysis - items covering user experience with AIPs
Rotated Component Matrix a
Component
1 2
I have the resources to use AIPs .834
I have the knowledge to use AIPs .874
I have access to the support I need to use AIPs .878
It is risky to use AIPs .821
Using AIPs comes with more risks than benefits .816
It is a mistake to use AIPs .874
I am afraid to use AIPs because of the possibility of losing personal data and privacy .829
AIPs provide a secure work environment where personal data are transmitted
securely
I find it risky to disclose personal data and information to use AIPs .742
Cronbach Alpha .884 .857
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalisation.
Note A. Rotation converged in 3 iterations.
Note B: Kaiser-Meyer-Olkin Measure of Sampling Adequacy: 0.744; Bartlett’s Test of Sphericity-
Sig.: 0.00
Note C: Cronbach Alpha: >0.70 for all factors and between factors.
In a second step, the two factors were regressed together with gender, education, income,
technology proficiency and equipment used for Internet access (smartphone, laptop, tablet,

personal computer, smartwatch) against the usage of AIPs in general and the usage of AIPs
for educational purposes. The evaluation of the use of each equipment to access the Internet
was performed on a different scale (Table no. 1). Smartphone usage was not introduced in
the regression models because this option was selected by all respondents, thus rendering
zero non-users. About the performed logistical regression, using a stepwise approach and a
non-significant value of the Hosmer and Lemeshow goodness-of-fit test (Wuensch, 2014),
the model with the most significant independent variables, explaining the use of AIPs in
general, initially included six variables, namely F1, F2, gender, education, smartwatch use
for accessing the Internet and technology proficiency. However, when checking the data
against the logistic regression assumptions, technology proficiency was significantly highly
correlated with gender, so it was removed from the model in order to meet the assumptions
(Zaharia et al., 2022). The final model for explaining the usage of AIPs in general is
displayed in Table no. 3.
Table no. 3. Logistic regression- the usage of AIPs in general as dependent variable
Variables in the Equation
95% C.I.
for EXP(B)
B S.E. Wald df Sig. Exp(B) Lower Upper
F1 (Risk and lack of trust associated
with AIPs) -.348 .134 6.730 1 .009 .706 .543 .919
F2 (Facilitating conditions for AIPs) .560 .133 17.657 1 .000 1.751 1.348 2.273
Gender (Prefer not to respond) 55.787 2 .000
(Women) 3.742 .534 49.149 1 .000 42.174 14.816 120.048
(Men) 2.617 .544 23.112 1 .000 13.690 4.711 39.785
Education (Undergraduate vs.
Postgraduate) -1.247 .465 7.192 1 .007 .287 .115 .715
Equipment used for Internet access
(Smartwatch- No vs. Yes) -.740 .270 7.491 1 .006 .477 .281 .811
Note A: Hosmer and Lemeshow Test- non-significant value (p>0.05)-adequate level of data fitting;
Chi-square = 202.892 (p<0.001); Nagelkerke R Square = 0.484; correctly classifying 80% of the
cases; Note B- Logistic regression assumptions met (according to Zaharia et al., 2022)
Variables not included in the equation
Score df Sig.
due to a significance level below .05
Income (less than or equal to 1900 RON) 5.405 2 .067
Income (1901-3565 RON) .633 1 .426
Income (3566 and above) 3.346 1 .067
Equipment used for Internet access (Laptop)
(No vs. Yes) .465 1 .495
Equipment used for Internet access (Tablet)
(No vs. Yes) .538 1 .463
(Personal computer) (No vs. Yes) .012 1 .914
The model includes five significant variables (Wald tests, p<0.001 for the second and third
variables and p<0.01 for the first, fourth and fifth variables) (Table no. 2). The impact of
each variable on the usage of AIPs in general is explained based on the odds ratio. F1 (Risk
and lack of trust associated with AIPs), with an odds ratio of 0.706, displays that a decrease
of one unit on the measurement scale of the predictor increases the odds of using AIPs in
general by a multiplicative factor of 1.416. F2 (Facilitating conditions for AIPs), with an
odds ratio of 1.751, shows that an increase of one unit on the measurement scale of the
Vol. 26 • No. 65 • February 2024 117

predictor increases the odds of using AIPs in general by a multiplicative factor of 1.751.
Gender (significant overall at p<0.001), with odds ratios of 42.174 (p<0.001) and 13.690
(p<0.001), displays that female students are 42.174, while male students are 13.690 times
more inclined to use AIPs in general than people who preferred not to respond to their
gender. Education, with an odds ratio of 0.287, shows that students enrolled in
undergraduate programmes are 3.484 times more inclined to use AIPs in general than those
enrolled in postgraduate programmes. The equipment used for Internet access
(Smartwatch), with an odds ratio of 0.477, shows that students not using this equipment to
access the Internet are 2.096 times more inclined to use AIPs in general than those using
smart watches.
The most comprehensive model explaining the usage of AIPs for educational purposes
includes four variables, namely F1, F2 gender, and the use of personal computers to access
the Internet. These variables are retained after checking the data against the logistic
regression assumptions (Zaharia et al., 2022). The final model for explaining the usage of
AIPs for educational purposes is shown in Table no. 4.
Table no. 4. Logistic regression - the usage of AIPs for educational purposes
as dependent variable
Variables in the Equation
95% C.I.
for EXP(B)
B S.E. Wald df Sig. Exp(B) Lower Upper
F1 (Risk and lack of trust associated
with AIPs) -.498 .124 16.048 1 .000 .608 .477 .776
F2 (Facilitating conditions for AIPs) .745 .127 34.405 1 .000 2.106 1.642 2.700
Gender (Prefer not to respond) 67.850 2 .000
(Women) 1.823 .225 65.717 1 .000 6.188 3.982 9.614
(Men) 1.252 .250 25.169 1 .000 3.498 2.145 5.704
(Personal computer- No vs. Yes) -.494 .254 3.789 1 .052 .610 .371 1.003
Note A: Hosmer and Lemeshow Test- non-significant value (p>0.05)-adequate level of data fitting;
Chi-square = 189.387 (p<0.001); Nagelkerke R Square = 0.458; correctly classifying 81.3% of the
cases; Note B- Logistic regression assumptions met (according to Zaharia et al., 2022)
Variables not included in the equation
Score df Sig.
due to a significance level below .05
Income (less than or equal to 1900 RON) 12.712 2 .302
Income (1901-3565 RON) 12.709 1 .302
Income (3566 and above) 1.415 1 .234
Education (Undergraduate vs. Postgraduate) 2.892 1 .089
Technology proficiency (Rather technologically
proficient vs. Rather not technologically
proficient) .672 1 .412
Equipment used for Internet access (Smartwatch)
(No vs. Yes) .115 1 .735
Equipment used for Internet access (Laptop)
(No vs. Yes) 3.841 1 .150
Equipment used for Internet access (Tablet)
(No vs. Yes) 0.484 1 0.486

The model includes four significant variables (Wald tests, p<0.001 for the first three
variables and p<0.1 for the fourth variables) (Table no. 3). The impact of each variable on
the usage of AIPs for educational purposes is explained based on odds ratio. F1 (Risk and
lack of trust associated with AIPs), with an odds ratio of 0.608, displays that a decrease of
one unit on the measurement scale of the predictor increases the odds of the usage of AIPs
for educational purposes by a multiplicative factor of 1.645. F2 (Facilitating conditions for
AIPs), with an odds ratio of 2.106, shows that an increase of one unit on the measurement
scale of the predictor increases the odds of the usage of AIPs for educational purposes by a
multiplicative factor of 2.106. Gender (significant overall at p<0.001), with odds ratios of
6.188 (p<0.001) and 3.498 (p<0.001), displays that female students are 6.188, while male
students are 3.498 times more inclined to use AIPs for educational purposes than people
who preferred not to respond regarding their gender. The equipment used for Internet
access (Personal computer), with an odds ratio of 0.610, shows that students not using
personal computers are 1.639 times more inclined to use AIPs for educational purposes
than those using personal computers.
In conclusion, hypotheses H1 to H6 were supported, while hypotheses H7 to H16 were not
supported by the findings. Furthermore, the two dependent variables, using AIPs in general
and AIPs for educational purposes, are strongly correlated based on Phi coefficient value
(0.789, p<0.001). Furthermore, the other can be explained in a similar proportion based on
Lambda coefficients (0.684, p<0.001).
4. Discussion
In both situations, the use of AIPs in general and the use of AIPs for educational purposes
show that facilitating conditions are a stronger predictor compared to the risk and lack of
trust associated with AIPs. The significant positive effect of facilitating conditions is in line
with previous works in AI adoption in general (Gansser and Reich, 2021) and in higher
education (Chatterjee and Bhattacharjee, 2020) or ChatGPT usage by students (Strzelecki,
2023), and extends the existing literature on their stimulating impact on AI acceptance by
consolidating the actual usage perspective. Thus, universities should focus their strategies
on displaying the resources linked to the infrastructure and on the support offered to
efficiently make use of AI and on building knowledge on how to use it.
The risk and lack of trust associated with AIPs have a significant influence on the usage of
AIPs in general and for educational purposes. The PCA performed in this research grouped
perceived risk and perceived lack of trust, offering a new angle to explore and understand
AI usage. The significant negative effect of this variable on AI usage in both research
streams is in sync with previous works on this issue (Chatterjee and Bhattacharjee, 2020;
Du and Xie, 2021; Gansser and Reich, 2021; Thiebes, Lins and Sunyaev, 2021).
Universities must respond by better communicating the benefits of AI, thoroughly
explaining the risks associated with them, and focussing on safety by clarifying personal
data and privacy issues.
Gender turned out to be a significant predictor in both cases. However, it is interesting that
female students turned out to be more inclined to use AIPs than male students, with the
reference base in the performed logistic regressions being those who did not want to declare
their gender. The finding contradicts even the view of Venkatesh et al. (2003) that supports
that men were more inclined towards technology. Furthermore, the result disproves the
Vol. 26 • No. 65 • February 2024 119

findings of Lin et al. (2021), who concluded that male students were more inclined to learn
AI. However, the result of this study expands on AI the views of Dziuban and Moskal
(2001) and Wrycza, Marcinkowski and Gajda (2017) who point to a growing interest
among women in IT education and in technology, in general. An explanation for this result
could be that AIPs are becoming more user-friendly, and thus more accessible. Moreover,
many AIPs, in general and more clearly for educational purposes, target both genders.
Education was a significant predictor in the case of the use of AIPs in general.
Undergraduate students showed a higher tendency to use AIPs in general compared to those
enrolled in postgraduate programmes. This result could be explained based on the
perspective that students enrolled in undergraduate programmes find themselves in a
knowledge-building stage, thus, looking for ways and methods to acquire knowledge, while
the others are in a refinement phase in which people try to specialise in various fields. The
findings of this study support the divergence found in the literature, with the significant
effect found on AIP usage in general that matches the conclusions of studies that determined
the impact of education on technology and/or AI use for learning purposes (Matzavela and
Alepis, 2021; Rodriguez-Hernandez et al., 2021), while the no-effect result on AIP use for
education converging with the conclusions of studies that found no effect of education on AI
applications used in teaching (Wang et al., 2023). However, through the results obtained,
this study highlights that the academic level deserves to be taken into consideration as a
predictor in analyses related to understanding the behavioural aspects of AI and technology.
Two types of equipment used for Internet access turned out to be significant predictors,
interestingly, both displaying an inverse relationship with the dependent variables. Thus,
people not using smartwatches tend to use AIPs in general more, while those not using
personal computers show a higher tendency to use AIPs for education. In the first instance,
an explanation could come from the fact that the screen of smartwatches is too small for a
proper experience, although they are deemed highly convenient devices (Gansser and
Reich, 2021), while in the second instance, individuals found in this age bracket tend to like
to access the Internet through mobile devices, such as smartphones, laptops, or tablets
(Petrosyan, 2023; Taylor, 2023). The results are even more interesting as tablets and
laptops did not show no effect on AIP usage, although 93.3% of the respondents mentioned
using laptops and 57.3% tablets (Table no. 1). However, all respondents mentioned that
they used smartphones for connectivity purposes, hence for AIPs too.
Conclusions
The scientific contributions of this study could be found on multiple levels. This study
expands the AI user behaviour literature by probing into the actual usage of AIPs, offering
a new perspective on the UTAUT model by retaining the only predictor that directly
explained usage, and including perceived risk and the negative view of trust in the form of
lack of trust. Additionally, the UTAUT model has been developed in this study by
converting gender and experience into independent variables, upgrading their role from the
moderating one in the original model (Venkatesh et al., 2003), and by including another
three descriptive variables to explain AIPs usage in general, and for educational purposes,
namely income, education and equipment used for Internet access. In this study, based on
the principal component analysis (PCA), perceived risk and perceived lack of trust were
combined into one variable (Risk and lack of trust associated with AIPs), thus prompting a

new theoretical perspective in the field. Moreover, as a result of using logistic regression,
the significant direct (positive) effect of facilitating conditions and the inverse (negative)
one of risk and lack of trust display a different view that can be applied to understanding AI
adoption or technology adoption in general. The significant results obtained in the case of
gender, education and types of equipment used for accessing the Internet indicate that these
variables could bring clarifications regarding the use of technology in general and AI, in
particular. Furthermore, the high correlation displayed between the two dependent variables
adds to the literature on AI in education, with this strong connection being rendered by
three common antecedents with similar impact, namely facilitating conditions, risk and lack
of trust related to AIPs and gender.
For managerial implications, the findings of this study show that universities must facilitate
the necessary means for users to exploit AIPs for educational purposes effectively. Thus,
higher education institutions should consider a comprehensive integration of AIPs to cover
teaching, learning, and faculty-student interaction (Wu et al., 2022), as well as assessment
and administrative tasks. Faculty and students have been accustomed to working remotely
since the time of the pandemic (Istudor et al., 2020). AIP providers should also attempt to
alleviate perceived risk and lack of trust in technology by providing comprehensive
information on the actual usage of AIPs and data protection and privacy. The providers
should create adequate instructional materials, provide support services and encourage
interaction with and between users about processes and outcomes of the implementation of
this procedure. Furthermore, AIPs and supportive resources should be tailored based on the
educational level of the target audience and the mobile equipment, as the results of this
study show that all respondents used smartphones to access the Internet. Regarding public
policies, the authorities should consider allocating funds for the purchase of equipment and
AIPs, but also for training academics on the usage of AIPs. In addition, as a country
strategy, funding of scientific research must be prioritised in this extremely dynamic field.
The study presents some limitations that could be addressed as directions in future research.
Firstly, the study focused on business students from Romania. Future research can employ
the model to investigate students from other fields, as well as for multi-country analysis.
Secondly, the research made use of a nonprobability sampling procedure, which makes the
generalisation of the findings difficult. Future studies should employ a probability sampling
approach. Third, the sample used presented an imbalance between students enrolled in
undergraduate study programmes and those enrolled in postgraduate programmes. Future
studies could seek to ensure representativeness from the perspective of distributing students
between university and postgraduate programmes. Fourthly, the nonsignificant results
rendered by the usage of tablets and laptops, as well as the inverse effects recorded by
smartwatches and personal computers, should be further analysed. Future studies should
investigate the effect of the equipment used for Internet access using Likert scales. Fourth,
the model can be developed by including other descriptive variables, such as AI usage
purposes or AI usage frequency.
References
Alshammari, S., 2021. Determining the Factors that Affect the Use of Virtual Classrooms:
A Modification of the UTAUT Model. Journal of Information Technology Education:
Research, 20, pp. 117-135. https://doi.org/10.28945/4709.
Vol. 26 • No. 65 • February 2024 121

Alzahrani, L., 2023. Analyzing Students’ Attitudes and Behavior Toward Artificial
Intelligence Technologies in Higher Education. International Journal of Recent
Technology and Engineering (IJRTE), 11(6), pp. 65-73. https://doi.org/10.35940/
ijrte.F7475.0311623.
Arfi, W.B., Nasr, I.B., Kondrateva, G. and Hikkerova, L., 2021. The role of trust in
intention to use the IoT in eHealth: Application of the modified UTAUT in a consumer
context. Technological Forecasting and Social Change, 167, article no. 120688.
Bearman, M. and Ajjawi, R., 2023. Learning to work with the black box: Pedagogy for a
world with artificial intelligence. British Journal of Educational Technology, 54(5),
pp. 1160-1173. https://doi.org/10.1111/bjet.13337.
Burton, E., Goldsmith, J., Koenig, S., Kuipers, B., Mattei, N. and Walsh, T., 2017. Ethical
Considerations in Artificial Intelligence Courses. AI Magazine, 38(2), pp. 22-34.
https://doi.org/10.1609/aimag.v38i2.2731.
Intelligence trends in education: a narrative overview. Procedia Computer Science, 136,
pp. 16-24. https://doi.org/10.1016/j.procs.2018.08.233.
Information Technologies, 25(5), pp. 3443-3463. https://doi.org/10.1007/s10639-020-
10159-7.
Du, S. and Xie, C., 2021. Paradoxes of artificial intelligence in consumer markets: Ethical
challenges and opportunities. Journal of Business Research, 129, pp. 961-974.
Dziuban, C. and Moskal, P., 2001. Evaluating Distributed Learning in Metropolitan
Universities. Metropolitan Universities: An International Forum, 12(1), pp .41-49.
Escotet, M.Á., 2023. The optimistic future of Artificial Intelligence in higher education.
Prospects, pp. 1-10. https://doi.org/10.1007/s11125-023-09642-z.
European Parliament, 2020. What is artificial intelligence and how is it used?, [online]
Available at: <https://www.europarl.europa.eu/news/en/headlines/priorities/artificial-
intelligence-in-the-eu/20200827STO85804/what-is-artificial-intelligence-and-how-is-it-
used> [Accessed 10 February 2023].
Field, A., 2009. Discovering Statistics using SPSS. 3rd edition, London: Sage Publications.
Gansser, O.A. and Reich, C.S., 2021. A new acceptance model for artificial intelligence with
extensions to UTAUT2: An empirical study in three segments of application. Technology
in Society, 65, article no. 101535. https://doi.org/10.1016/j.techsoc.2021. 101535.
Hasan, R., Shams, R. and Rahman, M., 2021. Consumer trust and perceived risk for voice-
controlled artificial intelligence: The case of Siri. Journal of Business Research, 131,
pp. 591-597. https://doi.org/10.1016/j.jbusres.2020.12.012.
Huang, C., Zhang, Z., Mao, B. and Yao, X., 2023. An Overview of Artificial Intelligence
Ethics. IEEE Transactions on Artificial Intelligence, 4(4), pp. 799-819.
https://doi.org/10.1109/TAI.2022.3194503.
Istudor, N., Dinu, V., Gogu, E., Prada, E.M. and Petrescu, I.E., 2020. Impact of Education
and Economic Growth on Labour Migration in the European Union. A Panel Data

Analysis. E&M Economics and Management, 23(4), pp. 55-67. https://doi.org/

10.15240/tul/001/2020-4-004.
Kamila, M.K. and Jasrotia, S.S., 2023. Ethical issues in the development of artificial
intelligence: recognizing the risks. International Journal of Ethics and Systems, ahead-
of-print. https://doi.org/10.1108/IJOES-05-2023-0107.
Kelly, S., Kaye, S.-A. and Oviedo-Trespalacios, O., 2023. What factors contribute to the
acceptance of artificial intelligence? A systematic review. Telematics and Informatics,
77, article no. 101925. https://doi.org/10.1016/j.tele.2022.101925.
Khechine, H. and Lakhal, S., 2018. Technology as a Double-Edged Sword: From Behavior
Prediction with UTAUT to Students’ Outcomes Considering Personal
Characteristics. Journal of Information Technology Education: Research, 17,
pp. 063-102. https://doi.org/10.28945/4022.
Leichtmann, B., Humer, C., Hinterreiter, A., Streit, M. and Mara, M., 2023. Effects of
Explainable Artificial Intelligence on trust and human behavior in a high-risk decision
task. Computers in Human Behavior, 139, article no. 107539. https://doi.org/10.1016/
j.chb.2022.107539.
Lin, P.-Y., Chai, C.-S., Jong, M.S.-Y., Dai, Y., Guo, Y. and Qin, J., 2021. Modeling the
structural relationship among primary students’ motivation to learn artificial
intelligence. Computers and Education: Artificial Intelligence, 2, article no. 100006.
https://doi.org/10.1016/j.caeai.2020.100006.
Markauskaite, L., Marrone, R., Poquet, O., Knight, S., Martinez-Maldonado, R., Howard,
S., Tondeur, J., De Laat, M., Buckingham Shum, S., Gašević, D. and Siemens, G., 2022.
Rethinking the entwinement between artificial intelligence and human learning: What
capabilities do learners need for a world with AI? Computers and Education: Artificial
Matzavela, V. and Alepis, E., 2021. Decision tree learning through a Predictive Model for
Student Academic Performance in Intelligent M-Learning environments. Computers
and Education: Artificial Intelligence, 2, article no. 100035. https://doi.org/10.1016/
j.caeai.2021.100035.
Nunnally, J.C., 1978. Psychometric theory. 2nd Edition, New York: McGraw-Hill.
Nurjanah, R.L. and Pratama, M.R.A., 2020. Self-regulated learning strategy instructions in
reading comprehension skill learning during outbreak era. Journal of English Language
Teaching and Linguistics, 5(2), pp.191-201. https://doi.org/10.21462/jeltl.v5i2.409.
Osborne, J.W. and Costello, A.B., 2009. Best Practices in Exploratory Factor Analysis:
Four Recommendations for Getting the Most from Your Analysis. Pan-Pacific
Management Review, 12(2), pp.131-146.
Papaspyridis, A., 2020. AI in Higher Education: opportunities and considerations, March
26. [online] Available at: <https://news.microsoft.com/apac/2020/03/26/ai-in-higher-
education-opportunities-and-considerations/> [Accessed 15 July 2023].
Park, S.H., Do, K.H., Kim, S., Park, J.H. and Lim, Y.S., 2019. What should medical
students know about artificial intelligence in medicine?. Journal of Educational
Evaluation for Health Professions, 16(1), article no. 18. https://doi.org/10.3352
jeehp.2019.16.18.
Vol. 26 • No. 65 • February 2024 123

Petrosyan, A., 2023. Internet usage penetration in the U.S. 2022, by age group and device.
[online] Available at: <https://www.statista.com/statistics/1360723/us-internet-usage-
penetration-by-age-group-and-device/> [Accessed 3 July 2023].
Pew Research Center, 2021. Experts Doubt Ethical AI Design Will Be Broadly Adopted as the
Norm Within the Next Decade, June 16. [online] Available at:
<https://www.pewresearch.org/internet/2021/06/16/experts-doubt-ethical-ai-design-will-
be-broadly-adopted-as-the-norm-within-the-next-decade/> [Accessed 3 April 2023].
Pillai, R., Sivathanu, B., Metri, B. and Kaushik, N., 2023. Students' adoption of AI-based
teacher-bots (T-bots) for learning in higher education. Information Technology &
People, ahead-of-print. https://doi.org/10.1108/ITP-02-2021-0152.
Pisica, A.I., Edu, T., Zaharia, R.M. and Zaharia, R., 2023. Implementing Artificial
Intelligence in Higher Education: Pros and Cons from the Perspectives of Academics.
Societies, 13(5), 118. https://doi.org/10.3390/soc13050118
Raffaghelli, J.E., Rodríguez, M.E., Guerrero-Roldán, A.E. and Baneres, D., 2022. Applying
the UTAUT model to explain the students' acceptance of an early warning system in
Higher Education. Computers & Education, 182, article no. 104468. https://doi.org/
10.1016/j.compedu.2022.104468.
Rico-Bautista, D., Medina-Cardenas, Y., Coronel-Rojas, L.A., Cuesta-Quintero, F.,
Maestre-Gongora, G. and Guerrero, C.D., 2021. Smart university: key factors for an
artificial intelligence adoption model. In: Advances and Applications in Computer
Science, Electronics and Industrial Engineering: Proceedings of CSEI 2020, pp. 153-166.
Singapore: Springer Singapore.
Rivers, D.J., Nakamura, M. and Vallance, M., 2022. Online self-regulated learning and
achievement in the era of change. Journal of Educational Computing Research, 60(1),
pp. 104-131. https://doi.org/10.1177/07356331211025108.
Robinson, S.C., 2020. Trust, transparency, and openness: How inclusion of cultural values
shapes Nordic national public policy strategies for artificial intelligence (AI).
Technology in Society, 63(C), article no. 101421. https://doi.org/10.1016/
j.techsoc.2020.101421.
Rodríguez-Hernández, C.F., Musso, M., Kyndt, E. and Cascallar, E., 2021. Artificial neural
networks in academic performance prediction: Systematic implementation and predictor
evaluation. Computers and Education: Artificial Intelligence, 2, article no. 100018.
Roy, R., Babakerkhell, M.D., Mukherjee, S., Pal, D. and Funilkul, S., 2022. Evaluating the
intention for the adoption of artificial intelligence-based robots in the university to
educate the students. IEEE Access, 10, article no. 125666. https://doi.org/10.1109/
access.2022.3225555.
Schmitt, B., 2019. From atoms to bits and back: A research curation on digital technology
and agenda for future research. Journal of Consumer Research, 46(4), pp. 825-832.
https://doi.org/10.1093/jcr/ucz038.
Shaya, N., Madani, R. and Mohebi, L., 2023. An Application and Extension of the UTAUT
Model: Factors Influencing Behavioral Intention to Utilize Mobile Learning in UAE
Higher Education. Journal of Interactive Learning Research, 34(1), pp. 153-180.

Strzelecki, A., 2023. To use or not to use ChatGPT in higher education? A study of
students’ acceptance and use of technology. Interactive Learning Environments,
pp. 1-14. https://doi.org/10.1080/10494820.2023.2209881.
Taylor, P., 2023. Devices with access to the internet at home in the Netherlands 2019, by
age group. [online] Available at: <https://www.statista.com/statistics/573347/devices-
with-access-to-the-internet-at-home-in-the-netherlands-by-age-group/> [Accessed 10
July 2023].
Thiebes, S., Lins, S. and Sunyaev, A., 2021. Trustworthy artificial intelligence. Electronic
markets, 31(2), pp. 447-464. https://doi.org/10.1007/s12525-020-00441-4.
Touretzky, D.S., Gardner-McCune, C., Martin, F. and Seehorn, D., 2019. K-12 guidelines
for artificial intelligence: what students should know. [online] Available at:
<https://upload01.uocslive.com/ISTE/ISTE2019/PROGRAM_SESSION_MODEL/HA
NDOUTS/112142285/ISTE2019Presentation_final.pdf> [Accessed 10 July 2023].
Venkatesh, V., Morris, M.G., Davis, G.B. and Davis, F.D., 2003. User acceptance of
information technology: Toward a unified view. MIS quarterly, 27(3), pp. 425-478.
https://doi.org/10.2307/30036540.
Wang, H., An, N., Ma, S. and Rajakani, K., 2022. Innovative Methods for the Integration of
Marketing Concepts and Art Design in Product Design in the Era of Internet of
Things. Wireless Communications and Mobile Computing, 2022, pp. 1-12.
https://doi.org/10.1155/2022/8435815.
Wang, S., Christensen, C., Cui, W., Tong, R., Yarnall, L., Shear, L. and Feng, M., 2023.
When adaptive learning is effective learning: comparison of an adaptive learning system
to teacher-led instruction. Interactive Learning Environments, 31(2), pp. 793-803.
https://doi.org/10.1080/10494820.2020.1808794.
Wrycza, S., Marcinkowski, B. and Gajda, D., 2017. The enriched UTAUT model for the
acceptance of software engineering tools in academic education. Information systems
management, 34(1), pp. 38-49. https://doi.org/10.1080/10580530.2017.1254446.
Wu, W., Zhang, B., Li, S. and Liu, H., 2022. Exploring Factors of the Willingness to
Accept AI-Assisted Learning Environments: An Empirical Investigation Based on the
UTAUT Model and Perceived Risk Theory. Frontiers in psychology, 13, article no.
870777. https://doi.org/10.3389/fpsyg.2022.870777.
Wuensch, K.L. 2014. Binary Logistic Regression with SPSS. [pdf] Available at:
<http://core.ecu.edu/psyc/wuenschk/MV/MultReg/Logistic-SPSS.pdf> [Accessed 1 April
2023].
Wuensch, K.L. and Poteat, G.M., 1998. Evaluating the Morality of Animal Research:
Effects of Ethical Ideology, Gender, and Purpose. Journal of Social Behavior and
Personality, 13(1), pp.139-150.
Zaharia, R., Zaharia, R.M., Edu, T. and Costel Negricea, I., 2022. Exploring student
satisfaction with online education during the Covid-19 pandemic in Romania: a logistic
regression approach. Transformations in Business & Economics, 21(2), pp. 41-62.
Zeide, E., 2019. Artificial intelligence in higher education: Applications, promise and
perils, and ethical questions. Educause Review, 54(3), pp. 1-11.
Zhang, X., Chan, F.T., Yan, C. and Bose, I., 2022. Towards risk-aware artificial
intelligence and machine learning systems: An overview. Decision Support Systems,
159, article no. 113800. https://doi.org/10.1016/j.dss.2022.113800.
Vol. 26 • No. 65 • February 2024 125

AE Artificial Intelligence Adoption in the Workplace and Its Impact on the
Upskilling and Reskilling Strategies
ARTIFICIAL INTELLIGENCE ADOPTION IN THE WORKPLACE

AND ITS IMPACT ON THE UPSKILLING AND RESKILLING STRATEGIES
Constanța-Nicoleta Bodea1* , Mario Paparic2 , Radu Ioan Mogoș3
and Maria-Iuliana Dascălu4
1)
Bucharest University of Economic Studies, Romania and National Institute
for Economic Research, Romanian Academy, Bucharest, Romania
2)
Alma Mater Europaea ECM, Maribor, Slovenia and Calor Gas Ireland, Dublin,
Ireland
3)
4)
National University of Science and Technology Politehnica Bucharest, Romania

Bodea, C.N, Paparic, M., Mogoș, R.I. and Dascălu, Received: 29 September 2023
M.I., 2024. Artificial Intelligence Adoption in the Revised: 28 November 2023
Workplace and Its Impact on the Upskilling and Accepted: 27 December 2023
Reskilling Strategies. Amfiteatru Economic, 26(65),
pp. 126-144.
Abstract
The technology innovation, especially in the case of artificial intelligence, has significantly
transformed the work processes and how they are organised and performed. Even if the
adoption of advanced technologies usually leads to a higher work performance, there are
risks of negative disruptions in the working systems, such as non-ethical use and social
negative effects. The paper presents the results of an ethnographic research conducted by the
authors, with the objective to identify the impact of the artificial intelligence adoption in the
workplace on the professional knowledge and skills requirements and on the upskilling and
reskilling strategies. Three different domains were considered: information technology,
education, and scientific research. One relevant conclusion of the research is that knowledge
and skills requirements should be studied from multiple perspectives, such as profession
dynamics, not only from the technology innovation perspective. The research originality
mainly consists in the way in which the concept of the level of upskilling/reskilling
importance is defined and applied, based on professional knowledge and skills development
requirements. By using the assessed level of upskilling/reskilling importance, strategies and
related actions may be defined and undertaken. By substantiating this manner of setting up
*
Corresponding author, Constanța-Nicoleta Bodea – e-mail: [email protected]

AE
the upskilling and reskilling strategies and actions, the research has a theoretical and practical
impact in the domain of talent management.
Keywords: upskilling, reskilling, artificial intelligence, competence, ethnography, talent
management
JEL Classification: I25, J21, I29, M53, O33
Introduction
The adoption of artificial intelligence (AI) in organisations reshapes the work processes,
leading to significant changes in the requirements regarding the professional knowledge and
skills. While AI adoption can lead to new opportunities and improved job satisfaction, it also
raises challenges and risks. Adopting AI can facilitate automation and improvement of the
recruitment strategies, but it can also engender fear and distrust among recruiters. It can also
change how employees perceive their role regarding expectations relating to AI adoption.
This may cause uncertainty for those who do not fully grasp information technology and
those whose job consists of repetitive low-engagement tasks such as data entry.
Using AI systems and tools in the workplace can revolutionise the integration of natural and
artificial learning (Laat et al., 2020). AI systems can be designed to provide real-time
feedback and workplace learning analytics to support complex problem-solving (Laat et al.,
2020). Integrating AI in the workplace can lead to a surge in a skilled workforce that can
effectively use AI in an ever-complex environment (Ameen et al., 2022). Integrating AI in
the workplace also has implications for skill acquisition and development. AI systems can
automate individualised adjustments to work environments and facilitate healthier worker
behaviours (Fukumura et al., 2021). This can lead to improvements in worker performance,
health, and well-being. However, it is essential to recognise that using AI systems does not
automatically lead to systematically improving employee skills. While AI systems can bring
many benefits to the workplace, their use must be accompanied by efforts to upskill and
reskill employees (Morandini et al., 2023).
To address the changes in the requirements regarding the professional knowledge and skills,
the professionals have the alternatives of upskilling and reskilling. Upskilling means the
expanding of existing skill set and reskilling represents the development of new skills, often
closed to the existing ones, but not necessarily. The new developed skills may lead the
professionals to embrace a completely different career path. Both upskilling and reskilling
are achieved through the means of different education and training actions; what is different
is only the result of these learning actions.
Although the professional development seems to be a well-defined endeavour, there is still a
need to research how the requirements about the professional knowledge and skills evolved
due to the AI adoption, in comparison with other factors also affecting the competence
requirements, such as the knowledge dynamics and other labour market specificities and,
based on that, how to assess the importance of upskilling/reskilling, in terms of priority and
urgence. To define a proper upskilling/reskilling strategy, it is not sufficient to identify the
needs for professional knowledge and skills development, but also to assess the importance
of these needs, in terms of their priority and urgency. A better understanding of the
connection between the upskilling/reskilling importance and different categories of actions
Vol. 26 • No. 65 • February 2024 127

for knowledge and skills development is still needed. To fill this research gap, the authors
decided to conduct the ethnographic research that is presented in the paper.
The paper is contributing to the existing theory of talent management by defining the concept
of level of upskilling and reskilling importance, in relation to different requirements for
professional knowledge and skills development. Many references to can be found in the
literature about the importance of upskilling and reskilling (Xu, 2011; Kim, 2017; Sawant et
al., 2022; Li, 2022)., but most of the time, the adopted approaches are very generic and
difficult to be operationalised and applied in the context of talent management activities. The
paper defines four levels of upskilling/reskilling importance, by combining the requirements
related to AI adoption with those independent to AI adoption. This is an original way of
substantiating the concept of level of upskilling/reskilling importance.
After the introduction section, the paper presents the main relevant concepts and topics
identified in the reviewed literature. The main sections of the literature review are AI
adoption and employment prospects, professional development/reskilling, and educational
and training actions. The paper presents and justifies the methodological approach of the
research, mentioning the characteristics and requirements of ethnographic research and the
main stages and methods of a research of this type. The research findings are then presented
and discussed. The research started by identifying the work processes that have undergone
transformations through the adoption of AI. According to the participants' statements, the
benefits obtained by using AI systems and tools vary depending on the field of work, but
most of the respondents referred to time savings and improved work quality. The pressure of
professional development requirements is clearly perceived by specialists, as it is directly
related to professional performance. A method based on requirement relevance classes was
defined in the research. Based on the relevance classes of requirements, the level of
importance of professional development and the strategy of professional development are
established, with the recommendation of educational actions. Three levels of importance of
upskilling/retraining were defined based on respondents' reflections: low importance,
medium importance, and high importance. This working method represents an original
contribution of the authors. Based on the research findings, conclusion and future research
directions are included. The reference section and Acknowledgements represent the final
parts of the paper.
1. Literature review
The use of AI systems and tools may potentially enhance the workplace productivity and
efficiency across various domains, including IT, research, and education (Fukumura et al.,
2021). AI may add value to managers’ work by supporting decision-making through
extensive data analysis and search and discovery activities (Yu et al., 2022). Additionally,
integrating AI in the workplace can drive future research and understanding of collaboration
with AI, particularly regarding employees' identity and relevant factors when introducing AI
(Mirbabaie et al., 2021). In research, AI systems have the potential to revolutionise the
integration of human and artificial learning, particularly in complex problem-solving
scenarios. Technological solutions and workplace learning analytics systems have been
designed to aid decision-making processes, and recent developments in AI can further
enhance this integration (Laat et al., 2020). In education, AI has the potential to significantly
impact the higher education. AI can improve teaching effectiveness by automatically scoring

AE
tasks such as English writing and supporting innovative approaches to teaching professional
courses (Slimi, 2021; Su et al., 2022). Integrating AI in education requires adopting modern
teaching methods and technologies, and it can significantly impact student subjectification
and pedagogical practices (Loftus and Madden, 2020; Ahmad et al., 2021). AI-assisted
writing has also been identified as a potential area for enhancing productivity and efficiency
in the workplace. The benefits include increased efficiency and better idea generation in
writing tasks, and students are encouraged to develop AI literacy to succeed in the workplace
(Cardon et al., 2023). In the IT sector, AI systems have been applied in various domains,
including retail, where intelligent automation of process change can improve operational
efficiency and productivity (Manasa and Jayanthila-Devi, 2022). The use of AI in the
workplace also raises ethical challenges, particularly regarding worker surveillance and
productivity-scoring tools. Implementing AI systems can extend and systematise ethical
failings and fundamentally change the relationship between workers and their managers
(Hickok and Maslej, 2023). Additionally, the use of AI in the gig economy has implications
for worker rights and the potential for bias and discrimination (Tan et al., 2021).
1.1. AI adoption and job prospects

Jetha (2023) discusses the work changes in the context of AI and proposes a research agenda
to examine how AI affects the relationship between socioeconomic, political, and working
conditions and worker health and employment outcomes. Braganza et al. (2021) focus on the
impact of AI adoption on psychological contracts, job engagement, and employee trust. The
study suggests that AI adoption can lead to new opportunities for productive employment
and decent work. It highlights how new online platforms can help employees and
organisations find more opportunities in the global marketplace. Naudé (2020) discusses the
potential impacts of AI adoption, highlighting that the effects are unlikely to be utopian or
apocalyptic soon. The study cautions against underestimating the long-term impacts of AI
and highlights the need to consider the potential consequences of AI adoption. Regona et al.
(2022) focus on the construction industry and explores AI's opportunities and adoption
challenges. Jain et al. (2021) focus on AI's employability implications in the healthcare
ecosystem. Ore and Sposato (2021) explore the opportunities and risks of AI adoption in
recruitment and selection processes. The study highlights how AI can facilitate the adequate
performance of routine tasks through automation. Tursunbayeva and Renkema (2022)
investigate the impact of AI applications in the healthcare sector on the job design of
healthcare professionals. Jetha (2023) highlights the implications of AI adoption for
employment opportunities, work environments, and worker health, safety, well-being, and
equity. Jaiswal et al. (2021) discuss the upskilling of employees for AI adoption in
multinational corporations. In the study conducted by (Nguyen and Malik, 2021), the
attitudes and readiness towards AI adoption and usage across different levels and geographies
of organisations were discussed.
1.2. Upskilling and reskilling

One of the challenges is organisational readiness. Alami et al. (2020) emphasise the
importance of studying organisational readiness to integrate AI into healthcare delivery. They
argue that organisational readiness should be carefully considered to ensure the successful
implementation of AI systems in healthcare organisations. Another challenge is the impact
Vol. 26 • No. 65 • February 2024 129

of AI on job attitudes and career behaviours. Presbitero and Teng-Calleja (2022) discuss how
incorporating AI in the workplace can lead to new types of jobs requiring new competencies.
Wilkens (2020) points out that while AI has the potential to support these processes, it can
also have unintended effects. AI systems may reinforce specific learning patterns, which can
be both beneficial and detrimental to individuals and organisations. AI can also support
lifelong learning and education. Poquet and Laat (2021) discuss the role of AI in lifelong
learning and highlight the opportunities for technology-mediated education. They argue that
AI can facilitate personalised learning experiences and provide learners access to various
resources and learning materials.
AI and machine learning technologies can enhance and replace specific job roles, leading to
changes in the demand for specific skills (Muhammad et al., 2023). The impact of AI on skill
development and learning is not limited to specific industries or sectors. The use of AI in the
workplace is on the rise across various domains, and it is essential to understand the
consequences of AI adoption and application (Yu et al., 2022). The use of AI in the workplace
requires organisations to understand its capabilities and shortcomings and address the
predicted impact on skills, roles, and employee morale (Treacy, 2022).
Integrating AI systems in the workplace significantly impacts skill development and learning.
AI technologies have the potential to revolutionise the integration of human and artificial
learning, leading to a surge in a skilled workforce that can effectively use AI in complex
environments. However, using AI systems does not automatically lead to a systematic
improvement in employees’ skills, and efforts to upskill and reskill employees are necessary.
The impact of AI on workforce skills and economic mobility has been studied in various
contexts, including developing countries and medical education.
1.3. Education & training actions

One important factor in organisational AI readiness is adopting AI technologies. Adoption
refers to the organisation's decision to use an innovation, such as AI, in its operations (Jöhnk
et al., 2020). To effectively adopt AI, organisations must consider various factors, including
the readiness of their employees and the organisational structure. A centralised organisational
structure has been found to support the effective adoption of AI principles (Kelley, 2022).
However, the design and implementation of AI in education require a learning sciences-
driven approach to ensure the effective use of AI algorithms (Luckin and Cukurova, 2019).
Adopting AI in human resource management (HRM) also impacts employee training and
support. AI adoption in HRM can enhance HR system effectiveness, but it requires careful
consideration of the impact on HR systems (Agarwal, 2022). Factors influencing the adoption
of AI in organisations have been studied using diffusion of innovation and technology-
organisation-environment theories (Kurup and Gupta, 2022). AI adoption can potentially
improve efficiency and effectiveness. AI-based hiring practices can contribute to the
underdeveloped, yet emerging paradigm of AI-based hiring in practice and research (Bhatt,
2022). Organisations must establish AI governance frameworks to ensure AI's ethical and
responsible use. AI governance at the organisational level should be defined and integrated
into existing governance areas, such as corporate governance, information technology
governance, and data governance (Mäntymäki et al., 2022). This ensures that AI adoption is
aligned with organisational values and principles.

AE
Although the professional development seems to be well known, there are still a need to
research how the requirements about the professional competences evolved due to the AI
adoption, in comparison with other factors affecting the competence requirements, such as
knowledge dynamics and other labour market specificities and, based on that, how the
importance of upskilling/reskilling might be assessed. It is not sufficient only to identify the
need for upskilling/reskilling but, for an adequate strategy in career development it is relevant
to assess the importance and urgency of this need. Also, there is still needed a better
understanding of the connection between the upskilling/reskilling importance and different
categories of competence development actions. To fill this research gap, the authors defined
the following two research questions:
RQ1: How are the requirements regarding the professional knowledge and skills changing
due to the AI adoption in the workplace?
RQ2: How is the level of upskilling/reskilling importance assessed and applied for defining
the upskilling/reskilling strategies and related educational actions due to the AI adoption at
workplace?
To properly address these research questions, the authors defined the research model
presented in Figure no. 1, which highlight the main factors under consideration. The research
was conducted at the level of working places, by the means of ethnography.
Considering that the topic under investigation is the impact of AI on the competence
requirements, we consider that the most adequate level for the research is the working place,
because only at this level it is possible to better understand the impact of AI adoption in a
large variety of contexts and for many professional competence categories.
Ethnography was chosen as the research approach, due to the possibility to collect a large
variety of data on the experiences, thoughts, and feelings from the professionals who are
using the AI technologies in their professional activities and were impacted by the AI
adoption. The experiences as reflected by the participants in this research are a valuable
information source that may reveals many relevant aspects of the topic.
Figure no. 1. The research model

Ethnography has been applied for the study of the AI emergence, especially for the study of
AI social implications. Hagerty and Rubinov (2019) conducted a literature review on the
Vol. 26 • No. 65 • February 2024 131

social impact of AI and connected technologies. They concluded that it is very likely AI to
socially impact differently the specific regions, and to lead to the increase of social inequality,
especially for the marginalised population. The AI negative social impact is more likely in
the countries with low income than in the countries with middle and high income. Marda and
Narayan (2021) advocate the importance of ethnography to understand the impact of AI at
society level, due to the focus on the societal actors and institutions that gain more influence
by using the AI technologies.
Any profession has its own dynamic, in terms of associated knowledge and professional
practices. The profession dynamics requires professional to develop their working skills and
abilities. Technological innovation is an important factor of transforming the profession. It is
widely accepted that when disruptive technologies, such as AI technologies are adopted, the
requirements regarding professional knowledge and skills may become even more relevant
and urgent.
The ethnographic investigation a qualitative research method that allows the study of
behaviours and events, within a day-by-days context/setting, based on data collected from
multiple sources (Bardi, 2021). The researchers must immerse in the research context for a
significant period. This prolonged involvement in the investigated context allows researchers
to conduct observations and to gain a deep understanding of behaviours and events, even in
the absence of knowledge included in the existing theory.
The research activities carried out are presented in figure no. 2 (Xu, 2011; Yu et al., 2022).
Figure no. 2. The research activities

AE
Attempting to establish the impact of AI adoption on competence requirements, the research

collected data based on the observations performed by the authors during February-June
2023, the reflective reports filled by 12 responders selected from the participants group
during August 2023 and interviews conducted in early September 2023, mostly for clarifying
some of the reflective reports. The validation of the research findings was also performed in
early September 2023.
Observations were made of the activities carried out by 45 specialists who use AI systems
and tools in the following three fields: education, scientific research and information
technology (IT), including the authors of this paper and 41 specialists with whom the authors
have coordination responsibilities, or they are their colleagues. Of the 45 participants, 30 are
from Romania, 5 are from Ireland and 10 are from Slovenia.
From the participants group, 12 professionals were selected mainly based on their availability
to contribute to the research, but also to have a balance between different work domains and
work experience. The selected professionals received a reflection guide, and they were asked
to reflect on their experience in using AI systems and tools at the working places. The
respondent profiles are shown in Table no. 1.
Table no. 1. The respondents’ profile
Career AI systems/tools used No. of years using
ID Job title Domain
stage* at work AI systems/tools
R1 IT system IT LC Chat-GPT plus 1
analyst
R2 IT technical IT LC Python, Chat-GPT plus 1
leader
R3 IT consultant IT LC Python, Chat-GPT plus, 1
Google Bard
R4 Data scientist IT MC Python, SQL, Neo4j, 4
Airflow, Kubernetes,
Docker
R5 Researcher Research LC Oracle products for BI 10
Suite
R6 Researcher Research EC Chat-GPT, Google Bard 1
R7 Researcher Research MC Power BI 2
R8 Researcher Research LC Biopython, InterMine, 6
R Suite
R9 Professor Education MC ChatGPT, SnatchBot 4
(Chatbot)
R10 Professor Education MC SnatchBot (Chatbot) 11
Flow XO, IBM
Watson(x)
Assistant/Services,
WEKA (datamining),
ZEUS (intelligent
software agents)
R11 Professor Education MC Custom-based software 2
to build students’ team
for projects, Turnitin
with AI functions
Vol. 26 • No. 65 • February 2024 133

Career AI systems/tools used No. of years using

ID Job title Domain
stage* at work AI systems/tools
R12 Career Education MC CAS++ (psychological 2
counsellor assessment platform),
CareProfSys
(recommender system
using AI and virtual
reality)
Note: *Career stage: EC–Early Career, MC-Mid Career, LC-Late Career
The main data collection instrument was the reflection guide, applied to capture reflections
(personal experience, thoughts, and feelings). The structured of the reflection guide include
one section with questions about the professional profile and the following sections, one for
each reflection topic:
 Changes in the work processes (job content);
 Requirements for reshaping the professional competences’ profile;
 Strategies for reshaping the professional competences’ profile;
 Incentives and rewards offered by the organisation;
 Perceived outcomes;
 Further expectations about AI adoption at the working place.
The interview protocol follows the structure of the reflection guide, considering that the
interviews were performed only for getting a better understand of the reflective reports. The
structure of the observation diary was a standard one, including a section for the field notes
and one section for questions and answers. The collected data collections were the following:
twelve reflective reports, four interview transcripts, and five observation diaries.
In the case of reflective reports and interview transcripts, data analysis was performed using
the coding technique. Table no. 2 presents an example for the transition from codes to
categories and themes. The observation diaries were inductively and deductively analysed.
First, the content was annotated using codes, that were later grouped to set up larger sematic
units. The semantic units were compared and integrated with themes identified from
reflective reports and interview transcripts.
Table no. 2. Sample of the results obtained from coding
Codes Category Theme
- Adoption (AI systems/tools) Knowledge and skills requirements Upskilling/reskilling
- Change (work processes) criticality, before AI adoption importance
- Revise (procedures) Knowledge and skills requirements
- Extend (knowledge) criticality, after AI adoption
- Improve (communication)
- Improve (critical thinking)
To validate the data analysis results, they were sent to three respondents. Based on the
received feedback, another data analysis cycle was performed.

AE
3. Research findings and discussion

This section is organised by themes, as they were identified during the data coding. For each
theme, the associated categories guided the interpretation and discussion of the research
findings. The themes include the answers to the two research questions.
3.1. Work processes transformation
Tables no. 3 and no. 4 present the respondents’ testimonials about the changes in their
workplace.
Table no. 3. The impact of AI adoption on the work processes in the IT domain
Domain/ Work processes affected by the adoption of AI systems/tools (with quotes
role included)
IT/system  Translating business requirements into technical requirements
analyst  Summarisation of large text that need to be digested (for example, agreements,
contracts)
 Analysing project reports, for discovering patterns
 Communication. „English is not my first language. I use Chat-GPT-4 to reframe
my writing and adapt it to a specific audience”. (R1)
 Producing professional presentations, including text and diagrams/images.
IT/  Organising testing and acceptance testing internally within the IT team and
technical business unit
leader  Generating testing scenarios and adapting to testing systems.
 Writing and debugging complex SQL queries to create reports from a database.
“Chat-GPT-4 is not perfect in this regard, so it is not always a time-saving
method, but it is helpful most of the time”. (R2)
IT/  Managing external development partners (R3)
consultant  Summarisation of large text that need to be digested (R3)
and data  Improving the analytical machine learning (ML) models that allow stakeholders to
science make better decisions (R4)
 “Due to the increasing complexity of a ML project, the focus is on best software
development practices and in keeping up to date the new releases”. (R4)
Table no. 4. The impact of AI adoption on the work processes in the scientific research
and education domains
Domain/ Work processes affected by the adoption of AI systems/tools
role (with quotes included)
Scientific  Identification of research topics/sub-topics and associated references
research/  Summarisation of text, for structuring the literature review
PhD student  Producing professional presentations, including not only text but also
diagrams/images
Scientific  Improving the analytical machine learning (ML) models (R8)
research/  Analysing large data sets, for discovering patterns (R5, R8)
Experienced  Producing professional presentations, including text and diagrams/images (R7)
researcher
(MC or LC)
Education/  Assessment of students (based on the annual exams, tests during the semester,
teaching bachelor’s, or master's theses). “I periodically change the student evaluation
methods to avoid situations in which the result of the exam/verification test, of
Vol. 26 • No. 65 • February 2024 135

Domain/ Work processes affected by the adoption of AI systems/tools

role (with quotes included)
the bachelor's thesis or master's degree may not truly reflect the student's
competences”. (R9)
 “I introduced chatbots use, during tests or exams, to ask students some
questions based on predefined scenarios. I update periodically these scenarios
to create the student's impression that the set of questions he is answering is
unique”. (R9)
 “I am teaching AI discipline for almost 15 years, and I constantly updated the
content of my lectures, considering that a lot of AI tools and platforms have
appeared. To teach this discipline, I performed a constant documentation
regarding the latest news in the AI field as well as the identification of tools that
could be used for the lab activities”. (R10)
 “In order to use a new AI system/tool for educational purpose, I have to check if
the platform provider requests fee for certain facilities at any time. Also, I must
develop training data sets”. (R10)
 “I am teaching Systems Engineering discipline. The course is a complex one, in
which students are learnt how to develop and manage interdisciplinary and
challenging projects, such as Apollo aircraft. At the practical part, they must
work in teams and to develop a report and a prototype of a complex project. For
a better composition of teams, I decided to use a specialised AI tool”. (R11)
 “I started using Turnitin to check if the students’ reports were generated with AI
tools, such as ChatGPT. I had to use this plagiarism checker, as the students
extensively exploited ChatGPT for any academic task”. (R11)
Education/  “I ask students who come to CCOC (Career Counselling and Guidance Centre
Services on of our university) to use CAS++ platform, which gives them an individualised
Career feedback report based on Jackson Vocational Interest Survey and MLQ –
Counselling Multifactor Leadership Questionnaire. Further on, we will discuss this report
and face-to-face. The students are more interested in discovering their potential
Guidance using modern tools and more open to further discussions”. (R12)
 “A tool was developed by our fellow colleagues from IT departments, which
offers virtual reality games containing specific activities for the recommended
profession: CareProfSys automatically assesses the user profile and
recommends suitable professions, then allows the user to test activities which
he/she may do in that profession. I noticed that the automated feedback given by
these advanced platforms are accurate enough”. (R12)
In addition to the identification of the transformed work processes, the participants made an
estimation of the percentage of the total work time that is spent on transformed processes and
the impact on the work productivity.
“Around 20% of the working time is related to the performance of transformed processes.
And I can say that using AI tools, I save around one-two hours per day on average, depending
on the type of work.” (R1)
The benefits achieved by using AI systems and tools varies according to the domain of work,
but most of the responders have referred to saving time and improving the quality of their
work.

AE
3.2. Organisational formal policy regarding the AI systems and tools use
Some of the responders declared that there is no formal policy regarding the AI tools usage.
But, in relation to the AI systems, there are always clear responsibilities and rules.
“In my company, there is not a formal policy on the AI tools usage, and, for this reason,
nobody knows how many people are using such tools and in what capacity. I have informed
my colleagues and my manager from the IT department that I am using Chat-GTP, with
different add-ons. The reaction was mixed; some were interested and wanted to know more,
but others found it interesting, but weren’t keen on using it without a formal announcement
from the management. My manager was OK with that.” (R2)
3.3. Upskilling/Reskilling importance

Usually, the professionals perceive the difference between the regular upskilling/reskilling
and that related to technological innovation.
“In IT, the work methods and tools are changing all the time, so I constantly must learn. It is
OK, because I perceive this as an opportunity for personal and professional development.”
(R3)
The AI adoption may have a different impact on the knowledge and skills development.
“For using an AI tool, I must learn how this tool is working, which are the benefits but also
the risks of using it.” (R3)
“The AI systems and tools usage represents the core of my professional responsibilities. For
improving my work performance, I must constantly learn new AI analytical methods and
procedures.” (R4)
The requirements’ criticality for upskilling/reskilling is clear perceived by the professionals
because it relates to work performance.
“Some of the new AI methods that I adopt in my daily work impose to revisit my mathematical
or statistical background, which is not easy to be done. But I perceive it as being a critical
requirement, so I dedicate the time that is needed for do that.” (R8)
But technological innovation is not the only factor that determines the emergence of new
requirements for upskilling/reskilling. Suppose that a new method of quantitative analysis of
project risks has been introduced in project management, the application of which requires
advanced knowledge of statistics. The requirement to expand the knowledge of the project
manager is justified by ensuring better results in estimating the probability of occurrence of
risks. The more risks associated with the project, the greater the relevance of the knowledge
expansion requirement might be, as it has a greater impact on the professional performance
of the project manager. We also assume that the project manager also wants to use an AI tool
to ensure better risk identification in projects. Therefore, the project manager must learn to
use that tool and improve his communication with the relevant parties to build confidence
that he is using AI ethically. These latter requirements are synchronised with the acquisition
of new statistical knowledge, so they must be considered together as a set of professional
development requirements. But they are not as critical as the first requirement. In this case,
Vol. 26 • No. 65 • February 2024 137

the question is how to determine the relevance of the whole set of requirements and how to
decide the professional development strategy?
The separate identification of professional development requirements in relation to the
factors involved, allows comparisons to be made between the impact of different factors, so
that different decisions to make organisational changes can be better substantiated, from the
perspective of the professional development requirements of employees. Most of the
respondents involved in the research stated that the professional development requirements
associated with the adoption of AI in the workplace are greater than the requirements
independent of the adoption of AI.
To combine upskilling/reskilling requirements, a method based on the requirements’
relevance classes was defined (Figure no. 3). The method involves the use of four classes of
requirements, defined based on the scale “low relevance, medium relevance and high
relevance” for each of the two types of requirements (those resulting from the adoption of AI
and those independent of AI). Requirements with the same level of relevance are included in
each class. Let us assume that the requirement to expand statistical knowledge belongs to
class 3, while the accumulation of knowledge to use the AI tool belongs to class 2, and the
requirement to improve communication with stakeholders regarding the ethical use of AI
belongs to class 3.
Starting from the relevance classes of the requirements, the level of upskilling/reskilling
importance is established, according to figure no. 3. For the given example, the level of
importance is “medium importance”. Based on the level of importance, the professional
development strategy may be established, and educational actions may be recommended.
Figure no. 3. The classes of professional development requirements and the levels
of upskilling/retraining importance

AE
Based on the respondents' reflections, three levels of upskilling/retraining importance were

defined: low importance, medium importance, and high importance. This method represents
an original contribution of the authors, although it bears some similarity to the work of
Jaiswal et al. (2021), but they defined the importance of upskilling/reskilling as a
combination of future and current competence requirements, there being no levels of
importance, but a continuum of values of the upskilling/reskilling importance.
According to the respondents’ testimonials, the low importance level of upskilling/reskilling
includes most of the hard skills, the medium importance level includes most of the transversal
skills, and the high importance level includes skills, such as: critical thinking, especially for
discriminating between ethical and non-ethical professional behaviour, building trust in
teams, effective communication, agility, and resilience.
3.4. Upskilling/Reskilling strategies

The upskilling/reskilling strategies are different in relation to the importance given to
improvement/retraining, but also to the practices in a certain field. For example, in the field
of IT most improvement efforts are based on courses and trainings paid for by the
organisation.
“The company where I work now actively supports my professional development and offers
me access to various courses (for example on Udemy) that can help me improve my skills
and knowledge and pays the fees for AI/ML conferences.” (R4)
In the research and education domains, the preference is for conferences, discussions with
peers, in formal or informal context. The expected support coming from organisation is
related to the attendance publication fees. Only for high importance level of upskilling, the
professionals attend courses and trainings.
“I prefer free opportunities to reshape my competence profile. My preferred strategy is to
participate at workshops, conferences, and ad-hoc discussions with my colleagues. Also, I
am reading scientific literature and software documentations. Regarding the university
support, I expect to pay the conference fees and the publication fees, for disseminating my
professional results.” (R11)
“I read many scientific papers. I have extended meetings with my colleagues, for exchanging
ideas. I also participate in international conferences and projects meetings. I expect my
company to pay the access fee to the scientific database, to pay for data that I use for testing
the analytical models, and to pay for conference fee and to cover the attendance costs at
project meetings when they are taking place outside my city/country.” (R8)
No respondents declared that their organisations are running mentorship and consultancy
programs.
According to the statements of the research participants, in the case of a high level of
importance of upskilling/reskilling, long-term professional development strategies with high
associated costs are preferred, while in the case of a low importance of upskilling/reskilling,
the strategy of discussions and peer feedback is preferred, and the self-learning improvement
strategy.
Vol. 26 • No. 65 • February 2024 139

Conclusions
The research was conducted by the authors to gain a deeper understanding of the implications
of adopting AI systems and tools in the workplace. Using an ethnographic approach allowed
the phenomenon to be investigated within those workplaces where AI systems and tools have
already been adopted. The research was based on the collection of a significant volume of
data with the help of observations and interviews organised by the authors, as well as with
the help of reflective reports written by 12 people selected from among the 46 research
participants.
First, the ethnographic investigation allowed to identify an important number of work
processes that were affected by the adoption of AI, in three important fields, namely: IT,
scientific research and the educational field. Secondly, it was possible to characterise the
level of involvement of organisations in the adoption of AI. If in the case of AI systems, they
are implemented in most cases through decisions of the organisation, regarding the use of AI
tools, in many cases the decision belongs directly to specialists, who may or may not inform
colleagues and/or managers.
The research enabled a clear differentiation between the knowledge and work skills
requirements due to the adoption of AI versus those due to other factors. This is relevant
when defining professional development and retraining strategies. Also, to substantiate this
type of strategies, the authors identified the need to introduce the concept of the level of the
importance of professional knowledge and skills development. This concept is different from
that proposed by Jaiswal et al. (2021), who defined the importance of upskilling/retraining
as a combination of future and current competence requirements, without considering
levels/steps of importance, but only a continuum of importance, which makes that concept
more difficult to operationalise. For this reason, the research is original and has a theoretical
and practical relevance in the talent management domain. And finally, based on the data
collected in the research, the authors analysed the conditions under which the various
strategies for professional development and retraining can be adopted.
Regarding the research limitations, the first one is related to the targeted fields, respectively:
IT, education, and scientific research. As is known, AI has been adopted in many fields of
activity, which justifies the extension of research to other fields as well. Secondly, the
research did not aim at an analysis of country differences, even though the research
participants come from three different countries, namely: Romania, Ireland, and Slovenia.
And finally, the research only covers the situation in 2023.
As future research, the authors consider the use of this methodological approach for other
fields of activity where AI systems and tools have begun to be used on a large scale, such as
journalism and the field of artistic creation. Also, the authors consider conducting
longitudinal research, for a duration of 5-7 years, to see to what extent the accelerated trend
of adopting AI systems and tools will be maintained in the future.
Acknowledgements
We are grateful to the professionals agreeing to reflect on their experiences in using AI
systems/tools, to participate in interviews and to validate the research findings.

AE
This work was supported by a grant of the Ministry of Research, Innovation and Digitization,
CNCS–UEFISCDI, project number TE 151 from 14/06/2022, within PNCDI III: “Smart
Career Profiler based on a Semantic Data Fusion Framework”.
References
Agarwal, A., 2023. AI adoption by human resource management: a study of its antecedents
and impact on HR system effectiveness. Foresight, [e-journal] 25(1), pp. 67-81.
https://doi.org/10.1108/FS-10-2021-0199.
Ahmad, S.F., Rahmat, Mohd.K., Mubarik, M.S., Alam, M.M. and Hyder, S.I., 2021.
Artificial Intelligence and Its Role in Education. Sustainability, [e-journal] 13(22),
Alami, H., Lehoux, P., Denis, J.-L., Motulsky, A., Petitgand, C., Savoldelli, M., Rouquet, R.,
Gagnon, M.-P., Roy, D. and Fortin, J.-P., 2020. Organizational readiness for artificial
intelligence in health care: insights for decision-making and practice. Journal of Health
Organization and Management, [e-journal] 35(1), pp. 106-114. https://doi.org/
10.1108/JHOM-03-2020-0074.
Ameen, N., Sharma, G.D., Tarba, S., Rao, A. and Chopra, R., 2022. Toward advancing theory
on creativity in marketing and artificial intelligence. Psychology & Marketing, [e-journal]
39(9), pp. 1802-1825. https://doi.org/10.1002/mar.21699.
Bardi, M., 2021. ‘How can I make something interesting for me relevant for the wider
community?’ – An ethnographic exploration of Romanian researchers’ adjustment to
research communication standards. Journal of English for Academic Purposes,
[e-journal] 49, article no. 100943. https://doi.org/10.1016/j.jeap.2020.100943.
Bhatt, P., 2023. AI adoption in the hiring process – important criteria and extent of AI
adoption. Foresight, [e-journal] 25(1), pp. 144-163. https://doi.org/10.1108/FS-07-2021-
0144.
Braganza, A., Chen, W., Canhoto, A. and Sap, S., 2021. Productive employment and decent
work: The impact of AI adoption on psychological contracts, job engagement and
employee trust. Journal of Business Research, [e-journal] 131, article no. 485-494.
Cardon, P., Fleischmann, C., Aritz, J., Logemann, M. and Heidewald, J., 2023. The
Challenges and Opportunities of AI-Assisted Writing: Developing AI Literacy for the AI
Age. Business and Professional Communication Quarterly, [e-journal] 86(3),
pp. 257-295. https://doi.org/10.1177/23294906231176517.
De Laat, M., Joksimovic, S. and Ifenthaler, D., 2020. Artificial intelligence, real-time
feedback and workplace learning analytics to support in situ complex problem-solving: a
commentary. The International Journal of Information and Learning Technology,
[e-journal] 37(5), pp. 267-277. https://doi.org/10.1108/IJILT-03-2020-0026.
Fukumura, Y.E., Gray, J.M., Lucas, G.M., Becerik-Gerber, B. and Roll, S.C., 2021. Worker
Perspectives on Incorporating Artificial Intelligence into Office Workspaces:
Implications for the Future of Office Work. International Journal of Environmental
Research and Public Health, [e-journal] 18(4), article no. 1690. https://doi.org/
10.3390/ijerph18041690.
Vol. 26 • No. 65 • February 2024 141

Hagerty, A. and Rubinov, I., 2019. Global AI Ethics: A Review of the Social Impacts and
Ethical Implications of Artificial Intelligence. [online] Available at: <http://arxiv.org/abs/
1907.07892> [Accessed 18 October 2023].
Hickok, M. and Maslej, N., 2023. A policy primer and roadmap on AI worker surveillance
and productivity scoring tools. AI and Ethics, [e-journal] 3(3), pp. 673-687.
https://doi.org/10.1007/s43681-023-00275-8.
Jain, M., Goel, A., Sinha, S. and Dhir, S., 2021. Employability implications of artificial
intelligence in healthcare ecosystem: responding with readiness. Foresight, [e-journal]
23(1), pp. 73-94. https://doi.org/10.1108/FS-04-2020-0038.
Jaiswal, A., Arun, C.J. and Varma, A., 2022. Rebooting employees: upskilling for artificial
intelligence in multinational corporations. The International Journal of Human Resource
Management, [e-journal] 33(6), pp. 1179-1208. https://doi.org/10.1080/09585192.
2021.1891114.
Jetha, A., Bakhtari, H., Rosella, L.C., Gignac, M.A.M., Biswas, A., Shahidi, F.V., Smith,
B.T., Smith, M.J., Mustard, C., Khan, N., Arrandale, V.H., Loewen, P.J., Zuberi, D.,
Dennerlein, J.T., Bonaccio, S., Wu, N., Irvin, E. and Smith, P.M., 2023. Artificial
intelligence and the work–health interface: A research agenda for a technologically
transforming world of work. American Journal of Industrial Medicine, [e-journal] 66(10),
pp. 815-830. https://doi.org/10.1002/ajim.23517.
Jöhnk, J., Weißert, M. and Wyrtki, K., 2021. Ready or Not, AI Comes – An Interview Study
of Organizational AI Readiness Factors. Business & Information Systems Engineering,
[e-journal] 63(1), pp. 5-20. https://doi.org/10.1007/s12599-020-00676-7.
Kelley, S., 2022. Employee Perceptions of the Effective Adoption of AI Principles. Journal
of Business Ethics, [e-journal] 178(4), pp. 871-893. https://doi.org/10.1007/s10551-022-
05051-y.
Kim, J.H., 2017. A Review of Cyber-Physical System Research Relevant to the Emerging IT
Trends: Industry 4.0, IoT, Big Data, and Cloud Computing. Journal of Industrial
Integration and Management, [e-journal] 02(03), article no. 1750011. https://doi.org/
10.1142/S2424862217500117.
Kurup, S. and Gupta, V., 2022. Factors Influencing the AI Adoption in Organizations.
Metamorphosis: A Journal of Management Research, [e-journal] 21(2), pp. 129-139.
https://doi.org/10.1177/09726225221124035.
Li, L., 2022. Correction to: Reskilling and Upskilling the Future‑ ready Workforce for
Industry 4.0 and Beyond. Information Systems Frontiers, [e-journal].
https://doi.org/10.1007/s10796-022-10318-w.
Loftus, M. and Madden, M.G., 2020. A pedagogy of data and Artificial Intelligence for
student subjectification. Teaching in Higher Education, [e-journal] 25(4), pp. 456-475.
https://doi.org/10.1080/13562517.2020.1748593.
Luckin, R. and Cukurova, M., 2019. Designing educational technologies in the age of AI:
A learning sciences‐ driven approach. British Journal of Educational Technology,
[e-journal] 50(6), pp. 2824-2838. https://doi.org/10.1111/bjet.12861.
Manasa, R. and Jayanthila Devi, A., 2022. Amazon’s Artificial Intelligence in Retail Novelty
- Case Study. International Journal of Case Studies in Business, IT, and Education,
[e-journal] pp. 787-804. https://doi.org/10.47992/IJCSBE.2581.6942.0233.

AE
Mäntymäki, M., Minkkinen, M., Birkstedt, T. and Viljanen, M., 2022. Defining
organizational AI governance. AI and Ethics, [e-journal] 2(4), pp. 603-609.
https://doi.org/10.1007/s43681-022-00143-x.
Marda, V. and Narayan, S., 2021. On the importance of ethnographic methods in AI research.
Nature Machine Intelligence, [e-journal] 3(3), pp. 187-189. https://doi.org/10.1038/
s42256-021-00323-0.
Mirbabaie, M., Brünker, F., Möllmann Frick, N.R.J. and Stieglitz, S., 2022. The rise of
artificial intelligence – understanding the AI identity threat at the workplace. Electronic
Markets, [e-journal] 32(1), pp. 73-99. https://doi.org/10.1007/s12525-021-00496-x.
Morandini, S., Fraboni, F., De Angelis, M., Puzzo, G., Giusino, D. and Pietrantoni, L., 2023.
The Impact of Artificial Intelligence on Workers’ Skills: Upskilling and Reskilling in
Organisations. Informing Science: The International Journal of an Emerging
Transdiscipline, [e-journal] 26, pp. 039-068. https://doi.org/10.28945/5078.
Muhammad, A., Umar, U.A. and Adam, F.L., 2023. The impact of Artificial Intelligence and
Machine learning on workforce skills and economic mobility in developing countries:
A case study of Ghana and Nigeria. Journal of Technology Innovations and Energy,
[e-journal] 2(1), pp. 55-61. https://doi.org/10.56556/jtie.v2i1.466.
Naudé, W., 2021. Artificial intelligence: neither Utopian nor apocalyptic impacts soon.
Economics of Innovation and New Technology, [e-journal] 30(1), pp. 1-23.
https://doi.org/10.1080/10438599.2020.1839173.
Nguyen, T. and Malik, A., 2022. A Two‐ Wave Cross‐ Lagged Study on AI Service Quality:
The Moderating Effects of the Job Level and Job Role. British Journal of Management,
[e-journal] 33(3), pp. 1221-1237. https://doi.org/10.1111/1467-8551.12540.
Ore, O. and Sposato, M., 2022. Opportunities and risks of artificial intelligence in recruitment
and selection. International Journal of Organizational Analysis, [e-journal] 30(6),
pp. 1771-1782. https://doi.org/10.1108/IJOA-07-2020-2291.
Poquet, O. and De Laat, M., 2021. Developing capabilities: Lifelong learning in the age of
AI. British Journal of Educational Technology, [e-journal] 52(4), pp. 1695-1708.
https://doi.org/10.1111/bjet.13123.
Presbitero, A. and Teng-Calleja, M., 2023. Job attitudes and career behaviors relating to
employees’ perceived incorporation of artificial intelligence in the workplace: a career
self-management perspective. Personnel Review, [e-journal] 52(4), pp. 1169-1187.
https://doi.org/10.1108/PR-02-2021-0103.
Regona, M., Yigitcanlar, T., Xia, B. and Li, R.Y.M., 2022. Opportunities and Adoption
Challenges of AI in the Construction Industry: A PRISMA Review. Journal of Open
Innovation: Technology, Market, and Complexity, [e-journal] 8(1), article no. 45.
https://doi.org/10.3390/joitmc8010045.
Sawant, R., Thomas, B. and Kadlag, S., 2022. Reskilling and Upskilling: To Stay Relevant
in Today’s Industry. International Review of Business and Economics, [e-journal] 7(1).
https://doi.org/10.56902/IRBE.2022.7.1.4.
Slimi, Z., 2021. The impact of AI implementation in higher education on educational process
future: A systematic review. Preprint. https://doi.org/10.21203/rs.3.rs-1081043/v1.
Su, J., Zhong, Y. and Ng, D.T.K., 2022. A meta-review of literature on educational
approaches for teaching AI at the K-12 levels in the Asia-Pacific region. Computers and
Vol. 26 • No. 65 • February 2024 143

Education: Artificial Intelligence, [e-journal] 3, article no. 100065.

https://doi.org/10.1016/ j.caeai.2022.100065.
Tan, Z.M., Aggarwal, N., Cowls, J., Morley, J., Taddeo, M. and Floridi, L., 2021. The ethical
debate about the gig economy: A review and critical analysis. Technology in Society,
[e-journal] 65, article no. 101594. https://doi.org/10.1016/j.techsoc.2021.101594.
Treacy, S., 2022. A Roadmap to Artificial Intelligence: Navigating Core Impacts to
Successfully Transform Organisations. European Conference on the Impact of Artificial
Intelligence and Robotics, [e-journal] 4(1), pp. 85-92. https://doi.org/10.34190/
eciair.4.1.923.
Tursunbayeva, A. and Renkema, M., 2023. Artificial intelligence in health‐ care:
implications for the job design of healthcare professionals. Asia Pacific Journal of Human
Resources, [e-journal] 61(4), pp. 845-887. https://doi.org/10.1111/1744-7941.12325.
Wilkens, U., 2020. Artificial intelligence in the workplace – A double-edged sword. The
International Journal of Information and Learning Technology, [e-journal] 37(5),
pp. 253-265. https://doi.org/10.1108/IJILT-02-2020-0022.
Xu, L.D., 2011. Enterprise Systems: State-of-the-Art and Future Trends. IEEE Transactions
on Industrial Informatics, [e-journal] 7(4), pp. 630-640. https://doi.org/10.1109/
TII.2011.2167156.
Yu, X., Xu, S. and Ashton, M., 2023. Antecedents and outcomes of artificial intelligence
adoption and application in the workplace: the socio-technical system theory perspective.
Information Technology & People, [e-journal] 36(1), pp. 454-474. https://doi.org/
10.1108/ITP-04-2021-0254.

AE
ARTIFICIAL INTELLIGENCE IN EDUCATION:

NEXT-GEN TEACHER PERSPECTIVES
Nicoleta Sămărescu1 , Robert Bumbac2 , Alin Zamfiroiu3
and Maria-Cristina Iorgulescu4
1)
National University of Science and Technology Politehnica Bucharest,
University Centre Pitești, Pitești, Romania.
2)3)4)
3)
National Institute for Research & Development in Informatics – ICI Bucharest,
Bucharest, Romania.

Sămărescu, N., Bumbac, R., Zamfiroiu, A. and Received: 30 September 2023
Iorgulescu, C.M., 2024. Artificial Intelligence in Revised: 23 November 2023
Education: Next-Gen Teacher Perspectives. Amfiteatru Accepted: 25 December 2023
Economic, 26(65), pp. 145-161.
Abstract
The progress made by artificial intelligence (AI) worldwide has led, through the Beijing
Consensus, UNESCO to recommend to governments, in 2019, to include AI in educational
policies and processes. While numerous studies inventory students' and teachers’ opinions
on the use of AI in education (AIEd), this article differs by analysing the factors that influence
the behavioural intention to use AI from the training stage of future primary and secondary
teachers in Romania. Through exploratory quantitative research, carried out on a sample of
270 students from the Faculty of Education, Social Sciences and Psychology, the interaction
of subjects with AI and the intention of integrating AIEd are followed using binary logistic
regression. The analysis conducted shows that among the six variables of the model, “the
confidence in personal ability to use AI” and “the perception of a greater number of
advantages” have a positive and significant impact on the willingness to use AI in the
educational process, more than “being previously used”, “the knowledge level” or “student
requirements”. These findings are of particular importance for the revision of teacher training
programmes and the development of educational policies that increase the confidence of
future teachers in the ability to use AI, eliminating fears or misconceptions about AI.
Keywords: artificial intelligence; education; technology; future teachers; teacher training;
behaviour intention; Technology Acceptance Model (TAM).
JEL Classification: O33, I20

Corresponding author, Bumbac Robert – e-mail: [email protected]
Vol. 26 • No. 65 • February 2024 145

AE Artificial Intelligence in Education: Next-Gen Teacher Perspectives
Introduction
Artificial intelligence has seen considerable development since 1943, when the first
mathematical model that computerised the biological neurone was developed (McCulloch
and Pitts, 1943), and 1956, when the Dartmouth conference, considered the birthplace of AI,
took place (McCarthy, 1998). Recent developments in AI caused UNESCO (2019b), through
the Beijing consensus, an event attended by 500 representatives of more than 100 member
states, to provide guidance and recommendations on developing policies and practices in
education that harness AI technologies.
Based on the identified literature, it is considered that the perception of teachers' behavioural
intention to use AI (identified in the Technology Acceptance Model - TAM) has been under
researched since the training stage (Davis, 1989). According to Choi, Jang and Kim (2023),
“teachers hardly integrate AI educational tools and little is known about their perceptions of
AI tools”. There is also limited knowledge about how students and teachers use AI systems
and how AI can be integrated into schools, colleges, and other educational institutions
(Williamson and Eynon, 2020). Given the role of future teachers in implementing AI tools in
their teaching and how they will inspire and influence generations of students, it is essential
to understand their perceptions of AI and the factors that influence their behavioural intention
to use the technology. Current research contributes to the development of the existing
literature by identifying and testing a first set of factors that influence the use of AI in
education. Based on these findings, as AI evolves and is better integrated into education,
additional factors can be identified and a more detailed perspective can be generated on the
best way to develop teacher training programmes so that teachers become interested and
prepared to use AI in their work.
For a proper integration of AI elements in education, the aim is to get a clear picture of what
stakeholders involved in the educational process want, by conducting as many studies as
possible. AI has made considerable progress in fields such as healthcare and business, while
in education AI is still in an incipient stage (Singh and Miah, 2019). Unlike most of the papers
and books published so far, which focus on students’ attitudes and perceptions of using AI
in the medical field in the UK (Sit et al., 2020; Chan and Hu, 2023; Kairu, 2023), the present
study investigates the perception of current students - prospective primary or pre-school
teachers regarding the integration of AI in the educational process. Thus, the objective of the
article is to explore the behavioural intention of future teachers to use AI and to identify
factors that influence this intention. To this end, a model based on binary logistic regression
was developed to determine the main factors that lead prospective teachers to use AI. The
specific objectives of our research are: (1) to understand the degree of familiarity of future
teachers with AI; (2) to study the perceptions of future teachers of the change brought about
by AI in the educator’s role; (3) to create a model that predicts the behavioural intention of
future teachers to use AI and to identify the main factors that influence them. For this
purpose, exploratory research was carried out with 270 respondents, current students of the
Faculty of Education Sciences of the National University of Science and Technology
Politehnica Bucharest - Pitești University Centre and the Râmnicu Vâlcea Territorial Centre.
Most of the respondents attend these teacher training courses, being current or prospective
teachers in primary or pre-school education in Romania.
One of the main desiderata identified worldwide is to increase the degree of “positive use”
of AI elements in education (UNESCO, 2019b). Although there are policies and programmes
that seek to integrate elements of AI into the education system, the success of this approach

AE
will largely depend on the knowledge and perception of teachers at the bottom of the
education system. Therefore, studying the intention to use AI is important, as it will be passed
on, directly or indirectly, to the generations of children that future teachers will train.
The article is divided into three parts: (1) review of the existing scientific literature on the
use of AI in education; (2) presentation of the research methodology, including a description
of the method, sample, and its characteristics; (3) presentation of the results and discussions
of the intention to use AI in future teaching careers and the factors that influence it, followed
by the conclusions of the article.
Today, the elements of AI are a reality that future teachers must understand and appropriately
integrate into their teaching practices to ensure the efficiency and effectiveness of the
educational process. Research by Kim, Soyata and Behnagh (2018) shows how AI can
provide teachers with real-time feedback during presentations, analysing audio and visual
elements of the presenter to determine the quality of the presentation and improve the
audience experience. A similar approach is found in Woolf et al. (2013) who appreciate that
AI can help improve education in the long term by personalising learning, improving student
experiences by evaluating datasets of teaching behaviour, motivation, and social interaction.
However, there are also negative aspects of AI, identified by Păvăloaia and Necula (2023),
such as the high costs of integrating technology, the fact that it can lead to job losses and
higher electricity consumption. It should also be considered that prolonged virtual time can
cause dependency, lack of empathy, and communication deficiencies; moreover, there is
concern about the security of the data used by AI, in case the data have vulnerabilities, can
be compromised or stolen (Pisica et al., 2023).
Although most stakeholders understand the benefits and progress brought about by AI, its
implementation is not as clear and there are many challenges and pressures on educational
institutions to set guidelines and standards in the use of AI (Bonsu and Baffour-Koduah,
2023). Moorhouse and Kohnke (2023) showed that AI tools will generate substantial changes
in curriculum and that teacher educators need support to develop the necessary skills to use
AI appropriately. Integrating AIEd is a complex process, which requires from the outset to
establish the correct working coordinates, so that teachers and students alike perceive the
benefits of AI, the safe and constructive ways of using technology in their work.
Designing the right way to integrate AI into educational institutions requires first a clear
understanding of the knowledge, skills, and perceptions of teachers and students alike.
Numerous scientific papers explore the knowledge and preconceptions of teachers related to
technologies, robots, or elements that generate artificial intelligence (GAI). Kaplan-
Rakowski et al. (2023) show that the more teachers have a better outlook on GAI, the more
often they use it. Based on this finding, the present research will analyse the extent to which
future teachers in Romania perceive more advantages or disadvantages in using AI. The
benefits of AI for teachers do not seem to be confined to generating content, Chounta et al.
(2022) highlight that, despite their limited knowledge of this type of technology, Estonian
teachers use AI as a tool to facilitate the access and use of multilingual content.
Unfortunately, there is still no consensus on methods and techniques to use AI in teaching
and research. Fahrman et al. (2020) explain this situation by the constantly changing skills of
teachers and the fact that “researching, exploring and understanding this knowledge and skills
Vol. 26 • No. 65 • February 2024 147

is also complex”. Teacher technical skills, knowledge, and perceptions greatly influence
student knowledge; therefore, this research will assess the extent to which prospective
teachers used AI and will consider that they have sufficient knowledge of it.
The differences in perceiving AI can also be explained by age; Chan and Lee (2023) show
that Gen Z (born 1990-2000) students are optimistic about the potential of AI technologies,
while Gen X (born 1960-1980) and Gen Y (born 1980-1990) teachers acknowledge its
benefits, but also express concerns about overreliance on, ethical and pedagogical
implications of this technology. From this point of view, this research will examine teachers’
perceptions of students’ requirements to use AI and, respectively, the appreciation of students
for teachers who use this technology. Therefore, AIEd is at the same time a challenge in
respecting the authenticity of computer-generated content, which can lead to the “sacrifice of
uniqueness and creativity” of the human mind, so it is imperative to adopt standards for
responsible and ethical behaviour in the use of this technology (Dalalah and Dalalah, 2023).
Holmes et al. (2022) point out that the low interest in the ethics of using AI in education may
lead to an underdevelopment of this field, limiting the impact of AI and the usefulness of this
technology.
A second broad category of scientific papers investigates the perception and interaction of
students with AI. Chao et al. (2021) observe that attitudes toward AI among students tend to
be predominantly positive. However, as Ravi Kumar and Raman (2022) note, students
consider that AI can be used in teaching and learning activities or other administration
processes, but do not trust its use in the assessment processes such as admissions or
examinations. Doumat et al. (2022) complement this finding by pointing out that, although
more than 57% of students believe that “assessment by AI is more objective”, only 26% of
them “want to be assessed by AI”.
As with teachers, students understand the potential of technology, but see its implementation
and use in their future careers as a distant and long process. Sit et al. (2020) confirm this with
the example of medical students who feel unprepared but understand the growing importance
of AI in their future work and therefore would like to learn more about it. Advances in
artificial intelligence (AI) have also revealed problems with the accuracy of computer-
generated data. Research by Dalalah and Dalalah (2023) shows that the use of AI in various
professions, such as medicine, can be harmful by providing erroneous positive or negative
test results, leading to delays in caring for health problems or the suggestion of unnecessary
procedures or therapies. Such errors can also occur in the use of AI in student assessment.
This may explain the opinion of students that “AI cannot replace the teacher”, but at the same
time they want more AI resources to help them achieve their educational goals (Zou et al.,
2020). The difference between different types of AI resources stems from the perceived
usefulness and simplicity of communication, as Kim et al. (2020) show, and the use of AI
can work especially when face-to-face interaction is not possible.
Unlike previous works, the present research investigates the perception of AI among future
teachers during their training stage and focusses on identifying the main factors and how they
influence the behavioural intention to use AI. Future teachers will play a crucial role in
integrating and using AI in the development and education of children and students from an
early age, helping them to develop cognitive, social, and communication skills. This study
takes place in a context where modern early childhood and school education is considered a
key stage for the cultivation and AI literacy among children, as demonstrated by various

AE
initiatives undertaken by organisations such as Stanford - AI4ALL or the International

Society for Technology in Education (ISTE).
Studies such as those by Al Darayseh (2023), Limna, Kraiwanit and Jangjarat, (2023) or
Leong and Chaichi, (2021) prove the importance of the scientifically established Technology
Acceptance Model (TAM) used to explain the acceptance and intention to use new
technologies (Davis, 1989). Building on this model, the present research will examine the
intention of the behavioural to use AI (identified in the TAM) AI among prospective teachers.
Just as Hasib et al. (2022) use logistic regression to predict student performance based on
historical data, the present research aims to use binary logistic regression as a method to
predict teachers’ intention to use AI based on attitude and perception of its usefulness.
The purpose of the present research is to explore the intention of prospective teachers to use
AI and to identify its determining factors. For this purpose, a quantitative exploratory
research was carried out, using a questionnaire-based survey, which tests the perceptions of
current students enrolled in teacher training programmes at the Pitești University Centre, the
Faculty of Education, Social Sciences and Psychology, and its regional branch in Râmnicu
Vâlcea. The questionnaire included 10 questions using the five-step Likert scale, from “to a
very small extent” to “to a very large extent”, 4 multiple choice questions, 1 open question,
and 4 demographic questions. Given the difficulty of conducting research on a representative
random sample, the sample included respondents based on the criterion of relevance, i.e.,
people studying education sciences and teacher education who are aware of the evolution of
AI technology and use (Jurconi et al., 2022). The choice of quantitative research is justified
by the objective of creating a model that predicts the use of AI by future teachers and
identifies factors that influence this behaviour. This required recording as many responses as
possible, in an organised way, allowing evaluation and validation of the data using statistical
methods of analysis (Bell and Waters, 2018). In addition, quantitative research has the
advantage of being more objective, by generating quantifiable results that allow for a possible
generalisation of findings for the whole population.
The questionnaire was sent by e-mail to all undergraduate students enrolled in the Pedagogy
of Primary and Pre-school Education section and to those enrolled in the Early Childhood
Education Master’s programme at the Faculty of Education, Social Sciences and Psychology,
National University of Science and Technology Politehnica Bucharest - Pitești University
Centre and Râmnicu Vâlcea Territorial Centre, most of them working or intending to work
in primary or pre-school education in Romania. The responses were collected between May
and August 2023 using the Google Forms platform, and data analysis and interpretation were
performed using SPSS statistical software. Therefore, of a total of 370 students, 270 valid
responses were recorded, resulting in a response rate of 73%.
Given that the integration of AI in education is still in its infancy stage, especially in primary
and secondary education in Romania, the present research aimed to open the door for
prospective teachers to use AI, and the respondents were not required to have practical
experience teaching with the help of AI. Participation in the research was optional, voluntary,
and completely anonymous in order to protect the identity of the respondents. When the
questionnaire was distributed, the information message included the information that the
completion of the questionnaire implied consent to the aggregate dissemination of the
Vol. 26 • No. 65 • February 2024 149

research results. As illustrated in Table 1, 97.4% of the sample is female dominated, which
is a characteristic of pre-school and primary education not only in Romania but worldwide.
When it comes to technology, the age variable is seen as important in openness to new
technologies. Thus, a balance between Generation Z respondents (41.48%) and Generation
Y respondents (45.18%), and only a small share of Generation X participants (13.33%) is
observed. Therefore, it can be said that the study is largely conducted in the age segment
familiar with new technologies. In terms of education level, all research participants are
pursuing Bachelor’s or Master’s degrees and are in the process of training for their future
careers. Participants in the study included 21.1% of current students and teachers in pre-
school and primary education, 39.62% of current students who want to become teachers in
pre-school or primary education, and 39.25% of respondents who do not yet have a job and
do not know what career path they will follow.
Table no. 1. Distribution of the sample by gender, age, level of education
and current occupation or career plan
Age Level Current occupation or
Gender
(years) of education career plan
Currently a
18-23 112 undergrad 38
preschool teacher
Female 263 uate 238
Currently a primary
24-39 122 student 19
school teacher
Future pre-school
40-60 36 66
teacher
MA Future primary
Male 7 32 41
student school teacher
> 60 0
Not yet employed
106
and no career plan
270 270 270 270
Source: Authors’ own calculations
The study compares the perspective of two distinct groups: Group 1 – those students who
have previously used AI and Group 2 – those who did not use AI. Several factors influencing
the relationship between teacher and AI were examined, such as the level of knowledge of
AI, the degree of interaction and previous exposure to AI elements, the readiness to use AI
in teaching activity, the desire of future teachers for further training in AI, changes brought
about by AI in student-teacher interaction, in improving the learning experience, the
simplification of the administrative activities, the expectations of pre-school children /
schoolchildren for the use of new technologies and their vision of integrating AI in the
teaching and learning process. Following this analysis, a binary logistic regression was
performed to predict the intention of future teachers to use AI and to understand its
influencing factors. The following variables were established for this purpose:
● dependent variable: I17 – Behavioural Intention to Use AI in future teaching career
(identified in the TAM);
● independent variables: I1 – Use of AI up to now; I3 – I have sufficient knowledge and
can explain what AI means; I4 – I know how to exploit the benefits of AI in
school/kindergarten; I6 – Pupils appreciate more the teachers who use AI; I8 – AI changes
the role of the educator; I10 – AI will help improve student-teacher interaction; I14 –
I consider myself able to teach with AI (identified in TAM as ease of use); I15 – A future

AE
teacher needs to master and use AI; I16 – I consider there are more advantages to AI than
disadvantages (identified in TAM as perceived usefulness).
The study of behavioural intention to use AI is of particular importance in the TAM model
because it is the precursor to actual use (Davis, 1989). As can be seen in Table 2, for both
constructs (attitude toward AI use and perception of AI usefulness), Cronbach Alpha
coefficient (α) > 0.700, Average Variance Extracted (AVE) > 0.500, and Composite
Reliability (CR) > 0.700, demonstrating an appropriate level of internal consistency between
the variables used in the study (Henseler and Sarstedt, 2013; Nemțanu et al., 2021).
Table no. 2. Validity of the data used in the questionnaire
α / AVE /
Construct Variables Loading
CR
Attitude I3 – I have sufficient knowledge and can explain what 0.764 α: 0.737
towards AI means AVE: 0.504
the use of AI I4 – I know how to exploit the benefits of AI in school/ 0.860 CR: 0.826
kindergarten
I6 – Pupils appreciate more the teachers who use AI 0.764
I8 – AI changes the role of the educator 0.332
I10 – AI will help to improve pupil-teacher interaction 0.712
Perception of I1 – Use of AI up to now 0.179 α: 0.744
AI usefulness I14 – I consider myself able to teach using AI 0.751 AVE: 0.533
I15 – A future teacher must master and use AI 0.874 CR: 0.833
I16 – I think there are more advantages to AI than 0.850
disadvantages
I17 – Willingness to use AI in future teaching career 0.764
Source: The authors’ own calculations
Several hypotheses have been established on the basis of the papers published to date. The
research conducted by Labrague et al. (2023) shows that previous exposure to AI
technologies, knowledge about AI and competence in using AI lead to a positive perception
and implicitly to a greater willingness to incorporate AI into subsequent practices. On the
basis of this finding, the following two hypotheses are formulated:
H1: The previous use of AI significantly and positively influences the intention of
behavioural use of AI in the future teaching career.
H2: Knowledge of AI significantly and positively influences the behavioural intention to use
AI in future teaching careers.
Furthermore, the study by Ali (2017) supports the idea that teachers’ willingness to use AI is
directly influenced by what their students want and need. Therefore, teachers exposed to the
enthusiasm of students and their positive experiences of using AI led to increased chances of
incorporating AI into teaching methods, resulting in the formulation of the following
hypothesis:
H3: Students’ expectations and requirement to use new AI technologies significantly and
positively influence the behavioural intention to use AI in future teaching career.
It is interesting to note how some teachers understand their role in helping students use
technology, as well as how they view the use of AI in their own work (Ali, 2017).
Specifically, teachers will play a key role in ensuring that AI is used correctly and ethically
Vol. 26 • No. 65 • February 2024 151

to measure and assess student performance, that the data obtained is accurate, reliable, and
true (Owan et al., 2023). All these observations led us to formulate the fourth hypothesis:
H4: The perception of the change brought about by AI in the educator’s role significantly
and positively influences the behavioural intention to use AI in the future teaching career.
The use of AI is slowed by external barriers (lack of easy access to hardware, software, tools
and training) and internal barriers (lack of trust, negative beliefs and attitudes about using AI
in school), thus, beliefs, confidence and attitude significantly influence the teachers’ decision
to use AI in their work (Rowston, Bower and Woodcock, 2022). The following hypothesis
thus emerges:
H5: The perception of own ability /confidence to use AI in teaching significantly and
positively influences the behavioural intention to use AI in the future teaching career.
A survey of 399 Hong Kong university students found that perceptions of the usefulness and
willingness to use GAI technology increased as they identified its benefits in learning, writing
and studying activities (Chan and Hu, 2023). This finding prompts the formulation of the
following hypothesis:
H6: The perception of more advantages than disadvantages of AI in education (perceived
usefulness) significantly and positively influences the behavioural intention to use AI in a
future teaching career.

3.1. Degree of familiarity of future teachers with AI
Upon analysis of the data, it clearly appears that up to the time of the research, 77.41% of the
students enrolled in teacher education programmes have heard about AI, 27.04% have further
researched AI, and only 20% used AI. To understand to what extent the use of AI contributes
to changing perceptions, the results of the two groups were analysed: Group 1 – those who
have actually used AI; and Group 2 – those who did not use AI, although they have heard or
read about this technology. Thus, 59.26% of the respondents included in Group 1 used AI for
educational purposes, 48.15% out of curiosity, or 37.04% for various work or even
entertainment tasks. The low rate of use of AI is explained by the early stage of development
of the technology.
The survey respondents rate their knowledge and understanding of AI as average, with higher
values for Group 1 – those who have previously used this technology, as shown in Table 3.
It is important to note that, regardless of the group, there is a high interest in improving the
use of AI in the teaching process, with more than 76% of the respondents being interested in
attending training courses on this topic.
Table no. 3. Level of knowledge of AI
Group 1 – Group 2 –
Category of results
Used AI Did not use AI
How well can they define/explain what AI entails 3.12 2.85
I know how to take advantage of AI in school/kindergarten 3.12 2.64
Note: arithmetic weighted average on a scale from: 1 – to a very small extent; 5 – to a very large extent

AE
3.2. The change brought by AI in the educator’s role

When it comes to the changes that AI brings to education, about half of the respondents,
regardless of which group they belong to, are sure that AI will be used as a teaching tool in
the future. An example in this regard is identified in China, where the collaboration of
60.000 schools has helped to implement a system of automatic essay marking, achieving a
remarkable 92% accuracy rate, comparable to human assessment (UNESCO, 2019a).
The computer uses an AI neural network and deep learning algorithms to compare student
essays against human assessment.
That activity is followed, according to Table no. 4, by using AI as a virtual teacher assistant,
being a possibility among 30% of those who used AI. Respondents in Group 2 – those who
did not use AI – believe that AI, more than being a virtual assistant, will help personalise the
learning experience (27%). The percentage of those who believe that AI will not change the
educational process is insignificant, from 6% to 7%. However, it is interesting to note the
difference between the two groups when it comes to the possibility that AI will replace the
teacher. More specifically, 5% of those who did not use AI (Group 2) perceive AI as a
possible threat and are considering the scenario of AI replacing the teacher, while it should
be noted that none of the Group 1 respondents, those who used AI, considers this a plausible
scenario. This result is also supported by the study by Edwards and Cheok (2018), which
shows that machines lack the necessary capabilities for social interaction and emotional
communication with students.
Table no. 4. Changes brought by AI in education and the role of an educator
(% of total respondents)
Category of results
Used AI Did not use AI
Changes brought by AI to education
- Teachers will use AI as a teaching tool 43% 46%
- AI will be the virtual assistant of the teacher 30% 16%
- AI will be used by learners to customise the learning experience 20% 27%
- AI will not change the educational process 7% 6%
- AI will replace the teacher 0% 5%
Changes brought by AI to the educator’s role
- shift from traditional instruction to facilitation and mentoring 43% 38%
- shift from general and delayed feedback to instant and 39% 22%
personalised feedback
- shift from mass learning to personalised learning 37% 25%
- instant lesson design at the time of teaching, without pre- 31% 25%
planning 26% 19%
- extension of learning time outside the classroom 17% 12%
- relief from administrative tasks (e.g. parent feedback,
centralisation of various situations)
To better understand how exactly the educator’s role will change with the integration of AI
into education, the main estimates of the respondents are presented in Table 4. There is a
common perception between the two groups regarding the ability of AI to transform the role
of the educator from traditional instruction and knowledge delivery to that of facilitator and
mentor for students in the learning process. It is also estimated, especially among those who
have used AI before, that there is an increase in the speed of response in terms of feedback
Vol. 26 • No. 65 • February 2024 153

to students and personalisation. At the same time, AI will help relieve some of the routine
tasks of the teacher (e.g., administrative tasks of centralising documents and results) and will
be able to extend learning time by answering questions from students outside of the
classroom.
The acceleration of the process of change, as shown in Table no. 5, can occur due to student
demands. Thus, there is a common perception between the two groups regarding the
expectations of preschool and schoolchildren of teachers to use AI (with an average of 3.87
and 3.49). This makes the integration of AI a necessity for future teachers to adapt to the
requirements and maintain the interest of the students in teaching activities. Other possible
benefits of using AI in education include improving student-teacher interaction, enhancing
the preschooler/schooler learning experience, and transitioning to personalised education.
Table no. 5. Reasons, causes and effects for/of using AI in education
Group 2 -
Group 1 -
Category of results Did not
Used AI
use AI
Reasons
Preschoolers or schoolchildren value the teacher who applies AI 3.87 3.49
AI will help to improve pupil-teacher interaction 3.31 3.13
AI will contribute to improving the learning experience of the pre- 3.24 3.32
schooler / schoolchild and transitioning to personalised education.
I believe that there are more future advantages generated by AI than 3.35 3.23
there are disadvantages
Causes and effects
I believe that a future teacher should master and use AI in teaching 3.77 3.41
I consider myself capable of teaching using artificial intelligence. 3.00 2.79
I can see myself using AI in my future profession 3.62 3.29
Note: arithmetic weighted average on a scale from: 1 – to a very small extent; 5 – to a very large extent
Regardless of the group to which they belong, respondents agree that a future teacher should
master and use AI in his teaching activity. The fact that the values recorded for the current
ability to use AI in teaching activity (3.00 and 2.79) are lower than the predicted use of AI in
their future activities (3.62 and 3.29) may indicate a tendency to openness and confidence in
the ability of AI technology and its usefulness in future teaching activity. According to the
recorded values, there is a non-significant difference between the intention of using AI in
future profession between the two groups (average of 3.62 of those who used AI vs. 3.29 of
those who did not use AI). This shows that future teachers are eager to implement elements
of AI in the teaching-learning process, even when they have had less exposure to such
technology and belong to generations with different digital skills.
A more detailed understanding of how to integrate AI into teacher work is presented in Table
no. 6. Here again there is a common awareness of the two groups when it comes to the need
to use AI in generating educational content and presenting it in the most interactive way for
pre-schoolers and schoolchildren. AI is also expected to be a real support in personalising the
learning experience of students and can be used as a virtual assistant available 24/7. This is
consistent with the finding of Kim et al. (2020) that the use of AI will work, especially when
face-to-face interaction is not possible.

AE
Table no. 6. Ways of using AI in education (% of total respondents)

Category of results Used AI Did not
use AI
- AI can create intelligent/interactive content and boost student 69% 50%
engagement
- AI can help personalise the learning experience 52% 42%
- AI can help as a ready-to-use virtual assistant 24/7 41% 27%
- AI can automate administrative processes 37% 21%
- Based on the collected data, AI can be used to update curriculum 35% 11%
and instructional methods
- AI can automate the assessment process 20% 13%
Significant differences can also be observed between the two groups regarding the use of AI
in updating curriculum and training methods, a scenario agreed by 35% of Group 1
respondents – who used AI vs. 11% of Group 2 respondents – who did not use AI. This
phenomenon can be explained by the ability of the respondents in Group 1 to understand the
power of AI to use and interpret large amounts of data. What is also interesting to note are
the differences between the groups in terms of the reluctance to involve AI in the evaluation
process, particularly for the respondents in Group 2. These results are similar to those
recorded by Doumat et al. (2022) who claim that only 26% of students “want to be assessed
with AI”. Samarescu (2021) presents a different perspective, showing that the use of AI in
the assessment process would be beneficial, as it could provide detailed information and
personalised feedback to students. The respondents were also given the opportunity to add to
these usage options and complete them, by answering the open question “What is the most
difficult task AI can help you?”. Therefore, a large proportion of respondents see the
usefulness of AI in “designing teaching activities”, followed by reducing teachers
“administrative tasks” by automating processes that currently take up a lot of time and would
allow teachers to focus more on the teaching process. The respondents believe that AI could
help to “capture and hold the attention of students for a long time”, making the educational
process more engaging and dynamic. Last but not least, respondents see value in integrating
AI into “developing and managing online courses”.
3.3. Future teachers’ behavioural intention to use AI
More than the ways in which AI can contribute to improving teaching activities, the study
aims – by means of a binary logistic regression – to explore the future teachers' behavioural
intention to use AI and to identify its determining factors. Therefore, to identify the regression
equation that predicts the extent to which future teachers are willing or not to use AI in their
teaching process, the dependent variable I17 (Behavioural intention to use AI in future career)
was considered, which registered the value of 1 (Yes) if the answer was positive and 0 (No)
if the respondents did not intend to use AI in the future. Compared to the baseline version –
Model 0 – where the correct estimation percentage was 80.7%, Model 1 includes the
independent variables established in the research methodology and shows an improvement
in accuracy, explaining the variation of the dependent variable in 89.3% of cases.
To validate the model, several tests were conducted, and certain parameters were checked.
Thus, the initial regression model, Model 0, constructed using only the constant, is found to
be valid (B=1.433; S.E.=0.154; Wald=86.246; Sig.=0.000; Exp(B)=4.192). The Chi-square
Vol. 26 • No. 65 • February 2024 155

test shows that Model 1 is improved over the baseline model and is thus statistically
significant (Sig.=0.000; Chi-square=120.206; df=6), and the value recorded -2 Log
likelihood=144.373 and Cox & Snell R Square=0.359. The value recorded by Nagelkerke
R2=0.575 shows that 57.5% of the predictors included in Model 1 explain the variation in
the dependent variable (I17 - behavioural intention to use AI in the future teaching career).
The last test carried out to verify the model is Hosmer and Lemeshow, and the recorded p-
value > 0.05 confirms the validity and explanatory power of Model 1 (Chi-square=14.960;
Sig.=0.60; df=8).
Table no. 7 shows the results of Model 1 regarding the likelihood that a future teacher will
use AI in his/her future career. Thus, two predictors with a significant level of influence on
the preference to use AI in the future can be observed, namely I14 and I16. The values
confirm that self-confidence about the ability to use AI (perceived ease of use) and the
prospect of more advantages of AI compared to possible disadvantages (perceived
usefulness) significantly increase the likelihood that a future teacher will use AI in his/her
career. Interestingly, other factors such as previous use of the technology (I1), background
knowledge about AI (I3), the requirements of the students to use these technologies (I6) or
perception of the changing role of the educator (I8) do not influence the future desire to use
AI in the educational process.
Table no. 7. Variables in equation for Model 1
95% C.I. pt EXP(B
B S.E. Wald df Sig. Exp(B)
Inferior Superior
S I1. Use of AI 1.421 .769 3.417 1 .065 4.141 .918 18.682
T I3. Knowledge of AI -.005 .220 .001 1 .980 .995 .646 1.530
E I6. Required by students .321 .208 2.379 1 .123 1.378 .917 2.073
P I8. Change of .141 .217 .420 1 .517 1.151 .753 1.760
1 educator’s role
I14. Able to teach using AI .716 .249 8.285 1 .004 2.046 1.257 3.331
I16. More advantages than 1.684 .291 33.537 1 .000 5.387 3.047 9.526
disadvantages
Constant -6.526 1.149 32.288 1 .000 .001
a. Variable(s) introduced for step 1: I1, I3, I6, I8, I14, I16.
The results recorded by I14 (B=0.716; Exp(B)=2.046; Sig.=004) contribute to validate
hypothesis H5: The perception of your own ability / confidence in using AI in teaching
significantly and positively influences the desire to use AI in a future teaching career. Thus,
a person who considers himself / herself able to teach with AI will be twice as determined to
use AI in the educational process. Similarly, the values recorded by I16 (B=1.684;
Exp(B)=5.387; Sig.=0.000) lead to the validation of hypothesis H6: Perceiving more
advantages than disadvantages of AI in education (perceived usefulness) significantly and
positively influences the behavioural intention to use AI in the future teaching career. This
highlights the importance of also presenting the perception of using AI in a positive sense, as
accurately as possible, from the very beginning. This feature increases the chances of a future
teacher to integrate AI technology into his/her work fivefold. These results are also confirmed
by Al Darayseh’s (2023) research, which states that behavioural intentions are determined by
two types of factors: “expected benefits” and “ease of use”. The values recorded for variables
I1. Use of AI (B=1.421; Exp(B)=4.141; Sig.=0.065); I6. Required by the students (B=0.321;
Exp(B)=1.378; Sig.=0.123); I8. The change in the educator's role (B=0.141; Exp(B)=1.151;

AE
Sig.=0.517) suggest a positive relationship between these and the dependent variable I17.
Behavioural intention to use AI in future career, yet the relationships between them are not
significant, marking a value of Sig. > 0.05. As a result, the following hypotheses are partially
validated: H1: Previous use of AI significantly and positively influences the willingness to
use AI in future teaching career; H3: Students’ expectations and requirement to use new AI
technologies significantly and positively influence the behavioural intention to use AI in
future teaching career; H4: Perception of the change brought about by AI in the educator’s
role significantly and positively influences the behavioural intention to use AI in future
teaching career.
Knowledge of AI and the value recorded by the variable I3. The knowledge of AI (B = -
0.005; Exp(B)=0.995; Sig.=0.980) prove that, in the case of the sample, the mere
accumulation of knowledge of AI had no positive effect on the desire to use AI in a future
teaching career. This invalidates H2: Knowledge of AI significantly and positively influences
the behavioural intention to use AI in the future teaching career. Therefore, the above
recorded result confirms that regardless of the level of knowledge or digital skills, there is a
desire to improve and use AI in the educational process given the many benefits this
technology can generate.
Conclusions
AI can complement and enhance the role of educators, ultimately contributing to more
effective personalised learning experiences. The results of the present study, conducted on
270 future teachers, emphasise the fact that although 77% of them have heard of AI, only
20% have actually used AI to date, mainly for personal educational purposes. There is high
interest among 76% of the respondents in taking further training in the field of AI, showing
an understanding of the potential AI has for educational development.
On the contrary, the small percentage (5%) of those who did not use AI (Group 2) to date
and perceive this technology as a threat should be noted, considering the scenario where the
teacher is replaced by AI. The vast majority of the respondents value the use of AI as a
teaching tool or virtual teacher assistant to enhance the student learning experience.
Although currently only 62.96% of the respondents consider themselves able to use AI in
their future teaching work, it is noteworthy that 81.48% of them are convinced of the need to
master and use AI technology in their future teaching career. Among the most important
benefits identified by respondents in using AI are its use to create content and increase the
level of interactivity for students, extend learning time and use AI as a virtual teacher
assistant, as well as automating some administrative and evaluation processes. This will
change the role of the teacher and accelerate the transition from traditional instruction and
knowledge transfer to the role of facilitator and mentor for students in the learning process,
as well as the transition from general to personalised instantaneous feedback. In the future,
teachers are likely to do less administrative work and focus on teaching while receiving real-
time suggestions to improve teaching methods and increase audience interest.
The regression model shows that the behavioural intention to use AI in the future teaching
career is significantly and positively influenced by the perceived confidence in using AI
(perceived ease of use) and assigning a positive role to AI by identifying more advantages
than disadvantages (perceived usefulness). Examples of this could be the use of AI in
Vol. 26 • No. 65 • February 2024 157

automatic writing assessment or giving teachers real-time feedback during presentations to

improve the student experience. Surprisingly, these predictors seem more important than
previous use of AI; student requirements and expectations of technology use; awareness of
the educator’s changing role as more and more AI elements are integrated into education.
The results show that, regardless of the level of knowledge about AI or the digital skills
possessed, there is a desire for further development and future use of AI in education. This
reinforces the firm belief that the future of education will be fundamentally improved with
the help of AI, and the time required to transition to this process is largely influenced by the
self-confidence of teachers and the need for more and more positive examples in interacting
with the said technology.
Theoretical contributions: This paper brings together and tests in a single regression model
several predictors of behavioural intention to use AI in education and identifies those with
the greatest impact. According to TAM, this intention is considered a precursor to actual use
of the technology, enhancing the relevance of researching it during the training stage. The
study provides valuable information that can lead to changes in attitude and behaviour in
favour of adopting AI in education in the future.
Management implications: Research highlights the importance of confidence in the ability to
use AI and understanding of the benefits that technology has to stimulate future intention to
use AI in education. This finding can form the basis for improving training programmes and
can contribute to the development of educational policies that accelerate the process of
integrating AI into teaching in Romania.
Limitations: This exploratory research covers a limited geographical area and a small number
of respondents to allow for a representative random sample. The still early stage of
development and integration of AI in education, and the insufficient research on this topic,
make it difficult to identify predictors of the intention of behavioural use of AI. The
regression model used partially describes the dependent variable, with only two significant
predictors identified.
Prospects: Integrating qualitative research would provide more information on motivations
and behavioural intentions to use AI in teaching careers. Further studies could identify and
analyse more factors that influence the intention to use AI. It would also be of interest to
identify a group of active AI users and thus understand how to integrate AIEd. As technology
advances and the educational context changes, future research is essential to update the
perceptions and factors that influence the use of AI in education.
References
Al Darayseh, A., 2023. Acceptance of artificial intelligence in teaching science: Science
teachers’ perspective. Computers and Education: Artificial Intelligence, 4, article no.
Ali, N., 2017. The influence of technology on the academic and social lives of students and
lecturers in Kuwaiti higher education (HE). [online] College of Basic Education.
Available at: <https://www.proquest.com/dissertations-theses/influence-technology-on-
academic-social-lives/docview/2083753176/se-2> [Accessed 26 July 2023].
Bell, J. and Waters, S., 2018. Doing your research project. New York: McGraw-Hill.

AE
Bonsu, E. and Baffour-Koduah, D., 2023. From the Consumers’ Side: Determining Students’
Perception and Intention to Use ChatGPT in Ghanaian Higher Education. SSRN
Electronic Journal. https://doi.org/10.2139/ssrn.4387107.
Chan, C.K.Y. and Hu, W., 2023. Students’ voices on generative AI: perceptions, benefits,
and challenges in higher education. International Journal of Educational Technology in
Higher Education, 20(1), article no. 43. https://doi.org/10.1186/s41239-023-00411-8.
Chan, C.K.Y. and Lee, K.K.W., 2023. The AI generation gap: Are Gen Z students more
interested in adopting generative AI such as ChatGPT in teaching and learning than their
Gen X and millennial generation teachers? Smart Learning Environments, 10(1), article no.
60. https://doi.org/10.1186/s40561-023-00269-3.
Chao, P.-J., Hsu, T.-H., Liu, T.-P. and Cheng, Y.-H., 2021. Knowledge of and Competence
in Artificial Intelligence: Perspectives of Vietnamese Digital-Native Students. IEEE
Choi, S., Jang, Y. and Kim, H., 2023. Influence of Pedagogical Beliefs and Perceived Trust
on Teachers’ Acceptance of Educational Artificial Intelligence Tools. International
Journal of Human – Computer Interaction, 39(4), pp. 910-922. https://doi.org/10.1080/
10447318.2022.2049145.
Chounta, I.-A., Bardone, E., Raudsep, A. and Pedaste, M., 2022. Exploring Teachers’
Perceptions of Artificial Intelligence as a Tool to Support their Practice in Estonian K-12
Education. International Journal of Artificial Intelligence in Education, 32(3),
pp. 725-755. https://doi.org/10.1007/s40593-021-00243-5.
Dalalah, D. and Dalalah, O.M.A., 2023. The false positives and false negatives of generative
AI detection tools in education and academic research: The case of ChatGPT. The
International Journal of Management Education, 21(2), article no. 100822.
https://doi.org/10.1016/j.ijme.2023.100822.
Davis, F.D., 1989. Perceived Usefulness, Perceived Ease of Use, and User Acceptance of
Information Technology. MIS Quarterly, 13(3), article no. 319. https://doi.org/10.2307/
249008.
Doumat, G., Daher, D., Ghanem, N.-N. and Khater, B., 2022. Knowledge and attitudes of
medical students in Lebanon toward artificial intelligence: A national survey study.
Frontiers in Artificial Intelligence, 5. https://doi.org/10.3389/frai.2022.1015418.
Edwards, B.I. and Cheok, A.D., 2018. Why Not Robot Teachers: Artificial Intelligence for
Addressing Teacher Shortage. Applied Artificial Intelligence, 32(4), pp. 345-360.
https://doi.org/10.1080/08839514.2018.1464286.
Fahrman, B., Norström, P., Gumaelius, L. and Skogh, I.-B., 2020. Experienced technology
teachers’ teaching practices. International Journal of Technology and Design Education,
30(1), pp. 163-186. https://doi.org/10.1007/s10798-019-09494-9.
Hasib, K.Md., Rahman, F., Hasnat, R. and Alam, Md.G.R., 2022. A Machine Learning and
Explainable AI Approach for Predicting Secondary School Student Performance.
In: 2022 IEEE 12th Annual Computing and Communication Workshop and Conference
(CCWC). IEEE. pp. 0399-0405. https://doi.org/10.1109/CCWC54503.2022.9720806.
Henseler, J., Sarstedt, M., 2013 Goodness-of-fit indices for partial least squares path
modeling. Computational Statistics, 28, pp.565-580. https://doi.org/10.1007/s00180-012-
0317-1.
Vol. 26 • No. 65 • February 2024 159

Holmes, W., Porayska-Pomsta, K., Holstein, K., Sutherland, E., Baker, T., Shum, S.B.,
Santos, O.C., Rodrigo, M.T., Cukurova, M., Bittencourt, I.I. and Koedinger, K.R., 2022.
Ethics of AI in Education: Towards a Community-Wide Framework. International
Journal of Artificial Intelligence in Education, 32(3), pp.504-526. https://doi.org/
10.1007/ s40593-021-00239-1.
Jurconi, A., Iacobescu, I.M.P., Manea, D.-I., Mihai, M. and Pamfilie, R., 2022. The Impact
of the Green Transition in the Field of Food Packaging on the Behavior of Romanian
Consumers. www.amfiteatrueconomic.ro, 24(60), article no. 395. https://doi.org/
10.24818/EA/ 2022/60/395.
Kairu, C., 2023. Students’ Attitude Towards the Use of Artificial Intelligence and Machine
Learning to Measure Classroom Engagement Activities. In: Proceedings of EdMedia +
Innovate Learning. [online] The Netherlands: Association for the Advancement of
Computing in Education (AACE). pp. 793-802. Available at:
<https://www.learntechlib.org/primary/p/217382/> [Accessed 15 July 2023].
Kaplan-Rakowski, R., Grotewold, K., Hartwick, P. and Papin, K., 2023. Generative AI and
Teachers’ Perspectives on Its Implementation in Education. Journal of Interactive
Learning Research, 34(2), pp. 313-338.
Kim, J., Merrill, K., Xu, K. and Sellnow, D.D., 2020. My Teacher Is a Machine:
Understanding Students’ Perceptions of AI Teaching Assistants in Online Education.
International Journal of Human–Computer Interaction, 36(20), pp. 1902-1911.
https://doi.org/10.1080/10447318.2020.1801227.
Kim, Y., Soyata, T. and Behnagh, R.F., 2018. Towards Emotionally Aware AI Smart
Classroom: Current Issues and Directions for Engineering and Education. IEEE Access,
6, pp. 5308-5331. https://doi.org/10.1109/ACCESS.2018.2791861.
Labrague, L.J., Aguilar-Rosales, R., Yboa, B.C. and Sabio, J.B., 2023. Factors influencing
student nurses’ readiness to adopt artificial intelligence (AI) in their studies and their
perceived barriers to accessing AI technology: A cross-sectional study. Nurse Education
Today, 130, article no. 105945. https://doi.org/10.1016/j.nedt.2023.105945.
Leong, M.K. and Chaichi, K., 2021. The Adoption of Technology Acceptance Model (TAM)
and Trust in Influencing Online Purchase Intention During the Covid-19 Pandemic:
Empirical Evidence from Malaysia. International Journal of Academic Research in
Business and Social Sciences, 11(8), pp. 468-478. https://doi.org/10.6007/IJARBSS/v11-
i8/10422.
Limna, P., Kraiwanit, T. and Jangjarat, K., 2023. Adopting the technology acceptance model
(TAM) to explore online purchase intention via Facebook live streaming: Empirical
evidence from Bangkok, Thailand. ASEAN Journal of Management & Innovation, 10(1),
pp. 1-13.
McCarthy, J., 1998. What is artificial intelligence. [online] Available at:
<https://api.semanticscholar.org/CorpusID:261070249> [Accessed 12 June 2023].
McCulloch, W.S. and Pitts, W., 1943. A logical calculus of the ideas immanent in nervous
activity. The Bulletin of Mathematical Biophysics, 5(4), pp. 115-133. https://doi.org/
10.1007/BF02478259.
Moorhouse, B. and Kohnke, L., 2023. The Effects of Generative Ai on Initial Language
Teacher Education: The Perspectives of Teacher Educators. [online] SSRN. Available
at: <https://ssrn.com/abstract=4532479> [Accessed 16 August 2023].

AE
Owan, V.J., Abang, K.B., Idika, D.O., Etta, E.O. and Bassey, B.A., 2023. Exploring the
potential of artificial intelligence tools in educational measurement and assessment.
Eurasia Journal of Mathematics, Science and Technology Education, 19(8), article no.
em2307. https://doi.org/10.29333/ejmste/13428.
Păvăloaia, V.-D. and Necula, S.-C., 2023. Artificial Intelligence as a Disruptive Technology
– A Systematic Literature Review. Electronics, 12(5), article no. 1102. https://doi.org/
10.3390/electronics12051102.
Societies, 13(5), article no. 118. https://doi.org/10.3390/soc13050118.
Ravi Kumar, V. V and Raman, R., 2022. Student Perceptions on Artificial Intelligence (AI)
in higher education. In: 2022 IEEE Integrated STEM Education Conference (ISEC).
IEEE. pp. 450-454. https://doi.org/10.1109/ISEC54952.2022.10025165.
Rowston, K., Bower, M. and Woodcock, S., 2022. The impact of prior occupations and initial
teacher education on post-graduate pre-service teachers’ conceptualization and
realization of technology integration. International Journal of Technology and Design
Education, 32(5), pp. 2631-2669. https://doi.org/10.1007/s10798-021-09710-5.
Samarescu, N., 2021. Objective online evaluation. eLSE 2021, pp .546-551.
https://doi.org/10.12753/2066-026X-21-137.
Singh, H. and Miah, S.J., 2019. Design of a mobile-based learning management system for
incorporating employment demands: Case context of an Australian University. Education
and Information Technologies, 24(2), pp. 995-1014. https://doi.org/10.1007/s10639-018-
9816-1.
Sit, C., Srinivasan, R., Amlani, A., Muthuswamy, K., Azam, A., Monzon, L. and Poon, D.S.,
2020. Attitudes and perceptions of UK medical students towards artificial intelligence
and radiology: a multicentre survey. Insights into Imaging, 11(1), article no. 14.
https://doi.org/10.1186/s13244-019-0830-7.
UNESCO, 2019a. Artificial Intelligence in Education: Challenges and Opportunities for
Sustainable Development. [online] Paris. Available at: <https://unesdoc.unesco.org/
ark:/48223/pf0000366994> [Accessed 12 September 2023].
UNESCO, 2019b. Beijing Consensus on Artificial Intelligence and Education, Outcome
document of the International Conference on Artificial Intelligence and Education
‘Planning education in the AI era: Lead the leap. [online] Available at:
<https://unesdoc.unesco.org/ ark:/48223/pf0000368303> [Accessed 12 May 2023].
Williamson, B. and Eynon, R., 2020. Historical threads, missing links, and future directions
in AI in education. Learning, Media and Technology, 45(3), pp. 223-235.
https://doi.org/10.1080/17439884.2020.1798995.
Woolf, B.P., Lane, H.C., Chaudhri, V.K. and Kolodner, J.L., 2013. AI Grand Challenges for
Education. AI Magazine, 34(4), pp. 66-84. https://doi.org/10.1609/aimag.v34i4.2490.
Zou, B., Liviero, S., Hao, M. and Wei, C., 2020. Artificial Intelligence Technology for EAP
Speaking Skills: Student Perceptions of Opportunities and Challenges. pp. 433-463.
https://doi.org/10.1007/978-3-030-34212-8_17.
Vol. 26 • No. 65 • February 2024 161

AE Generation Z Students’ Perceptions on the Abilities, Skills
and Competencies Required in the Age of Artificial Intelligence Systems
GENERATION Z STUDENTS’ PERCEPTIONS ON THE ABILITIES, SKILLS

AND COMPETENCIES REQUIRED IN THE AGE OF ARTIFICIAL
INTELLIGENCE SYSTEMS
Sorin George Toma1 and Oana Simona Hudea2
1)2)
University of Bucharest, Bucharest, Romania

Toma, S.G. and Hudea, O.S., 2024. Generation Z Received: 30 September 2023
Students’ Perceptions on the Abilities, Skills and Revised: 27 November 2023
Competencies Required in the Age of Artificial Accepted: 27 December 2023
Intelligence Systems. Amfiteatru Economic, 26(65),
pp. 162-180.
Abstract
The knowledge economy promotes the use of information and communication technology
and highly skilled labour on a large scale. The past decades have witnessed a continuous
fall in the importance of unskilled sectors and a rise in skilled sectors combined with the
emphasis on skills and competencies. Being highly familiar with the use of technology,
Generation Z students are very interested in raising their chances of employability in the
current digital society. The purpose of this research was to identify and analyse
Generation Z students’ perceptions on the abilities, skills and competencies required in
the age of artificial intelligence systems. In this respect, the authors used a quantitative
research method through an online survey. A total number of 352 questionnaires from
students who graduated from full-time undergraduate programs within a Romanian
faculty were validated. The data were processed using the SPSS 17.0 statistical software,
the subsequent viability tests and analyses involving the use of Cronbach's Alpha
coefficient, the Principal Component Analysis, respectively, the correlational analysis.
From a theoretical point of view, the paper contributes to the enrichment of the scientific
literature regarding the new relationships between human capital and the era of artificial
intelligence systems, offering a new perspective on the structure of the needed abilities,
skills, and competencies in the digital society. From a practical point of view, the results
of the study provide solutions for improving the educational process by adapting
academic programs to the demands of the digital society. The results of the research show
that, on the one hand, sensorial, cognitive, and psychomotor abilities, and, on the other
 Corresponding author, Sorin George Toma – e-mail: [email protected]

AE
hand, general knowledge, business and management, and technical skills and
competencies, are playing a key role in today’s era of artificial intelligence systems.
Key words: students, Generation Z, abilities, skills, competencies, artificial intelligence.
JEL Classification: I20, I23, J24.
Introduction
In their long evolution, humans have made a fundamental step from being mostly
illiterate to being educated in a significant proportion. Education, in general, and higher
education, in particular, have proved to be critical pillars of economic growth and
sustainable development throughout the world. Ensuring inclusive and equitable quality
education for all represents one of the Sustainable Development Goals adopted by the
United Nations in its 2030 Agenda for Sustainable Development (UN, 2015). Thus,
education constitutes an important matter for any country in terms of both quantity and
quality. In this respect, education should not only be adequate but also fulfil the needs of
various stakeholders, such as society, students, organisations, or markets. This is why the
aim of education in the 21st century is to prepare its graduates for life, work, and
citizenship.
The roots of the competence-based movement are to be found in the 1920s, but the
demand for higher accountability in education emerged only in the 1960s in the United
States of America (USA), when an increased emphasis on economic issues was
ascertained (Tuxworth, 1990). The assessment criteria and the outcomes were more and
more linked to the competencies acquired during the years of studies. It was the period
when the so-called “information wave” dramatically impacted humanity and provoked the
birth of the information economy (Toffler, 1980)/digital economy (Tapscott, 2014) or the
knowledge-based economy (Drucker, 1969). Characterised not only by a deeper reliance
on knowledge and information (Aparicio, Iturralde and Rodriguez, 2023), but also on
intellectual capabilities (Powell and Snellman, 2004), this knowledge economy promotes
the use of information and communication technology (ICT) and highly skilled labour on
a large scale. Its expansion was facilitated by the collapse of the communist regime at the
end of the 1980s and the beginning of the 1990s when the globalisation process rapidly
spread and led to the liberalisation of markets and the free movement of people, capital,
and information all over the world. As a result, the labour markets have changed into
global ones and have begun to accurately match workers to jobs provided in a specific
industry (Davidson et al., 2010) as their needed skills and competencies have become
somehow standardised.
From these radical transformations that altered the nature and type of work, the role of
human capital has become much more prominent in any organisation, irrespective of its
size and sector. As more and more skills are computer-based, the ICT revolution favours
knowledge workers because their intellect represents today’s primary resource (J. Hope
and T. Hope, 1997). The past decades have witnessed a continuous fall in unskilled
sectors and a rise in skilled sectors (Jaumotte and Tytell, 2007) combined with the shift
from jobs to skills (Deloitte, 2022). Consequently, the skills and competencies-focused
Vol. 26 • No. 65 • February 2024 163

approach has become the rule for an increasing number of organisations in the age of
automation and artificial intelligence (AI) systems.
The existing rift between people’s capabilities and employers’ skill requirements
represents a major challenge for education. The way education succeeds in providing the
needed abilities, skills, and competencies to pupils and students in order to cope with
different trends (e.g., social, economic, political, environmental, technological) and
respond to multiple risks (e.g., crises, pandemics) is decisive for its future. Innovating and
adapting to the new realities of the labour market are important prerequisites for
education to meet the demand for skilled people in an increasingly digitalised society.
Being highly familiar with the use of ICT and eager to embrace technological
advancements, Generation Z (Gen Z) students are very interested in raising their chances
of employability in the age of AI by obtaining as many as possible skills and competences
during the years of studies.
Gen Z students' perceptions of the skills and competencies needed in the digital society
have begun to be studied in recent years by various researchers (Benítez-Márquez et al.,
2022). However, the authors identified a research gap regarding the existing connection
among the three above-mentioned research areas- Gen Z students, abilities, skills, and
competencies, and AI, that constituted the basis of the literature review. The purpose of
this research is to identify and analyse Gen Z students’ perceptions on the abilities, skills
and competencies required in the age of AI systems. In this regard, the authors used a
quantitative research method based on an online survey. This study seeks to contribute to
the development of the literature by answering the two research questions and providing
other possible directions for researchers. The results show that Gen Z graduates of
business and administration specialisations are highly aware of the importance of the new
abilities, skills, and competencies specific to the digital society. Also, they highlight the
key role played by some types of abilities (e.g., sensorial, cognitive), skills, and
competencies (e.g., general knowledge, business, and management) in their opinion. The
originality of the paper lies in the new perspective, integrating previous approaches, that it
provides on Gen Z students' perception of the skills, abilities, and competencies required
by the age of AI systems and how education meets this challenge. The structure of the
paper is as follows: the next section displays the literature review. The research
methodology is presented in the second section. The results of the study are exhibited and
analysed in the third section of the paper. The conclusions are illustrated in the last
section of the paper.
This section of the study shows the theoretical framework that seeks to clarify the
meaning in the key concepts of the paper and reveal their relationship. The scientific
literature related to the areas of human resources management, education, and economics
provides numerous definitions of the concepts of Gen Z students, abilities, skills, and
competencies, and the age of AI. As the first generation born in an increasingly digital
society and deeply shaped by technological disruptions, Gen Z or iGen encompasses the
people born in the period 1997-2012 (Dimock, 2019). Most of today’s students are
representatives of Gen Z. Gen Z students share several common characteristics as follows
(Seemiller and Grace, 2016; Cho, Bonn and Han, 2018; Schroth, 2019):

AE
 They are frequent users of ICT (e.g., smartphones, social media platforms,
computers) as they grew up with it.
 Gen Z students embrace the values and principles of sustainable development,
human rights, social justice, social responsibility, and volunteerism.
 Being well-informed, they are eager to collaborate with public institutions,
corporations, and NGOs to address various challenges of human society, such as climate
change, social inclusion, health, and AI.
 Gen Z students possess an innovative entrepreneurial spirit and are open to
integrating new ICT into their activities.
 As digital natives and highly Internet-connected, they are rather multitaskers with
limited attention spans.
 When it comes to education, Gen Z students want to ensure that what they are
learning is both applicable and useful in their everyday lives.
With a high level of technology incorporated into their lives and activities, Gen Z students
are pragmatic, career-oriented and skill-focused (Schwieger and Ladwig, 2018).
As natural and long-lasting attributes of a person, abilities represent “the capacity or
power to do something” (Ivanovic and Collin, 2006, p.1) and influence individual
performance. In their turn, skills are individual abilities to perform something based on
education and/or training. They are personal qualities that have three features: productive,
expandable, and social (Green, 2011). Both abilities and skills can be improved over time.
In order to meet the needs of the age of AI systems, a plethora of taxonomies has emerged
in the scientific literature, such as the Global Skills Taxonomy (WEF, 2021) and the
O*NET Content Model (O*NET, 2023). Considering that these taxonomies are close one
to another, the authors referred to the latter because it is more actual and comprehensive
(Table no. 1). In its turn, competence constitutes “the ability to do the tasks required in a
job” (Collin, 2004, p.75).
Today’s technological disruptions (Dinu, 2022), specific to the age of AI systems, are
abruptly modifying the demand for skills and competencies. An AI system is seen as “a
machine-based system that is capable of influencing the environment by producing an
output (predictions, recommendations, or decisions) for a given set of objectives”
(OECD, 2023, p.1) and the age of AI is defined as the one of “the emergence of a
different kind of organisation, structured for a business environment shaped by digital
networks, analytics, and AI” (Iansiti and Lakhani, 2020, p.ix). Arising from the mixture
of big data, cloud computing, Internet of Things, and AI (Siebel, 2019), this continuous
digital transformation is more demanding than ever in relation to the qualities an
individual should possess to be successful in the labour market. As an information society
aiming to provide a sustainable, smart and prosperous human-centered environment for
its citizens, today’s Society 5.0 accomplishes the fusion between physical and cyber space
to address its multiple needs (Fukuyama, 2018). The wide use of digital technologies and
the accelerated implementation of AI systems in the Fourth Industrial Revolution impose
Gen Z students to possess various abilities and acquire specific skills and competencies,
both hard (know how to do) and soft (know how to be), such as programming techniques
and personal flexibility (Agolla, 2018). Recent studies show that some authors highlight
Vol. 26 • No. 65 • February 2024 165

the fact that Gen Z students perceive the age of AI systems as an era of profound
transformations in which success in the labour market is achieved by accumulating
knowledge in higher education institutions that will develop their skills and competencies
(Toma, 2011), while others emphasise how academic programs respond to the challenges
of the digital society (Schwieger and Ladwig, 2018).
Table no. 1. Abilities and skills taxonomy
Level 1 Level 2 Levels 3+4
Abilities Cognitive Verbal abilities: Oral Comprehension, etc.
Abilities Idea Generation and Reasoning Abilities: Fluency of Ideas, etc.
Quantitative Abilities: Mathematical Reasoning, etc.
Memory: Memorisation, etc.
Perceptual Abilities: Perceptual Speed, etc.
Spatial Abilities: Spatial Orientation, etc.
Attentiveness: Selective Attention, etc.
Psychomotor Fine Manipulative Abilities: Finger Dexterity, etc.
Abilities Control Movement Abilities: Control Precision, etc.
Reaction Time and Speed Abilities: Reaction Time, etc.
Physical Physical Strength Abilities: Static Strength, etc.
Abilities Endurance: Stamina, etc.
Flexibility, Balance, and Coordination: Dynamic Flexibility, etc.
Sensory Visual Abilities: Near Vision, etc.
Abilities Auditory and Speech Abilities: Hearing Sensitivity, etc.
Skills Basic skills Content: Reading Comprehension, etc.
Process: Critical Thinking, etc.
Cross- Social: Social Perceptiveness, etc.
Functional Complex Problem Solving: Problem Identification, etc.
skills Technical: Operations Analysis, etc.
Systems: Systems Analysis, etc.
Resource Management: Management of Financial Resources, etc.
Source: O*NET, 2023
As in most cases Gen Z students have entered or are just entering the labour market, they
are highly interested in the abilities, skills, and competencies that can be acquired during
their education. This is why numerous higher education institutions worldwide have
understood the need to reinvent themselves to face the continuous changes provoked by
ICT in the age of AI systems and to adapt to the demand for skilled people in the labour
market (Hernandez-de-Menendez et al., 2020).
This paper investigates Gen Z students' perceptions of the skills, abilities, and
competencies required in the age of AI systems. Starting from the results of the
specialised literature review and considering the contributions that the authors wish to
make to its development, they formulated the following four research hypotheses:
 H1: Gen Z students consider sensorial, cognitive, and psychomotor abilities to be
the most important in the age of AI systems.

AE
 H2: Gen Z students consider that general, Business and Management, and technical
skills and competencies are important in the age of AI systems.
 H3: Gen Z students consider that academic programs provide the general skills and
competencies, those in the fields of Business and Management, as well as the technical
ones needed in the age of AI systems.
 H4: Gen Z students consider that the study program followed exerts certain
influences on the consolidation of abilities, skills, and competencies needed in the context
of the age of AI systems.
To achieve the purpose of the research, the authors used a quantitative research method
based on an online survey because it gave them the opportunity to investigate a large
sample of Gen Z students to obtain the information needed for the statistical analysis. The
survey was conducted online because it does not involve considerable human resources,
its costs are low, and its duration is short (Hague, 2022).
The authors developed a structured questionnaire, starting from the specialised literature
review and based on the O*NET Content Model (O*NET, 2023). It comprised closed
questions with single and multi-response questions. The attitudinal questions sought to
understand what actual Gen Z graduates from full-time undergraduate programs think in
relation to the needed abilities, skills, and competencies in the age of AI systems. The
structure of the questionnaire contained 19 questions, grouped as follows:
 7 questions regarding Gen Z current graduates’ perceptions toward the abilities,
skills, and competencies required in the age of AI systems.
 12 issues related to socio-demographic data (e.g., gender, specialisation).
The measurements were made through the use of a five-point Likert scale (where 1 =
strongly disagree and 5 = strongly agree).
The target population (Table no. 2) and the sample (Table no. 3) were selected from the
graduates of the five undergraduate programs of the Faculty of Administration and
Business (FAB), University of Bucharest (UB). The following main reasons were the
basis of the authors' choice:
a) The field of higher education was chosen as most of its graduates belong to Gen Z. The
graduates represent the main stakeholders of the age of AI as they will be tomorrow’s
employees, entrepreneurs, and businessmen. Moreover, some of them already
accumulated work experience during their study years.
b) The target population comprised the graduates from the FAB, UB, one of the leading
Romanian higher education institutions. Only two public higher education institutions
located in Bucharest provide both Business and Administration specialisations.
c) As graduates, the respondents were highly familiar with the key concepts of the
research (e.g., skills, AI). These concepts were subjects of various disciplines (e.g.,
Human Resources Management, Informatics).
Sampling was done starting from Gen Z students who graduated from bachelor's
programs, full-time education, the sample being composed only of graduates aged
between 21-23 years. Due to the fact that its size was quite small, the sample included the
Vol. 26 • No. 65 • February 2024 167

entire graduate population. The fieldwork research was deployed in the period 6-27 June
2023. In this sense, an online questionnaire was applied to the whole population of
graduates within the day-course undergraduate programs. Initially, the questionnaire was
tested on a small sample of 28 respondents, consisting of employees working in human
resources departments, entrepreneurs, and businessmen, mostly graduates of the FAB,
UB.
A total number of 352 questionnaires were validated from graduates (80 out of 432
graduates were not Gen Z members, sent incomplete responses or did not answer), with a
response rate of 81.48%, confirming the quality of the survey (Hox and De Leeuw, 1994).
The rather small size of the population allowed the authors to obtain a higher response
rate and greater reliability of the data (Wu, Zhao and Fils-Aime, 2022). The gender
structure of the respondents (61.93% female and 38.07% male) was very close to the
gender structure of the total population (62.73% female and 37.27% male), ensuring the
representativeness of the sample. In the end, the authors centralised and systematised all
the information in a database. Then, they were processed through the SPSS 17.0 statistical
software.
Table no. 2. Total number of graduates and their gender within the full-time
undergraduate programs, at the end of the academic year 2022-2023
Number Gender
Specialisation
of Graduates Male Female
Business Administration 125 54 71
(in Romanian)
Business Administration (in English) 40 13 27
Marketing 71 19 52
Economic Cybernetics 62 37 25
Public Administration 134 38 96
Total 432 161 271
Table no. 3. Total number of respondents and their gender within the full-time
undergraduate programs
Number of Gender
Specialisation
Respondents Male Female
Business Administration (in Romanian) 100 43 57
Business Administration (in English) 32 10 22
Marketing 61 19 42
Economic Cybernetics 50 30 20
Public Administration 109 32 77
Total 352 134 218
In order to understand the grounds of the present analysis, it is necessary to provide some
pieces of information relating to the data collection and processing, as well as to the
econometric methods and techniques used so as to test the hypotheses formulated, for
reaching the pre-established goal of the research.
Considering that this study deals with three very large datasets grouped around three
ordinal variables, based on common characteristics, namely: (1) abilities potentially
necessary in the age of AI systems, relating to cognitive, psychomotor, physical issues

AE
etc. (Q1.2) (41 items); (2) skills and competencies potentially necessary in the age of AI
systems, relating to contents, processes, systems, complex problem solving, resource
management, knowledge, technical aspects, etc. (Q2.2) (76 items); (3) level of
contribution of the study programme undergone to the strengthening of the skills and
competences potentially necessary in the age of AI systems, relating to contents,
processes, systems, complex problem solving, resource management, knowledge,
technical aspects etc. (Q3.2) (76 items), the authors decided to check for the reliability
and internal consistency of each series of items on the associated scale, by resorting to
Cronbach's Alpha coefficient, so as to open the road, in case of validation of the same, for
the Principal Component Analysis (PCA), meant for reducing the number of such items,
while preserving reasonably high the information level, therefore facilitating any further
analysis. Provided that Cronbach's Alpha is validated, PCA, a statistical process used for
converting some correlated items, by virtue of an orthogonal transformation, into a
reduced series of linearly uncorrelated ones, becomes applicable, being useful for
outlining different constructs that might be generated based on the series of interest.
Several steps, consisting of the standardisation of the related items, the computation of the
arising covariance matrix or the determination of the eigenvalues and of the explained
variance, precede the identification of the pertinent principal components. The
communalities, also called herein, reflect how much information is extracted from each
initial item. Specifically, the initial communalities indicate the total initial variance of
each variable used in PCA (having a unitary value, unlike the Exploratory Factor
Analysis case), while the post-extraction communalities reveal the share of the total initial
variance of each variable explained by the deriving components. Having in mind that
variance means information, higher the values provided, more consistent the information
package generated as result of the analysis.
Last but not least, the rotated component matrix, based on the rotation of the factorial
axes, helps in easily construing the retained variables, namely, the principal components
or, otherwise said, the synthetic or condensed variables. In this specific case, we resorted
to Varimax, a rotation method that preserves the factors fully uncorrelated, this allowing
for using such arising items (factors) as explanatory variables in regressions, without
fearing that they might be characterised by multi-collinearity.
Once PCA finalised and any further process facilitated by virtue of the decrease of the
consistent number of variables, the correlational analysis is taken into consideration. As
the collected data are ordinal and given the specificity of the same, the recommended
correlational techniques are the rank-based Spearman coefficient (ρ), respectively Kendall
coefficient (τ), both revealing the bivariate correlation between variables. Such an
approach helps the authors to identify the relationships existing among the principal
components of the related items and to construct the said aspects from a theoretical
perspective.

The arising output, involving the testing of the internal consistency of data, based on
Cronbach's Alpha coefficient, the identification of the principal components for the large
series of items, grouped around the above-mentioned three variables considered, and the
Vol. 26 • No. 65 • February 2024 169

analysis of the possible correlations between the same, provided us with a series of
information of interest for our study.
In order to simplify the analysis output rendering and interpretation process, we resorted
to their progressive organisation, by stages, at the level of the three groups of items. After
a brief description of the case processing summary, with 352 observations, all of them
validated, the reliability statistics is considered, indicating a coefficient of 0.952 for the
first group of items, 0.979 for the second one, respectively 0,980 for the third one (> 0.7)
for Cronbach's Alpha, confirmed by its values computed when based on standardised
items, and, therefore, standing for a significant internal consistency of the items in the
related scale. It is to be mentioned that the values of the scale mean / variance /
Cronbach's Alpha if the related item is deleted reveal the high contribution of each
element in the list to the above-mentioned overall consistency, none of them having to be
removed from the group. The corrected item total correlation (> 0.4), also provided by
item total statistics, supports the significant relationship between the items of interest, at
the level of all analysed groups.
Once the reliability test has been validated, the authors proceed to the Principal
Component Analysis, identifying the specific constructs to be considered in this research.
The arising correlation matrix reveals the positive relationships between items, medium to
high in average, statistically significant, with a p-value < 0.01, in most of the cases. The
value of the correlation matrix determinant, amounting to 1.34E-012, 3.81E-032,
respectively 3.38E-038 (> 0), shows the absence of multi-collinearity for the three groups
of items considered.
The KMO and Barlett's test reveals, by virtue of the former, with a value of 0.928 for the
first group of items, respectively, 0.955 for the second and third group of items (> 0.6-
0.7), the adequacy of the database, the samples being fit for analysis in terms of PCA,
while, by virtue of the latter, testing the sphericity, with a statistical p-value < 0.01, the
inequality between the correlation matrix and the unitary matrix, namely, the existence of
a significant correlation among at least two initial items. In terms of communalities, as the
arising extracted variance is large enough (> 0.3-0.4), the authors assume that all items
are adequate and should be preserved for finalising PCA.
Knowing that the eigenvalues, representing the variances of the factors, should exceed 1
for the principal components extracted from the series of items to be considered,
according to the Kaiser Rule (Kaiser, 1960), the authors conclude that they are going to
select, out of the series of 41 items relating to the first group, 7 newly generated elements,
standing for a cumulative variance of 63.82%, out of the 76 items specific to the second
group, 11 components synthesising the information comprised by the initial variables,
cumulating a significance variance of 69.42%, respectively 10 components rendering the
information incorporated into the 76 initial items, for the third group of items analysed,
cumulating a very high variance, amounting to 73.15%. The high number of arising
principal components is due to the exhaustive initial approach, with an increased level of
originality, not encountered by authors at the level of other studies treating similar topics,
resulting in the cumulation of an impressive volume of variables (according to the
previously specified issues).
An interesting aspect is that the first three components contain 50.37% out of the
information incorporated in the initial set of 41 items for the first group considered,

AE
52.34% for the second one, out of the 76 initial items, respectively 58.1% for the third
group, with eigenvalues clearly detached in relation to the ones of the remainder of
components, for each and every separate case (> 2.4 in relation to 1.6, for the fourth
component, > 4.3 in relation to 2.8 for the next component, respectively > 4.1 relative to
2.7, specific to component no. 4), suggesting their significant consistency in revealing the
essence of the preliminary sets of variables. The number of such components might be
also detected by visualising the scree plot, relevant in this regard (Cattell, 1966),
generated by SPSS, strengthening the previously mentioned aspects, namely the
dimensionality of the significant factors for the current analysis, the first three of them
outstanding before the remainder of the graphically rendered elements. In this context,
given the consistent informational concentration at the level of the first three components,
aspect identified also from a graphical perspective, considering the bi-dimensional or tri-
dimensional representation and, especially, the fact that an increased number of factors
might excessively complicate the related structure, generating the so-called ”bloated
specifics”, representing factors generated as result of the artificial superposition of
variables, due to the phrasing of similar items (Cattell, 1961), the authors decided to
retrain the newly detected items (Braeken and Van Assen, 2017) exclusively to the first
three factors for each group of analysed factors.
Given the association of the related components with the initial items, deriving based on
the rotated component matrix output (table no. 4), the authors might name the three arisen
factors in the most convenient and suggestive way, such as: PC_1.1 – Sensorial abilities,
PC_1.2 – Cognitive abilities, and PC_1.3 – Psychomotor abilities, this reflecting the
abilities deemed necessary in the age of AI systems. The results of this research show that
Gen Z students perceive sensorial, cognitive and psychomotor abilities as the most
important abilities in the age of AI systems, thus validating hypothesis H1. In this sense,
they consider qualities, such as peripheral vision, depth perception, written comprehension,
fluency of ideas, finger dexterity, manual dexterity etc. as essential in the current digital
society. In the scientific literature, there are other studies that confirm this statement
(Iorgulescu, 2016; McKinsey, 2022; Mahajan, Gupta and Misra, 2022; Deloitte, 2023).
Table no. 4. Rotated Component Matrixa for the group of items (1)
Items PC_1.1 Items PC_1.2 Items PC_1.3
Peripheral 0.791 Written 0.715 Finger Dexterity 0.832
Vision Comprehension
Sound 0.784 Fluency of Ideas 0.715 Manual 0.807
Localisation Dexterity
Depth 0.784 Written Expression 0.669 Arm-Hand 0.797
Perception Steadiness
Hearing 0.763 Selective Attention 0.645 Multilimb 0.657
Sensitivity Coordination
Auditory 0.760 Originality 0.608 Control 0.544
Attention Precision
Night Vision 0.745 Problem Sensitivity 0.599
Visual Color 0.719 Mathematical 0.583
Discrimination Reasoning
Information Ordering 0.573
Deductive Reasoning 0.521
Note: a. Rotation converged in 7 iterations.
Source: Authors' computation in SPSS
Vol. 26 • No. 65 • February 2024 171

The rotated component matrix (Table no. 5) facilitated the clear delimitation and interpretation
of the arisen factors. As mentioned above, out of the number of components proposed by the
SPSS output (11), only the first three were considered, specifically: PC_2.1 – General
knowledge skills and competences (Engineering and Technology, Mathematics and Science,
Arts and Humanities), PC_2.2 – Business and Management skills and competences,
respectively PC_2.3 – Technical skills and competences, these ones consistently covering the
information contained in the initial set of 76 items. The related components stand for the skills
and competencies deemed necessary in the age of AI systems.
These findings reveal that a plethora of general knowledge, business and management,
and technical skills and competencies, such as AI, Big Data, Programming Languages,
Computer Knowledge, Machine Learning, Data Analysis, human resources management,
business administration, and equipment/tools installation, maintenance, use, and control
etc. are considered as valuable in the age of AI systems by the respondents. These
outcomes confirm the validity of hypothesis H2 and are in line with previous research that
highlight their importance, especially digital skills and competencies (Deloitte, 2019;
Meister, 2020; Mărginean, 2021; Benítez-Márquez et al., 2022).
Artificial 0.862 Human 0.701 Equipment/Tools 0.785
Intelligence Resources Maintenance
Management
Big Data 0.851 Customer 0.686 Equipment/Tools Use 0.779
Relationship and Control
Management
Programming 0.832 Sales 0.669 Equipment/Tools 0.775
Languages Installation
Geography 0.814 Leadership 0.659 Equipment/Tools 0.760
Operation Monitoring
History 0.813 Economics 0.652 Equipment/Tools Repair 0.752
Physics 0.797 Accounting 0.651 Equipment/Tools Failure 0.723
Cause Identification
Machine 0.788 Financial 0.629 Equipment/Tools 0.693
Learning Resources Selection
Management
Computer 0.797 Management of 0.619 Technology Design 0.572
Knowledge Products and/or
Services
Philosophy 0.730 Material 0.616 Quality Control Analysis 0.551
Resources
Management
Theology 0.726 Management 0.599
Programming 0.660 Marketing 0.584
Competencies
Law 0.625 Business 0.568
Administration
Psychology 0.611
Sociology 0.576
Data Analysis 0.549

AE
The rotated component matrix (Table no. 6) outlines the arising factors, helping us in
identifying their overall meaning. Thus, the number of components, indicated based on
the eigenvalues by SPSS (10), is finally reduced to three of them namely: PC_3.1 –
General knowledge skills and competences (Engineering and Technology, Mathematics
and Science, Arts and Humanities) acquired during higher education studies, PC_3.2 –
Technical skills and competences acquired during higher education studies and PC_3.3–
Business and Management skills and competences acquired during higher education
studies, such items covering the information contained in the initial set of 76 items, at a
minimum individual level exceeding 5%. These components reveal the level of
contribution of the study programme undergone to the strengthening of the skills and
competences deemed necessary in the age of AI systems, validating the hypothesis H3.
Artificial 0.901 Equipment/Tools Use 0.832 Financial 0.751
Intelligence and Control Resources
Management
Geography 0.893 Equipment/Tools 0.830 Customer 0.720
Operation Monitoring Relationship
Management
Big Data 0.886 Equipment/Tools 0.821 Marketing 0.682
Failure Cause
Identification
Programming 0.885 Equipment/Tools 0.811 Management 0.673
Languages Maintenance
Public Policies 0.875 Equipment/Tools 0.806 Material 0.670
Installation Resources
Management
Physics 0.866 Equipment/Tools 0.777 Management of 0.668
Selection Products and/or
Services
Philosophy 0.845 Equipment/Tools 0.768 Economics 0.659
Repair
Computer 0.840 Operations Analysis 0.731 Human Resources 0.658
Knowledge Management
Machine 0.819 Equipment/Tools 0.580 Business 0.652
Learning Testing Administration
Programming 0.801 Technology Design 0.565 Leadership 0.586

Competencies
History 0.762 Sales 0.585
Sociology 0.712 Time 0.562
Management
Psychology 0.668 Accounting 0.522
Law 0.615
Mathematics 0.596
Data Analysis 0.595
Vol. 26 • No. 65 • February 2024 173

Interesting enough is the fact that, when it comes to the academic curricula, Gen Z
students consider that they provide in a high proportion the skills and competencies
required in the age of AI systems in accordance with their opinions already mentioned.
These results are consistent with those of other studies that emphasise the need for
permanent skilling and reskilling (OECD, 2019; BCG, 2021; Tamayo et al., 2023).
The detection of the principal components helped us in easily testing the possible
correlation between the variables of interest for this study. Therefore, the authors selected
the first three components for each such variable, specifically: PC_1.1, PC_1.2 and
PC_1.3 (abilities deemed necessary in the age of AI systems), PC_2.1, PC_2.2 and
PC_2.3 (skills and competences deemed necessary in the age of AI systems),
respectively, PC_3.1, PC_3.2 and PC_3.3 (level of contribution of the study programme
undergone to the strengthening of the skills and competences deemed necessary in the age
of AI systems).
Hereinafter, are rendered the correlational analysis output, got based on Spearman and
Kendall coefficients, for which statistically significant values were obtained. This
selective aspect is due to the irrelevance of construing results for which the statistical
validation is null. As the declared intention was to determine the effective impact of the
study programme undergone to the strengthening of the abilities, skills and competences
deemed necessary in the age of AI systems, the authors resorted to the study of the
behaviour of PC_1.1, PC_1.2 and PC_1.3 in relation to PC_3.1, PC_3.2 and PC_3.3
(Table no. 7), respectively of PC_2.1, PC_2.2 and PC_2.3 in relation to PC_3.1, PC_3.2
and PC_3.3 (Table no. 8), elements exhaustively rendered during the previous stages of
the current research.
Thus, there is, as expected, a quite low correlation between PC_1.1 and PC_3.1,
respectively, PC_3.3, between PC_1.2 and PC_3.1, respectively, PC_3.3 and between
PC_1.3 and PC_3.2, such values being statistically significant, at a significance level of
1%, respectively, 5%, the results being similar for both coefficients, with slightly higher
values for Spearman's rho. This means that there is a correlation, although not strong or
even medium, between specific sensorial abilities and general knowledge skills and
competences (Engineering and Technology, Mathematics and Science, Arts and
Humanities) acquired during higher education studies, respectively Business and
Management skills and competences acquired during higher education studies, on one
hand, and between psychomotor abilities and general knowledge skills and competences
(Engineering and Technology, Mathematics and Science, Arts and Humanities) acquired
during higher education studies, respectively Business and Management skills and
competences acquired during higher education studies, on the other hand, the same kind
of relationship being encountered when coming about cognitive abilities in relation to
technical skills and competences acquired during higher education studies.
Table no. 7. Correlation between
PC_1.1, PC_1.2, PC_1.3 and PC_3.1, PC_3.2, PC_3.3
PC_3.1 PC_3.2 PC_3.3
Kendall's tau_b PC_1.1 Correlation Coefficient .122** - .113*
Sig. .010 .016
PC_1.2 Correlation Coefficient -.143** - .144**
Sig. .003 .002

AE
PC_3.1 PC_3.2 PC_3.3

PC_1.3 Correlation Coefficient - .107* -
Sig. .021
Spearman's rho PC_1.1 Correlation Coefficient .137** - .132*
Sig. .010 .013
PC_1.2 Correlation Coefficient .162** - .164**
Sig. .002 .002
PC_1.3 Correlation Coefficient = .124* -
Sig .020
Notes: * Correlation is significant at the 0.05 level (2-tailed); ** Correlation is significant at the
0.01 level (2-tailed).
Table no. 8. Correlation between

PC_2.1, PC_2.2, PC_2.3 and PC_3.1, PC_3.2, PC_3.3
PC_3.1 PC_3.2 PC_3.3
Kendall's tau_b PC_2.1 Correlation Coefficient ,310** - -
Sig. (2-tailed) ,000
PC_2.2 Correlation Coefficient - - ,221**
PC_2.3 Correlation Coefficient - - -
Sig. (2-tailed)
Spearman's rho PC_2.1 Correlation Coefficient ,350** - -
PC_2.2 Correlation Coefficient - - ,249**
PC_2.3 Correlation Coefficient - - -
Sig. (2-tailed)
Note: ** Correlation is significant at the 0. 01 level (2-tailed).
In this case, the authors identify a medium, statistically significant correlation, at a
significance level of 1%, between PC_2.1 and PC_3.1 and between PC_2.2 and PC_3.3,
while, surprisingly, no statistically significant correlation is encountered between PC_2.3
and PC_3.1, PC_3.2, respectively, PC_3.3, situation confirmed by both coefficients used
in such respect. This reveals that there is a reasonable correlation between required
general knowledge skills and competences (Engineering and Technology, Mathematics
and Science, Arts and Humanities) and general knowledge skills and competences
(Engineering and Technology, Mathematics and Science, Arts and Humanities) acquired
during higher education studies, respectively, between required Business and
Management skills and competences and Business and Management skills and
competences acquired during higher education studies. However, this type of connection
is not statistically validated for the assumed relationship between the required technical
skills and the competences and technical skills and competences acquired during higher
education studies.
Vol. 26 • No. 65 • February 2024 175

Therefore, overall, the authors state that there is some statistically confirmed connection,
ranging from low to medium, between the perceived level of contribution of the study
programme undergone to the strengthening of the abilities, skills and competences
deemed necessary in the age of AI systems and the effective abilities, skills and
competences deemed necessary in the age of AI systems. Otherwise said, even if the
correlational analysis is not a cause-effect type of test, we understand that the idea behind
this result stands for the fact that the study programme undergone exerts some influences
on the strengthening of the abilities, skills and competencies deemed necessary in the age
of AI systems, thus confirming the validity of hypothesis H4.
Conclusions
The appearance and expansion of AI have profoundly changed the way human society, in
general, and the economy, in particular, have functioned in the last decades. This has led
to significant changes in the labour market that have imposed education to adapt in order
to provide pupils and students with the required abilities, skills and competencies in the
age of AI systems. From a theoretical point of view, this paper contributes to the
development of the scientific literature related to the new relationships that emerged
between human capital and the needs of a digital society, increasingly based on AI. It
provides fresh insights into the current concepts of abilities, skills and competencies by
considering updated taxonomies. Moreover, this study identifies and analyses the
perceptions of Gen Z students on these concepts in the age of AI systems and on the way
academic curricula cope with these new and challenging requirements. Also, it
demonstrates that, on the one hand, sensorial, cognitive, and psychomotor abilities, and,
on the other hand, general knowledge, business and management, and technical skills and
competencies, are playing a key role in today’s digital society.
From a practical point of view, students’ abilities, skills and competencies are
interconnected and their development should constitute a top priority for any higher
education institution. This is why academic curricula must be designed and implemented
in full accordance with the needs of the labour market. In this respect, other stakeholders,
such as society, governmental authorities, companies, public institutions, and NGOs,
should be involved in the educational process in order to better face the challenges of the
age of AI systems. Through higher investments in ITC and its deeper integration into the
activities of higher education institutions, students may become better prepared and raise
their employability in an increasingly competitive and turbulent labour market. Therefore,
policymakers, together with decision-makers from higher education institutions should
collaborate and analyse the results of this study that may provide solutions aimed at
improving the educational process.
The research results can be extended to other specialisations because, on the one hand, AI
has practically penetrated all fields of activity, and on the other hand, the questionnaire
can be successfully applied not only to the already mentioned specialisations (Business
Administration, Marketing, Economic Cybernetics, Public Administration), but also to
other specialisations because it also contains items specific to them (for example,
Informatics, Mathematics, Physics, Engineering, Sociology, Psychology, Finance,
Accounting, Management, etc.). However, the outcomes of this paper should be
interpreted by considering some of its limitations. First, in spite of the great number of

AE
items considered, the research does not analyse all of them as presented in various
taxonomies. Thus, future research could integrate other new and relevant items that
influence the employability of graduates. Second, another limitation of this study is given
by the size and structure of the sample. It is highly representative at the faculty level, but
it is rather difficult to assert that its results may be generalised in other Romanian or
foreign higher education institutions. Future research may use larger and, therefore, more
representative samples. Therefore, these results may constitute starting points for other
research on Gen Z students from other specialisations and/or other higher education
institutions located in Romania or abroad. Third, another limitation is the lack of a
qualitative research involving specialists and experts from higher education institutions,
human resource management departments, businessmen, and entrepreneurs. This type of
research may be beneficial in order to better understand the challenges of the age of AI
systems in future studies. In this regard, the inherent risks associated with the expansion
of AI systems in all fields of activity for Gen Z students can also be considered.
The originality of this research is twofold. Firstly, it provides a new perspective on the
structure of abilities, skills and competencies in the digital society. Secondly, it analyses
Gen Z students’ perceptions towards these concepts and the way education addresses
them in the age of AI systems.
References
Agolla, J.E., 2018. Human Capital in the Smart Manufacturing and Industry 4.0.
Revolution. In: A. Petrillo, R. Cioffi and F. de Felice, eds., 2018. Digital
Transformation in Smart Manufacturing. Rijeka: IntechOpen. Ch.3.
https://doi:10.5772/intechopen.73575.
Aparicio, G., Iturralde, T. and Rodriguez, A.V., 2023. Developments in the Knowledge-
Based Economy Research Field: A Bibliometric Literature Review. Management
Review Quarterly, 73, pp. 317-352. https://doi.org/10.1007/s11301-021-00241-w.
Benítez-Márquez, M.D., Sánchez-Teba, E.M., Bermúdez-González, G. and Núñez-
Rydman, E.S., 2022. Generation Z within the workforce and in the workplace: A
bibliometric analysis. Frontiers in Psychology, 12, article no. 736820. https://doi.org/
10.3389/fpsyg.2021.736820.
BCG, 2021. Decoding Global Reskilling and Career Paths. [online] Available at:
<https://www.bcg.com/publications/2021/decoding-global-trends-reskilling-career-
paths> [Accessed 28 August 2023].
Braeken, J. and Van Assen, M.A.L.M., 2017. An Empirical Kaiser Criterion.
Psychological Methods, 22(3), pp. 450-466.
Cattell, R.B., 1961. Theory of Situational, Instrument, Second Order, and Refraction
Factors in Personality Structure Research. Psychological Bulletin, 58, pp. 160-174.
http://dx.doi.org/10.1037/h0045221.
Cattell, R.B., 1966. The Scree Test for the Number of Factors. Multivariate Behavioral
Research, 1, pp. 245-276. http://dx.doi.org/10.1207/s15327906mbr0102_10.
Cho, M., Bonn, M.A. and Han, S.J., 2018. Generation Z’s sustainable volunteering:
Motivations, attitudes and job performance. Sustainability, 10(5), article no. 1400.
https://doi.org/10.3390/su10051400.
Vol. 26 • No. 65 • February 2024 177

Collin, P.H., 2004. Dictionary of Business. 4th edition. London: A&C Black.
Davidson, C., Heyman, F., Matusz, F., Sjoholm, F. and Zhu, S., 2010. Globalization and
Imperfect Labor Market Sorting. [online] Geneva: International Labour Office.
Available at: <https://www.ilo.org/wcmsp5/groups/public/@ed_emp/documents/
genericdocument/wcms_158383.pdf> [Accessed 25 August 2023].
Deloitte, 2019. Understanding Generation Z in the workplace. [online] Deloitte.
Available at: <https://www2.deloitte.com/us/en/pages/consumer-business/articles/
understanding-generation-z-in-the-workplace.html> [Accessed 30 August 2023].
Deloitte, 2022. What's driving the shift from job to skills? [online] Deloitte. Available at:
<https://action.deloitte.com/insight/2825/whats-driving-the-shift-from-job-to-skills>
Deloitte, 2023. 2023 Gen Z and Millennial Survey. [online] Deloitte. Available at:
<https://www.deloitte.com/global/en/issues/work/content/genzmillennialsurvey.html?i
cid=gx:2or:3dp:4genzandmillennialsurvey2023:5GC1000102:6wok:20230517::deloitt
einsights%3Ficid%3Dlearn_more_content_click> [Accessed 30 August 2023].
Dimock, M., 2019. Defining generations: Where Millennials end and Generation Z
begins. [online] Pew Research Center. Available at: <https://www.pewresearch.org/
short-reads/2019/01/17/where-millennials-end-and-generation-z-begins/> [Accessed
30 August 2023].
Dinu, V., 2021. Artificial Intelligence in Wholesale and Retail. Amfiteatru Economic, 23
(56), pp. 5-7. https://doi.org/10.24818/EA/2021/56/5.
Drucker, P.F., 1969. The Age of Discontinuity: Guidelines to Our Changing Society.
London: Heinemann.
Fukuyama, M., 2018. Society 5.0: Aiming for a New Human-Centered Society. [online]
Hitachi Review. Available at: <https://www.hitachi.com/rev/archive/2017/r2017_06/
trends/index.html> [Accessed 29 August 2023].
Green, F., 2011. What is Skill? An Inter-Disciplinary Synthesis. LLAKES Research Paper
20, pp. 1-25. [online] Available at: <https://www.llakes.ac.uk/wp-
content/uploads/2021/01/Green-What-is-Skill-final.pdf> [Accessed 12 August 2023].
Hague, P., 2022. Market Research in Practice. 4th edition. London: Kogan Page.
Hernandez-de-Menendez, M., Escobar Diaz, C.A. and Morales-Menendez, R., 2020.
Educational experiences with Generation Z. International Journal on Interactive
Design and Manufacturing, 14, pp. 847-859. https://doi.org/10.1007/s12008-020-
00674-9.
Hope, J. and Hope, T., 1997. Competing in the Third Wave: The Ten Key Management
Issues of the Information Age. Boston: Harvard Business School Press.
Hox, J.J. and De Leeuw, E.D., 1994. A comparison of nonresponse in mail, telephone,
and face-to-face surveys. Quality and Quantity, 28, pp. 329-344. https://doi.org/
10.1007/BF01097014.
Iansiti, M. and Lakhani, K.R., 2020. Competing in the Age of AI: Strategy and Leadership
when Algorithms and Networks Run the World. Boston: Harvard Business Review
Press.
Iorgulescu, M.-C., 2016. Generation Z and its Perception of Work. Cross Cultural
Management, 18(1), pp. 48-54.

AE
Ivanovic, A. and Collin, P.H., 2006. Dictionary of Human Resources and Personnel
Management. 3rd edition. London: A&C Black.
Jaumotte, F. and Tytell, I., 2007. How has the globalization of labor affected the labor
income share in advanced countries?. [online] IMF Working Paper 07/298. Available
at: <https://www.imf.org/external/pubs/ft/wp/2007/wp07298.pdf> [Accessed 21
August 2023].
Kaiser, H.F., 1960. The application of electronic computers to factor analysis.
Educational and Psychological Measurement, 20, pp. 141-151. http://dx.doi.org/
10.1177/001316446002 000 116.
Mahajan, R., Gupta, P. and Misra, R., 2022. Employability skills framework: a tripartite
approach. Education+Training, 64(3), pp. 360-379. https://doi.org/10.1108/ET-12-
2020-0367.
Mărginean, A.E., 2021. Gen Z perceptions and expectations upon entering the workforce.
European Review of Applied Sociology, 14(22), pp. 20-30. https://doi.org/10.1515/
eras-2021-0003.
McKinsey, 2022. How Does Gen Z See Its Place in the Working World? With
Trepidation. 19.10.2022. [online] Available at: <https://www.mckinsey.com/featured-
insights/sustainable-inclusive-growth/future-of-america/how-does-gen-z-see-its-place-
in-the-working-world-with-trepidation#/> [Accessed 30 August 2023].
Meister, J., 2020. Top 10 HR Trends that Matter Most in the 2020 Workplace. [online]
Available at: <https://www.forbes.com/sites/jeannemeister/2020/01/15/top-10-hr-
trends-that-matter-most-in-the-2020-workplace/> [Accessed 13 August 2023].
O*NET, 2023. The O*NET Content Model. [online] Available at:
<https://www.onetcenter.org/content.html> [Accessed 10 May 2023].
Organisation for Economic Cooperation and Development (OECD), 2019. OECD
Employment Outlook 2019: The Future of Work. Paris: OECD Publishing.
https://doi.org/10.1787/9ee00155-en.
Organisation for Economic Cooperation and Development (OECD), 2023. AI Terms and
Concepts. [online] OECD – AI Policy Observatory. Available at: <https://oecd.ai/en/
ai-principles> [Accessed 12 May 2023].
Powell, W.W. and Snellman, K., 2004. The Knowledge Economy. Annual Review of
Sociology, 30, pp. 199-220. https://doi.org/10.1146/annurev.soc.29.010202.100037.
Schroth, H., 2019. Are You Ready for Gen Z in the Workplace? California Management
Review, 61(3), pp. 5-18. https://doi.org/10.1177/0008125619841006.
Schwieger, D. and Ladwig, C., 2018. Reaching and Retaining the Next Generation:
Adapting to the Expectations of Gen Z in the Classroom. Information Systems
Education Journal, 16(3), pp. 45-54.
Seemiller, C. and Grace, M., 2016. Generation Z Goes to College. San Francisco: Jossey-
Bass.
Siebel, T.M., 2019. Digital Transformation: Survive and Thrive in an Era of Mass
Extinction. New York: RosettaBooks.
Tamayo, J., Doumi, L., Goel, S., Kovacs-Ondrejkovic, O. and Sadun, R., 2023. Reskilling
in the Age of AI. [online] Harvard Business Review. Available at:
<https://hbr.org/2023/09/reskilling-in-the-age-of-ai/> [Accessed 13 August 2023].
Vol. 26 • No. 65 • February 2024 179

Tapscott, D., 2014. The Digital Economy Anniversary Edition: Rethinking Promise and
Peril in the Age of Networked Intelligence. 2nd ed. New York: McGraw-Hill.
Toffler, A., 1980. The Third Wave. London: Pan Books.
Toma, S.G., 2011. Exploratory Research on Students` Opinions Regarding the Features of
the Educational Process in a Knowledge-Based Economy. Amfiteatru Economic,
13(30), pp. 484-498.
Tuxworth, E., 1990. Competence Based Education and Training: Background and
Origins. In: J. Burke, ed., 1990. Competency Based Education and Training. London:
The Falmer Press. pp. 10-25.
United Nations (UN), 2015. Transforming Our World: the 2030 Agenda for Sustainable
Development. General Assembly, 21.10.2015. [online] Available at: <https://documents-
dds-ny.un.org/doc/UNDOC/GEN/N15/291/89/PDF/N1529189.pdf?OpenElement>
World Economic Forum (WEF), 2021. Building A Common Language for Skills at Work:
A Global Taxonomy. [online] Available at: <https://www3.weforum.org/docs/
WEF_Skills_Taxonomy_2021.pdf> [Accessed 12 May 2023].
Wu, M.-J., Zhao, K. and Fils-Aime, F., 2022. Response Rates of Online Surveys in
Published Research: A Meta-Analysis. Computers in Human Behavior Reports, 7,
article no. 100206. https://doi.org/10.1016/j.chbr.2022.100206.

AE
ARTIFICIAL INTELLIGENCE AND THE MODELLING

OF TEACHERS’ COMPETENCIES
Ramona Simuț1 , Ciprian Simuț2 , Daniel Bădulescu3
and Alina Bădulescu4*
1)2)3)4)
University of Oradea, Oradea, Romania.

Simuț, R., Simuț, C., Bădulescu, D. and Bădulescu A., Received: 29 September 2023
2024. Artificial Intelligence and the Modelling of Revised: 21 November 2023
Teachers’ Competencies. Amfiteatru Economic, Accepted: 22 December 2023
26(65), pp. 181-200.
Abstract
This research explores educators' perceptions of their role in shaping competencies and also
presents the key challenges and measures perceived by educators in the AI system development
landscape. The study employs a partial least square – structural equation modelling (PLS-SEM)
as the primary analytical tool. A set of hypotheses was derived from the literature, formulated and
tested to substantiate the relationship between educators' attitudes and competencies. The findings
highlight a significant positive relationship between teachers' attitudes toward AI and competence
development, with the most pronounced impact on educational management competencies. Thus,
teachers' positive attitudes toward AI statistically significant influence cognitive, fundamental and
educational management competencies, underscoring the critical role of instilling positive attitude
towards AI among educators to foster the development of multifaceted competencies.
Furthermore, the research highlights key challenges to integrating AI into education, including
the imperative for professional development for educators and ensuring equitable access to
educational resources and technology. From a practical point of view, the study advocates for
initiatives aimed at bridging the digital divide and integrating AI education into curricula as
indispensable policy measures. The originality of our research lies in the unitary approach of the
relationship between positive attitudes of teachers toward artificial intelligence and a series of
competencies, such as cognitive, fundamental and educational management. Although the study
provides valuable information, it has certain limitations. The findings are based on educators'
perceptions and attitudes, which may vary among individuals, the study scope is limited to a
specific region, and the generalisability of findings to other settings may require further
investigation. This research contributes to the existing literature by exploring the relationship
between teachers' attitudes toward AI and the development of their skills, with the help of an
exploratory model using structural equations, a model that has the teachers in the foreground.
Keywords: Artificial intelligence (AI), educator, competencies, attitude.
JEL Classification: A21; I21; O33
*
Corresponding author, Alina Badulescu – e-mail: [email protected]
Vol. 26 • No. 65 • February 2024 181

AE Artificial Intelligence and the Modelling of Teachers’ Competencies
Introduction
Artificial intelligence (AI) refers to the ability of a digital machine to perform tasks by
imitating human intelligence, to learn and think (Xia et al., 2022; Chiu et al., 2023), to solve
problems and make decisions (Bellman, 1978), to do something that would normally be
thought to require the intelligence of a human (Mitchell, 2019), or “machines with minds, in
the full and literal sense” (Haugeland, 1989).
AI has experienced an exceptional spread in recent decades, continuously changing the way
people live, interact, learn and work (Chiu et al., 2022; Xia et al., 2022), even though the
understanding of the place and role of AI in economic development is still fragmented (Qin
et al., 2023). AI excels in the execution and improvement of specific tasks, continuously
modifying and improving each economic sector, taking sound decisions (Dong et al., 2020),
streamlining production processes, and strengthening the businesses in which it is present.
However, the use and expansion of AI in the economy requires the development of new skills
and abilities, from technological expertise to social, emotional and creative skills (Felea and
Vasiliu, 2005; Pelau, Dabija and Ene, 2021; Limna et al., 2022; Klimova, Pikhart and Kacetl,
2023) suggesting the vital relationship between AI and education, full of unexplored
development possibilities.
In the existing literature, the predominant focus has been on examining students' views on
the connection between AI and competencies within their scope and less on teachers'
competencies in developing AI. However, this study shifts the focus to educators and their
role in shaping competencies in the context of AI development. Therefore, the aim of this
research is to investigate the perception of teachers regarding the role of educators in
modelling competencies in the context of the development of artificial intelligence systems
by employing an exploratory model, using structural equation modelling. These
competencies, in turn, serve as vital tools for educators in educating students about AI's
evolution.
In this paper, we will refer to AI applications that are freely available on dedicated
educational platforms or are developed by companies focused on education. These
applications can be used by both teachers and students to develop new skills. The paper is
structured as follows: theoretical background regarding the role and expansion of AI in
education, respectively, the role of educators in modelling competencies in AI, methodology,
results and discussion, conclusions and policy implications.
1. Theoretical background
AI in education energises the evolution of teaching and learning practices and the modernisation
of programmes, being at the same time a fruitful topic for the progress of educational research.
All these reasons and promises offered by AI in education determined numerous initiatives and
clarifications at the national decision-maker level, or within international institutions (UNESCO,
2019; OECD, 2021; European Commission, 2022). The increased attention given to this field
has revealed, however, that much of the research has focused more on engineering aspects
(algorithms, machine learning techniques), on educational technologies (gamification, blended
learning) and less on the impact of AI on education. For many researchers, this impact is still
unclear (Miao et al., 2021), and the limited training and experiences of teachers and managers

AE
make it difficult to introduce and integrate AI into the education system (Hussin, 2018; Owoc,
Sawicka and Weichbroth, 2021).
According to Chiu et al. (2023) the potential of AI in education is huge: it can improve
learning activities by making them more personalised and adaptable; supports teaching by
encouraging them to understand students' learning process, providing anywhere and anytime
prompts interrogations and feedback; improves the evaluation processes and administrative
activities at the level of educational institutions (Luckin, 2017; González-Calatayud,
Prendes-Espinosa and Roig-Vila, 2021). AI is one of the key technologies prepared to
revolutionize education (Talan, 2021), to transform the traditionalist and somewhat rigid way
in which education is perceived and delivered, adapting it to the technological advances of
the modern world (Sadiku et al., 2021).
1.1. Concerns about the role and expansion of AI in education
Concerns and concerns about the impact of AI on education systems and society in general
are numerous and it would be superficial to categorise them only as prejudices and
misconceptions (European Commission, 2022). Some of them even seem serious once they
have led many education specialists and experts to ask for a “pause” in the development of
AI (Future of Life Institute, 2023; Malekos, 2023).
The first concern focusses on (1) the difficulty of understanding AI. The lack of knowledge in
computer science, the complicated and sometimes abstract terminology, and the pace of news
overlap with the real difficulties of understanding the inside workings of the AI system. The lack
of technological knowledge on the part of educators (Chiu and Chai, 2020) is often exacerbated
by the technical infrastructure in schools, often insufficient to support the implementation of AI
in educational processes (McCarthy et al., 2016; Ozdemir and Tekin, 2016).
The second concern and misunderstanding is that (2) AI has a limited or no real role in
education. The implementation of AI in education may be affected by the lack of interest of
teachers in AI, they may perceive AI as “uninteresting and unenjoyable for teaching”
(McCarthy et al., 2016; Celik et al., 2022), or that AI algorithms are not sufficiently
predictable and adapted to their practical needs (Schwarz et al., 2018; Qian, Zhao and Cheng,
2020). However, AI is continuously changing the way we work and live, and education
cannot stay away from these changes. It is important that everyone benefits from these
changes and to be part of them, and that AI systems and solutions in education “to be
developed and used in an ethical, trustworthy, fair and inclusive way” (European
Commission, 2022, p. 12).
Another problem is that (3) AI is not inclusive. It can aggravate inequality, discrimination, or
give birth to new forms of inequality and segregation (UNESCO, 2019; European
Commission, 2022). Experts note the possibility that (biased) algorithms in AI applications
used to evaluate student performance could alter the principles of accuracy and objectivity
on the basis of which they were created and promoted, favouring students who attended
private or well-rated schools, while students from under-represented groups may be more
affected (Akgun and Greenhow, 2022). AI-based scoring may sometimes imperfectly assess
students' performance (Lu, 2019) and can undermine teachers' confidence in their accuracy
(Qian, Zhao and Cheng, 2020). Perpetuating and validating these biased constructs can affect
students' (official) performance, self-confidence, and their future careers.
Vol. 26 • No. 65 • February 2024 183

Another objection relates to (4) trust issues in AI systems, that their rapid, universal
development and spread will lead to the replacement of human decisions in more and more
contexts, increasing the risks of a lack of ethical considerations. Akgun and Greenhow (2022)
consider privacy, surveillance, autonomy, bias, and discrimination to be the most important
ethical challenges of AI applications in education. Seen also as a surveillance system, AI is
alleged of generating problems related to personal autonomy, affecting the ability to judge
and act based on one's own interests and values, for both students and teachers (Regan and
Jesse, 2019). Algorithms that predict future actions based on collected information raise
questions about fairness and self-freedom (Akgun and Greenhow, 2022), and increase the
risks of perpetuating existing prejudices, discriminations and maintaining social stratification
(Murphy, 2019). The regulation of AI will therefore require public discussions regarding
ethics, responsibility, and security such as algorithmic bias, data privacy, content ownership,
and transparency (UNESCO, 2019; Malekos, 2023). Researchers support ethical approaches
more sensitive to context and pedagogy (Adams et al., 2023), cooperation and joint research
involving other stakeholders when it comes to the ethical implications of using AI in
education.
Another caution, specific to education, comes from the supposition that (5) AI will undermine
the role of the teacher and make him/her obsolete and useless. Fear is difficult to ignore and
can be real if educators and authorities in the field do not properly assimilate and understand
the use and impact of AI in education. For Dillenbourg (2016) or Hrastinski et al. (2019),
teacherless education is an overly pessimistic and speculative view, and they recommend
insisting on understanding the opportunities that AI offers to teachers and how these
advantages can change the roles of teachers in the classroom (Celik et al., 2022). Individual
efforts and institutional support are needed to understand that AI helps and complements the
efforts of teachers, generates learning experiences, encourages students' creativity and free
thinking and challenges them to solve real-world problems. However, this AI support cannot
be achieved outside of an authentic human framework of the educator, because AI projects
are tools created to support and strengthen the role of the educator rather than replace it
(Moisil et al., 2010; Pinkwart, 2016; Schiff, 2021). In the logical dismantling of this concern,
comes the argument that AI can simplify and ease administrative tasks (quite substantial in
education systems) allowing educators to focus more and more efficiently on the educational
process, to have time to adapt to new transformations, and implicitly “the role of the teacher
is likely to be augmented and evolve with the capabilities that new innovations for AI in
education will bring” (European Commission, 2022, p. 13). According to Holstein et al.
(2019), AI-based machines can support and complement educators to fulfil their unique role
of “orchestrator” in the learning and teaching process (Dillenbourg, 2013), but the
effectiveness of teaching depends on the ability of teachers to accept and assimilate
appropriate pedagogical methods in their training (Tondeur et al., 2020), to learn from errors
and from the difficulties of the beginning (Prieto et al., 2018).
This concern seems to be related to the concern of (6) Lack of Human Interaction, the
isolation and disconnection of students from peers, teachers and social interactions is often
explained by excessive (sometimes destructive) reliance on AI and the artificial substitution
of genuine social contexts. However, with all the informational and educational contribution
of AI, it cannot substitute and replace “human interaction and the social and emotional
learning that comes with it” (Malekos, 2023).

AE
Finally, (7) Content saturation, along with developing lazy minds and the lack of critical
thinking (Malekos, 2023) is another problem raised. The concern is not only about education,
“AI significantly impacts the loss of human decision making and makes humans lazy”
(Ahmad et al., 2023), it diminishes intuitive analysis and creative problem solving,
autonomy, replacing human decision with its choices (Danaher, 2018). Thus, it is argued that
AI does not simplify, but increases the amount of information, often un-systematised and
biased, and this can lead to fatigue and demotivation, decreasing learning performance. This
situation stimulates lazy minds, disinterest and the “complicity” of an uncritical thinking,
lacking the tools and reasoning against misinformation and manipulation. Although there are
assessments that these tendencies are blurred when AI is complementary and used as a
support and assistant to the teaching done by a real educator, they should not be ignored.
1.2. The role of educators in modelling competencies in AI
AI and education are a topic of significant interest for researchers, Mena-Guacas et al. (2023)
found a rise in the literature on the subject of developing competencies through collaboration.
In the context of student-to-student and AI collaboration, Li and Su (2020) argue that physical
meetings among students are of great importance, despite computer-mediated interactions
made human connections more accessible. Furthermore, research on cognitive,
communicative, collaborative, and generic competencies is also of great interest in the field
of AI and education. As Vinichenko, Melnichuk and Karácsony (2020) argue the issue of AI
development that enables student efficiency, focusses on collective and individual
performance analysis, information for decision making, and providing tools for diagnostic
and predictive analysis, especially in a context where students are tempted to use AI to cheat
in their assignments.
Lameras and Arnab (2022) found that there is a need to rationalise and delineate the extent
to which AI can be inserted and used in education. It seems of great importance to consider
the way in which evaluation tools, various instructional strategies, ethical implications and
teacher competencies (digital, cognitive, fundamental, and in educational management) are
designed, developed, and applied in educational settings (Ng et al., 2021). The means through
which AI would operate in these settings is to transform educational, psychological, and
social knowledge into computational language. This would allow AI to interpret data and
translate it in order to support, guide, and enhance student learning and the ability to think in
a more efficient way (Kim, Lee and Cho, 2022; Sobolu, Stanca and Bodog, 2023). The result
would be technological innovation through activity-based, adaptive, and student-oriented
strategies.
AI would also allow for a deeper understanding of the mechanics of education, through an
increasingly precise analysis of student knowledge, assessment, and feedback. AI is primarily
focused on helping students, but the teacher factor must also be taken into account. Vlasova
et al. (2019) show that one of the main issues is that, without proper and consistent training,
teachers might be overwhelmed by the complexity of AI educational tools that they are to
use in teaching. Also, Zhao, Guo and Liu (2021) suggest customised teacher training
programmes aligned with the local context to guide teachers to consciously improve their
professional development, in order to create professional competencies among teachers. The
solution could be a careful re-evaluation and reorganisation of AI education tools that would
be used in a contextualised manner, adapted for each classroom situation (Touretzky et al.,
2019).
Vol. 26 • No. 65 • February 2024 185

Polak, Schiavo and Zancanaro (2022) argue that AI can be used to develop citizenship skills,
in order to help students become responsible and confident, in maneuvering in an ever
increasingly difficult and complex digital world. Alexandre et al. (2021) found that teachers’
perspectives on AI and its use in educational settings were positive, willing and supportive
of using AI, despite having limited AI-related competencies (Chounta et al., 2022). It seems
that digital skills are not considered sufficient by teachers to use in AI-related educational
issues (Ng et al., 2023). However, teachers argued that in order to have an efficient AI
educational structure, there need to be various teacher training programmes that would
provide basic AI related skills and knowledge, relevant AI content to use in school settings,
connected to interactive and collaborative methods of teaching, accessible software and
hardware, and user motivation initiatives (Vlasova et al., 2019).
On the issue of student competencies in AI education, Sanusi et al. (2022b) determined that
cognitive, human tool collaboration, self-learning, skill competence, and ethics were
significantly influencing AI content. Of significant importance, as Sanusi et al. (2022a) was
the discovery that teamwork competence and AI curriculum need to be improved. Kim and
Kwon (2023) incorporate empirical observations originating from educators' direct
participation in the implementation of AI curricula within their class settings. This method
provides a holistic perspective that encompasses both the cognitive and emotional aspects of
their abilities. Huang (2021) showed that to efficiently accommodate AI course content, such
as programming knowledge, image processing knowledge, natural language processing
knowledge, ethics of artificial intelligence and machine learning, students need to develop
several key competencies, such as skill and cultural competencies, teamwork and human-tool
collaboration competencies, cognition and self-learning competencies.
Pedro et al. (2019) tackle the issue of the challenges AI poses to the general field of education,
such as comprehensive public policy on AI, but in the context of sustainable development,
inclusion and equity in AI, teacher training for AI-powered education, but also preparing AI
to understand education, the need for the development of quality and inclusive data systems,
implementation of significant AI research for education, and ethics and transparency in data
collection, but also in the use and dissemination. In such a context, teachers with a positive
perspective about AI in educational environments are more willing to develop the cognitive,
digital, fundamental, and educational management competencies.
Based on our theoretical framework and insights identified in previous research (Zhao, Guo
and Liu, 2021; Sanusi et al., 2022a; Kim and Kwon, 2023; Ng et al., 2023), we developed a
series of research hypotheses. Our fundamental premise assumes that a positive attitude
toward AI exerts a generally positive influence on educators' competencies in the context of
developing artificial intelligence systems. The research hypotheses are as follows.
 Hypothesis 1 (H1): The positive attitude of teachers towards AI has a positive and
significant impact on cognitive competencies.
significant impact on digital competencies.
significant impact on educational management competencies.
significant impact on fundamental competencies.

AE
3. Methodology
The main objective of this research is to develop and validate an exploratory model, with the
help of a structural equation, regarding the perception of teaching staff regarding the role of
educators in modelling competencies in the context of the development of artificial
intelligence systems. Moreover, starting from the results of the structural equation modelling
(SEM), we will follow the challenges and measures that teachers consider to be the most
important in the context of the development of artificial intelligence systems. Before the
implementation of this questionnaire, we conducted a pilot study on a group of 20 teachers,
consisting of 10 teaching staff from pre-university education and 10 from higher education
institutions. This served to verify the accuracy of the questions and subsequently the
questionnaire was revised according to the comments received. In carrying out the research,
we used the survey method, starting from the elaboration of the questionnaire and its
distribution in the on-line environment. The questionnaire comprised two sections, each
consisting of closed-ended questions, was designed by using the online application Google
Forms and was distributed online. The initial section of the questionnaire focused on
collecting demographic information from the participants, while the subsequent section
delved into specific inquiries concerning educators' roles in the context of the development
of artificial intelligence systems. Data collection occurred in August 2023, resulting in a
convenience sampling comprised of 138 respondents who are educators in both preuniversity
and university settings within Bihor County (see Table 1). The convenience sampling, which
typically relies on volunteer participants, is a widely used method in research (Maxwell and
Delaney, 2004), and it is also acknowledged for its relatively lower degree of rigour. This
approach does have its limitations, notably its potential to introduce bias since it does not
necessarily represent the broader population. To mitigate these limitations, we employ direct
email outreach for participant recruitment and implement the “snowball” technique to expand
our sample size.
Table no. 1. Profile of the respondents

Characte- Characte- Abs. Rel. fr.
Category Abs.fr. Rel. fr. (%) Category
ristics ristics fr. (%)
Male 26 20%
Gender
Female 111 80%
0-3 Years 10 7%
[20-30) Years 8 6%
Work 3-9 Years 13 9%
[30-40) Years 33 24%
experience 10-20 Years 45 33%
Age [40-50) Years 58 42%
Over 20 years 70 51%
[50-60) Years 35 25%
60 years or older 4 3%
Preschool 4 3%
Primary school 29 21%
Teaching Profile of Real 58 42%
Low secondary 15 11%
level specialisation Humanistic 80 58%
Higher secondary 44 32%
University 46 33%
Urban 119 86% Teaching Urban 125 91%
Residence
Rural 19 14% location Rural 13 9%
TOTAL 138 100% TOTAL 138 100%
Note: Abs. fr. – the absolute frequency, Rel. fr. – the relative frequency
Source: own contribution
Vol. 26 • No. 65 • February 2024 187

The structure of the sample is presented in Table no. 1 provides a broad perspective on the
aspects investigated, both through the area of specialisation in which the teacher is active and
through the professional experience of the respondents. 138 Bihor County teachers
participated in the study. The questionnaire that is the basis of this research contains 27 items
regarding the attitude of teachers toward AI and the competencies (digital, cognitive,
fundamental and educational management) they should develop in order to promote AI
among students. For a unified understanding, respondents considered AI as a generic term
for any educational application based on artificial intelligence that can be used in the
teaching-learning process. The respondents were invited to rate the items presented in Table
no. 2 using a five-point Likert scale (1 - Strongly Disagree; 5 - Strongly Agree). Additionally,
the questionnaire includes 8 items related to the challenges facing education in the
development of AI systems and 8 items related to a series of measures to manage the
consequences of the integration of AI into society. Regarding the global score of Cronbach's
Alpha, we obtained a good value of 0.890, which shows a good internal consistency of the
items. The results on the subscales can be found in Table no. 2.
Table no. 2. Description of the constructs and reliability analysis
Cronbach’
Factor rho_A CR AVE
Item Measure s Alpha
Loading (> 0.7) (> 0.8) (> 0.5)
(> 0.7)
Positive attitude of teachers towards AI (ATT)
adapted from Sanusi et al. (2022a); Kim and Kwon (2023)
AI will help improve learning
ATT1 0.871
processes
AI will help tailor learning to
ATT2 0.875 0.773 0.779 0.870 0.691
the individual needs of learners
AI will help reduce inequalities
ATT3 0.741
in education
Digital competencies (DC)
adapted from Ng et al. (2023); Kim and Kwon (2023)
Competence in the use of the
DC1 0.905
computer
Competencies in using the
DC2 0.936
Internet
DC3 Digital literacy 0.962 0.971 0.979 0.977 0.875
DC4 Information literacy 0.957
DC5 Blended learning 0.932
DC6 Online research competencies 0.919
Cognitive competencies (CC)
adapted from Sanusi et al. (2022a); Kim and Kwon (2023)
Critical thinking to explain
CC1 0.918
complex AI concepts
Adaptability and continuous
CC2 learning in relation to AI 0.941
advances
0.966 0.978 0.972 0.854
Encouraging creativity and
CC3 innovative thinking in AI 0.933
development
Motivating students in AI
CC4 0.911
development

AE
Cronbach’
Factor rho_A CR AVE
Item Measure s Alpha
Loading (> 0.7) (> 0.8) (> 0.5)
(> 0.7)
Hands-on experience in AI
CC5 0.920
development
Correlating AI concepts with
CC6 0.922
real-world applications
Educational management competencies (EMC)
adapted from Kim and Kwon (2023)
Implementing AI education for
EMC1 all age groups, from all walks of 0.879
life
Integrating AI concepts into
EMC2 0.914
different educational disciplines
Promoting a mindset of lifelong
EMC3 0.866
learning
0.955 0.965 0.964 0.817
Ensuring equitable access to AI
EMC4 0.905
education and resources
Developing AI training
EMC5 0.937
programmes for educators
Collaboration with the business
EMC6 environment to identify AI 0.919
skills
Fundamental competencies (FC)
adapted from Kim and Kwon (2023)
Encourage students to explore
FC1 0.886
unconventional AI solutions
Promoting experimentation and
FC2 0.926
risk-taking in AI projects
Providing students with
FC3 resources for AI ideas and 0.937
innovations
0.964 0.966 0.971 0.847
Mentoring students in the
FC4 development of AI applications 0.934
with social impact
Stimulating brainstorming
FC5 0.932
sessions of AI projects
Presentation of successful
FC6 0.906
experiences in the field of AI
Statistical hypotheses were tested using the Partial Least Square - Structural Equation
Modelling (PLS-SEM) method, and the data were analysed using the SmartPLS 3.3.9
statistical software. This method allows the identification of cause-effect relationship models
even when considering a complex set of elements (Benitez et al., 2020), respectively, when
we have a small sample size (Hair et al., 2017). The elements of the constructs, the attitude
of teachers to AI (3 elements), digital competencies (6 elements), cognitive competencies (6
elements), fundamental competencies (6 elements) and educational management
competencies (6 items) were adapted from Zhao, Guo, and Liu (2021), Sanusi et al. (2022a),
Kim and Kwon (2023), Ng et al. (2023).
Vol. 26 • No. 65 • February 2024 189

Regarding the examination of the reliability and validity of the items from the five constructs,
Cronbach's alpha, rho_A, and composite reliability were analysed. According to the results
presented in Table no. 2, all values of Cronbach’s alpha and rho_A were above the
recommended standard value of 0.70, and the composite reliability (CR) was above the
recommended standard value of 0.80 (Cronbach and Shavelson, 2004; Chin, 2010).
Therefore, the reliability of all factors analysed was established. We used the Average
Variance Extracted (AVE) in order to test the discriminant validity of the model. The values
of Average Variance Extracted for each construct are higher than the acceptable value of 0.50
(Chin, 2010) which confirms the discriminant validity of the proposed model.
To effectively harness the potential of AI in education, educators must understand AI
technologies, adapt teaching methods, and stay abreast of ever-evolving AI advances. Therefore,
the integration of AI in education is not without challenges. Starting from the results obtained
following the analysis of the impact of positive attitudes toward AI on the competencies of
educators, in this paper, we propose to present the challenges faced by educators and the
measures proposed by them to make the potential of AI in education more efficient.

4.1. The relationship between the positive attitude of teachers towards AI and the
competencies of teachers
The structural model with path coefficient is presented in Figure no. 1. The results obtained
and presented in Figure no. 1 show that within the proposed structural model, all the
estimated coefficients are positive, which means that the teachers' positive attitude towards
AI positively influences the 4 categories of competencies analysed.
Figure no. 1. Model of the relationship between the positive attitude of teachers
towards AI and their competencies (Structural model 1)

AE
Thus, among the four constructs, the construct most influenced by the positive attitude of
teachers toward AI is the construct related to educational management competencies (β =
0.465, R-squared = 0.216, p-value = 0.002), followed by fundamental competencies. (β =
0.448, R-squared = 0.201, p-value = 0.014), cognitive competencies (β = 0.369, R-squared =
0.136, p-value = 0.025), digital competencies (β = 0.170, R-squared = 0.029, p-value =
0.220). Since the R-squared in the case of digital competencies is not statistically significant
(p-value > 0.10), we removed this construct from the model and we can argue that the H2
hypothesis is rejected. This outcome is corroborated by several other studies, such as those
conducted by Ng et al. (2023), Vlasova et al. (2019) and Chounta et al. (2022). Despite their
limited proficiency in AI-related domains, educators have demonstrated a willingness to
endorse AI adoption. However, they have voiced concerns regarding the substantial effort
required to acquire the necessary skills for proficiently using AI technologies and potential
concerns about trust when employing AI tools. Therefore, it appears that teachers do not
regard their digital competencies as adequate to address educational issues related to AI. The
new proposed model is as follows:
Figure no. 2. Model of the relationship between teachers' positive attitude towards AI
and their cognitive, fundamental and educational management competencies
(Structural model 2)
By using the standardised mean square root (SRMR), we evaluate the model fit. The SRMR
value for the second estimated model is 0.045, less than 0.08, indicating a good model fit.
Therefore, regarding the positive attitude towards AI of teachers, the more they agree with
the statements that hold the AI development systems in high regard, nevertheless they
recommend more the use of cognitive, fundamental and educational management
Vol. 26 • No. 65 • February 2024 191

competencies in teaching development AI. Considering that all the indexes indicate a good
fit of the model, we further tested the hypotheses H1, H3 and H4 with the help of PLS-SEM.
We consider the procedure of bootstrapping (Table no. 3).
According to the results presented in Table no. 3, we can argue that the H1 hypothesis:
ATT → CC (teachers’ attitude towards AI influences cognitive competencies) is accepted
(β = 0.369; p - value < 0.01). Regarding the hypothesis: ATT → EMC (teachers’ attitude
towards AI influences educational management competencies), because β = 0.466 and
p - value < 0.01, we can confirm that the hypothesis H3 is accepted. Also, the H4 hypothesis:
ATT → FC (teachers’ attitude toward AI influences fundamental competencies) is accepted
(β = 0.448; p - value < 0.01). Thus, we can conclude that the H1, H3 and H4 hypotheses are
accepted.
Table no. 3. The influence of the positive attitude of teachers towards AI

on the teachers’ competencies (Path coefficients) – Model 2
Path
Standard
Path coefficient T-statistics p-value Hypothesis
deviation
(β)
ATT → CC 0.369 0.087 4.253 0.000* H1 - accepted
ATT → EMC 0.466 0.077 6.044 0.000* H3 - accepted
ATT → FC 0.448 0.097 4.599 0.000* H4 - accepted
* p-value < 0.01
The presence of AI in both current and future education is undeniable. Educators are
increasingly recognising AI's significance in education and its integral place within the
classroom. This dynamic engenders a dichotomy between teachers and AI: some perceive AI
positively, while others view it negatively. After testing the SEM model, we found a positive
relationship between the attitude of teachers towards AI and cognitive, fundamental and
educational management competencies. These findings align with the studies conducted by
Vinichenko, Melnichuk and Karácsony (2020), Pedro et al. (2019), indicating that educators
who maintain positive perspectives on AI within educational contexts are more predisposed
to increase cognitive, digital, fundamental, and educational management competencies.
Therefore, we can argue that positive proponents seek to cultivate the required competencies
for AI integration, helping students use ethical and judicious AI to enhance educational
outcomes.
4.2. Challenges and proposed measures regarding artificial intelligence

In the evolving context of education, the positive attitude of teachers toward AI plays an
important role in shaping their competencies related to AI development. However, the results
of the second structural model tested in the present research suggest that while this positive
attitude undoubtedly influences their competencies, its impact could be less pronounced in
competencies such as fundamental, cognitive, and educational management. Consequently,
there is a critical need to investigate the challenges faced by educators to improve their
attitudes toward strengthening these competencies. This aims to identify effective measures
used to strengthen teacher attitudes, ultimately equipping them to engage with AI tools in

AE
education more effectively. Therefore, we also looked at teachers' perspectives on the

challenges facing education with the development of AI systems. Eight challenges were
proposed, of which the respondents had to choose the first three most important challenges
education faced with the development of AI systems from their point of view. The results are
presented in Table no. 4.
Table no. 4. Challenges regarding artificial intelligence
No. of
Item Rank
respondents
Teacher professional development needs 87 1
Limited access to educational resources, technology and AI
72 2
infrastructure
Updating the curriculum and educational policies for AI 63 3
Assessing the impact of AI on learning 52 4
Data privacy and security concerns 40 5
Ensuring equitable access to AI education 36 6
Balancing the use of AI in the teacher-student relationship 35 7
Preparing students for the AI-powered workforce 29 8
Based on the results, we can argue that most of the respondents consider that one of the most
important challenges education faces with the development of AI systems is related to the
need for professional development of teaching staff regarding artificial intelligence systems
(87 respondents). A second important challenge regarding the relationship between education
and AI is the limited access to educational resources, technology and AI infrastructure
(72 respondents). Updates to AI curriculum and educational policies (63 respondents)
represent the third important challenge in terms of the challenges education faces with the
development of AI systems. Therefore, the professional development of educators requires
the prioritisation of comprehensive and continuous professional development programmes
tailored to equip educators with AI related competencies. This could include workshops,
training courses and collaborations with AI experts to ensure that educators are adapted to
using AI technologies for improved learning outcomes. These results are also supported by
other research identified and presented in the theoretical background (Hussin, 2018; Vlasova
et al., 2019; Owoc, Sawicka and Weichbroth, 2021; Zhao, Guo and Liu, 2021). There is also
a need for policy makers to direct efforts toward reducing the digital divide by ensuring
equitable access to educational resources, technology, and AI infrastructure. Initiatives to
establish well-equipped artificial intelligence laboratories, provide access to educational
platforms, and facilitate technological support in disadvantaged regions are essential. In
terms of curriculum update, the integration of AI into curricula requires proactive updates to
align educational content with evolutionary advances in AI. Policymakers should
collaboratively develop curriculum guidelines and educational policies that integrate AI
education across disciplines, encouraging comprehensive AI literacy among students.
Another aspect addressed in this research and which we consider important in the relationship
between education and the development of AI systems is related to the measures that can
manage the consequences of the integration of AI into society. Eight measures were proposed
Vol. 26 • No. 65 • February 2024 193

from which the respondents were asked to select the first three most important measures. The
results are presented in Table no. 5.
Table no. 5. Proposed measures regarding artificial intelligence
Item No. of respondents Rank
Continuous training and professional development of teachers 99 1
Creating AI educational resources 92 2
Quick updates to the curriculum 62 3
Personalised learning platforms 49 4
Government funding and support 38 5
Transparent communication 35 6
Digital inclusion initiatives 20 7
Curricular integration planning 19 8
Teachers believe that among the top 3 most important measures that can manage the
consequences of AI integration in society are: Continuous training and professional
development of teachers (99 respondents), Creation of educational AI resources (92
respondents) and Quick updates to the curriculum (62 respondents). We note that these
measures are consistent with the challenges selected by the respondents regarding the
development of AI systems. Therefore, the creation and dissemination of AI-focused
educational resources should be a central policy focus. This involves developing
comprehensive AI learning modules, digital libraries, and interactive platforms to facilitate
educators' access to up-to-date AI content. In addition, policymakers must prioritise the
establishment of comprehensive and targeted professional development programmes for
educators. These initiatives should encompass AI-specific training, workshops, and
collaborative partnerships with industry experts to empower educators with the requisite
skills to effectively incorporate AI into the learning process.
Conclusions
The results presented demonstrate a positive correlation between teacher attitudes towards
AI and the development of competencies across various categories. The most significant
influence is observed in the field of educational management competencies, followed by
fundamental and cognitive competencies. It follows that a positive attitude toward AI
corresponds to an emphasis on cognitive, fundamental, and educational management
competencies in educational practices. In the case of digital competencies, the results showed
that the positive attitude towards AI does not influence these categories of competencies.
According to the obtained results, we can argue that only the H1, H3 and H4 hypotheses are
accepted. The influence of teachers' positive attitudes towards AI on cognitive, fundamental,
and educational management competencies is statistically significant. This underscores the
importance of fostering positive attitudes towards AI among educators for the development
of various competencies.
The research identifies several challenges in integrating AI into education, such as the need
for educators' professional development and equitable access to educational resources and
technology. Policy makers should prioritise designing comprehensive professional

AE
development programmes tailored to equip educators with AI-related competencies.

Initiatives to bridge the digital divide and integrate AI education into curriculum are also
essential policy implications. Teachers' perceptions of effective measures to manage AI's
societal consequences underscore the importance of continuous training, creation of
educational AI resources, and curriculum updates. These findings align with the identified
challenges, underlining the need to develop and disseminate AI-focused educational
resources and establish targeted professional development programmes for educators. The
integration of AI in education requires a multifaceted approach encompassing legal
frameworks, resource allocation, tailored educator training, and curricular adaptations. Such
efforts are essential to harness AI's transformative potential and effectively prepare students
for an increasingly AI-driven future. This comprehensive approach ensures that AI becomes
an integral part of the educational landscape, allowing students to navigate the complexities
of a technology-rich world while preserving the invaluable human touch in the learning
process.
The original contribution of the research consists of a unified approach of the relationship
between the positive attitudes of teachers towards artificial intelligence and a series of skills
(cognitive, fundamental, and educational management), which addresses the fact that a large
part of the literature addresses only one category of competencies in relation to teachers'
attitude towards artificial intelligence and ignores the unity of competencies.
This research is not without challenges and limitations. The first limitation of this research is
the relatively small number of survey respondents. Other limitations of the research refer to
the way in which the items referring to the positive attitude of teaching staff with regard to
AI and the relatively small number of items for this construct were selected. Additionally,
the scope of the study is limited to a single region and generalisation of the findings to other
regions may require further investigation. In this research, we focus on the positive attitude
of teachers towards AI, and in future research, we propose to address a critical analysis from
teachers towards AI, including their reservations towards the expansion of AI in education.
Future research could address the limitations of this study by expanding the sample size and
improving the measurement of teaching staff attitudes and competencies on AI. Additionally,
further investigations could examine the specific content and delivery methods of
professional development programmes aimed at equipping educators with AI related
competencies and knowledge.
References
Adams, C., Pente, P., Lemermeyer, G. and Rockwell, G., 2023. Ethical principles for
artificial intelligence in K-12 education. Computers and Education: Artificial
Ahmad, S.F., Han, H., Alam, M.M., Rehmat, Mohd.K., Irshad, M., Arraño-Muñoz, M. and
Ariza-Montes, A., 2023. Impact of artificial intelligence on human loss in decision
making, laziness and safety in education. Humanities and Social Sciences
Communications, 10(1), p. 311. https://doi.org/10.1057/s41599-023-01787-8.
Akgun, S. and Greenhow, C., 2022. Artificial intelligence in education: Addressing ethical
challenges in K-12 settings. AI and Ethics, 2(3), pp. 431-440.
https://doi.org/10.1007/s43681-021-00096-7.
Vol. 26 • No. 65 • February 2024 195

Alexandre, F., Becker, J., Comte, M.-H., Lagarrigue, A., Liblau, R., Romero, M. and
Viéville, T., 2021. Why, What and How to Help Each Citizen to Understand Artificial
Intelligence? KI – Künstliche Intelligenz, 35(2), pp. 191-199. https://doi.org/10.1007/
s13218-021-00725-7.
Bellman, R., 1978. An Introduction to Artificial Intelligence: Can Computers Think? San
Francisco: Boyd & Fraser Publishing Company.
Benitez, J., Henseler, J., Castillo, A. and Schuberth, F., 2020. How to perform and report an
impactful analysis using partial least squares: Guidelines for confirmatory and
explanatory IS research. Information & Management, 57(2), article no. 103168.
https://doi.org/10.1016/j.im.2019.05.003.
Celik, I., Dindar, M., Muukkonen, H. and Järvelä, S., 2022. The Promises and Challenges of
Artificial Intelligence for Teachers: a Systematic Review of Research. TechTrends, 66(4),
pp. 616-630. https://doi.org/10.1007/s11528-022-00715-y.
Chin, W.W., 2010. How to Write Up and Report PLS Analyses. In: V. Esposito Vinzi, W.W.
Chin, J. Henseler and H. Wang, eds. Handbook of Partial Least Squares. Berlin,
Heidelberg: Springer Berlin Heidelberg. pp. 655-690. https://doi.org/10.1007/978-3-540-
32827-8_29.
Chiu, T.K.F. and Chai, C., 2020. Sustainable Curriculum Planning for Artificial Intelligence
Education: A Self-Determination Theory Perspective. Sustainability, 12(14), article no.
5568. https://doi.org/10.3390/su12145568.
Chiu, T.K.F., Meng, H., Chai, C.-S., King, I., Wong, S. and Yam, Y., 2022. Creation and
Evaluation of a Pretertiary Artificial Intelligence (AI) Curriculum. IEEE Transactions on
Education, 65(1), pp. 30-39. https://doi.org/10.1109/TE.2021.3085878.
Chiu, T.K.F., Xia, Q., Zhou, X., Chai, C.S. and Cheng, M., 2023. Systematic literature review
on opportunities, challenges, and future research recommendations of artificial
intelligence in education. Computers and Education: Artificial Intelligence, 4, article no.
Chounta, I.-A., Bardone, E., Raudsep, A. and Pedaste, M., 2022. Exploring Teachers’
Perceptions of Artificial Intelligence as a Tool to Support their Practice in Estonian K-12
Education. International Journal of Artificial Intelligence in Education, 32(3),
pp. 725-755. https://doi.org/10.1007/s40593-021-00243-5.
Cronbach, L.J. and Shavelson, R.J., 2004. My Current Thoughts on Coefficient Alpha and
Successor Procedures. Educational and Psychological Measurement, 64(3), pp. 391-418.
https://doi.org/10.1177/0013164404266386.
Danaher, J., 2018. Toward an Ethics of AI Assistants: an Initial Framework. Philosophy &
Technology, 31(4), pp. 629-653. https://doi.org/10.1007/s13347-018-0317-3.
Dillenbourg, P., 2013. Design for classroom orchestration. Computers & Education, 69,
pp. 485-492. https://doi.org/10.1016/j.compedu.2013.04.013.
Dillenbourg, P., 2016. The Evolution of Research on Digital Education. International
Journal of Artificial Intelligence in Education, 26(2), pp. 544-560. https://doi.org/
10.1007/s40593-016-0106-z.
Dong, Y., Hou, J., Zhang, N. and Zhang, M., 2020. Research on How Human Intelligence,
Consciousness, and Cognitive Computing Affect the Development of Artificial
Intelligence. Complexity, 2020, pp. 1-10. https://doi.org/10.1155/2020/1680845.

AE
European Commission, 2022. Ethical guidelines on the use of artificial intelligence (AI) and
data in teaching and learning for Educators. Luxembourg: Publications Office of the
European Union.
Felea, M. and Vasiliu, C., 2005. Aspects of the Commercial Education in the European
Higher Education. Amfiteatru economic, 7(17), pp. 117-120.
Future of Life Institute, 2023. Pause Giant AI Experiments: An Open Letter, [online]
Available at: <https://futureoflife.org/open-letter/pause-giant-ai-experiments/>
[Accessed 17 July 2023].
González-Calatayud, V., Prendes-Espinosa, P. and Roig-Vila, R., 2021. Artificial
Intelligence for Student Assessment: A Systematic Review. Applied Sciences, 11(12),
p. 5467. https://doi.org/10.3390/app11125467.
Hair, J.F., Hult, G.T., Ringle, C.M. and Sarstedt, M., 2017. A primer on partial least squares
structural equation modeling (PLS-SEM). 2nd ed. Thousand Oaks: SAGE.
Haugeland, J., 1989. Artificial Intelligence: The Very Idea. The MIT Press.
https://doi.org/10.7551/mitpress/1170.001.0001.
Holstein, K., McLaren, B.M. and Aleven, V., 2019. Co-Designing a Real-Time Classroom
Orchestration Tool to Support Teacher–AI Complementarity. Journal of Learning
Analytics, 6(2), pp. 27-52. https://doi.org/10.18608/jla.2019.62.3.
Hrastinski, S., Olofsson, A.D., Arkenback, C., Ekström, S., Ericsson, E., Fransson, G.,
Jaldemark, J., Ryberg, T., Öberg, L.-M., Fuentes, A., Gustafsson, U., Humble, N.,
Mozelius, P., Sundgren, M. and Utterberg, M., 2019. Critical Imaginaries and Reflections
on Artificial Intelligence and Robots in Postdigital K-12 Education. Postdigital Science
and Education, 1(2), pp. 427-445. https://doi.org/10.1007/s42438-019-00046-x.
Huang, X., 2021. Aims for cultivating students’ key competencies based on artificial
intelligence education in China. Education and Information Technologies, 26(5),
pp. 5127-5147. https://doi.org/10.1007/s10639-021-10530-2.
Hussin, A., 2018. Education 4.0 made simple: Ideas for teaching. International Journal of
Education and Literacy Studies, 6(3), pp. 92-98. https://doi.org/10.7575/aiac.ijels.
v.6n.3p.92.
Kim, J., Lee, H. and Cho, Y.H., 2022. Learning design to support student-AI collaboration:
perspectives of leading teachers for AI in education. Education and Information
Technologies, 27(5), pp. 6069-6104. https://doi.org/10.1007/s10639-021-10831-6.
Kim, K. and Kwon, K., 2023. Exploring the AI competencies of elementary school teachers
in South Korea. Computers and Education: Artificial Intelligence, 4, article no. 100137.
Klimova, B., Pikhart, M. and Kacetl, J., 2023. Ethical issues of the use of AI-driven mobile
apps for education. Frontiers in Public Health, 10, article no. 1118116.
https://doi.org/10.3389/fpubh.2022.1118116.
Lameras, P. and Arnab, S., 2021. Power to the Teachers: An Exploratory Review on Artificial
Intelligence in Education. Information, 13(1), article no.. 14. https://doi.org/10.3390/
info13010014.
Li, M. and Su, Y., 2020. Evaluation of Online Teaching Quality of Basic Education Based
on Artificial Intelligence. International Journal of Emerging Technologies in Learning
(iJET), 15(16), article no. 147. https://doi.org/10.3991/ijet.v15i16.15937.
Vol. 26 • No. 65 • February 2024 197

Limna, P., Somporch, J., Sutithep, S., Pichart, K. and Patcharavadee, S., 2022. A Review of
Artificial Intelligence (AI) in Education during the Digital Era. Advance Knowledge for
Executives, 1(3), pp. 1-9.
Lu, X., 2019. An Empirical Study on the Artificial Intelligence Writing Evaluation System
in China CET. Big Data, 7(2), pp. 121-129. https://doi.org/10.1089/big.2018.0151.
Luckin, R., 2017. Towards artificial intelligence-based assessment systems. Nature Human
Behaviour, 1(3), article no. 0028. https://doi.org/10.1038/s41562-016-0028.
Malekos, N., 2023. The Rise of Artificial Intelligence in Education: Will A.I. Disrupt
eLearning? [online] Available at: <https://www.learnworlds.com/artificial-intelligence-
in-education/> [Accessed 12 July 2023].
Maxwell, S. E. and Delaney, H. D., 2004. Designing Experiments and Analyzing Data: A
Model Comparison Perspective (2nd Edition). Mahwah, NJ: Lawrence Erlbaum
Associates Publishers.
McCarthy, T., Rosenblum, L.P., Johnson, B.G., Dittel, J. and Kearns, D.M., 2016. An
Artificial Intelligence Tutor: A Supplementary Tool for Teaching and Practicing
Braille. Journal of Visual Impairment & Blindness, 110(5), pp. 309-322.
https://doi.org/10.1177/0145482X1611000503.
Mena-Guacas, A.F., Urueña Rodríguez, J.A., Santana Trujillo, D.M., Gómez-Galán, J. and
López-Meneses, E., 2023. Collaborative learning and skill development for educational
growth of artificial intelligence: A systematic review. Contemporary Educational
Technology, 15(3), article no. ep428. https://doi.org/10.30935/cedtech/13123.
Miao, F., Holmes, W., Huang, R. and Zhang, H., 2021. AI and education: A guidance for
policymakers. Paris: UNESCO.
Mitchell, M., 2019. Artificial Intelligence: A Guide for Thinking Humans. London: Pelican
Books.
Moisil, I., Pitic, A., Dzitac, S. and Popper, L., 2010. Adaptive Web Applications for Citizens’
Education. Case Study: Teaching Children the Value of Electrical Energy. International
Journal of Computers Communications & Control, 5(5), article no.. 819.
https://doi.org/10.15837/ijccc.2010.5.2242.
Murphy, R., 2019. Artificial intelligence applications to support k–12 teachers and teaching:
a review of promising applications, challenges, and risks, [online] Santa Monica, CA:
RAND Corporation. Available at: <https://www.rand.org/pubs/perspectives/
PE315.html> [Accessed 14 June 2023].
Ng, D.T.K., Leung, J.K.L., Chu, S.K.W. and Qiao, M.S., 2021. Conceptualizing AI literacy:
An exploratory review. Computers and Education: Artificial Intelligence, 2, article no.
Ng, D.T.K., Leung, J.K.L., Su, J., Ng, R.C.W. and Chu, S.K.W., 2023. Teachers’ AI digital
competencies and twenty-first century skills in the post-pandemic world. Educational
technology research and development, 71(1), pp. 137-161. https://doi.org/10.1007/
s11423-023-10203-6.
OECD, 2021. Recommendation of the Council on Artificial Intelligence, [online] Available
at: <https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449> [Accessed
22 July 2023].

AE
Owoc, M.L., Sawicka, A. and Weichbroth, P., 2021. Artificial Intelligence Technologies in
Education: Benefits, Challenges and Strategies of Implementation. In: M.L. Owoc and
M. Pondel, eds. Artificial Intelligence for Knowledge Management. Cham: Springer
International Publishing. pp. 37-58. https://doi.org/10.1007/978-3-030-85001-2_4.
Ozdemir, O. and Tekin, A., 2016. Evaluation of the presentation skills of the pre-service
teachers via fuzzy logic. Computers in Human Behavior, 61, pp. 288-299.
https://doi.org/10.1016/j.chb.2016.03.013.
Pedro, F., Subosa, M., Rivas, A. and Valverde, P., 2019. Artificial intelligence in education:
Challenges and opportunities for sustainable development, [online] Available at:
<https://repositorio.minedu.gob.pe/handle/20.500.12799/6533> [Accessed 10 June
2023].
Pelau, C., Dabija, D.-C. and Ene, I., 2021. What makes an AI device human-like? The role
of interaction quality, empathy and perceived psychological anthropomorphic
characteristics in the acceptance of artificial intelligence in the service
industry. Computers in Human Behavior, 122(2), article no. 106855.
https://doi.org/10.1016/ j.chb.2021.106855.
Pinkwart, N., 2016. Another 25 Years of AIED? Challenges and Opportunities for Intelligent
Educational Technologies of the Future. International Journal of Artificial Intelligence
in Education, 26(2), pp. 771-783. https://doi.org/10.1007/s40593-016-0099-7.
Polak, S., Schiavo, G. and Zancanaro, M., 2022. Teachers’ Perspective on Artificial
Intelligence Education: an Initial Investigation. In: CHI Conference on Human Factors
in Computing Systems Extended Abstracts. New Orleans LA USA: ACM. pp. 1-7.
https://doi.org/10.1145/3491101.3519866.
Prieto, L.P., Sharma, K., Kidzinski, Ł., Rodríguez‐ Triana, M.J. and Dillenbourg, P., 2018.
Multimodal teaching analytics: Automated extraction of orchestration graphs from
wearable sensor data. Journal of Computer Assisted Learning, 34(2), pp. 193-203.
https://doi.org/10.1111/jcal.12232.
Qian, L., Zhao, Y. and Cheng, Y., 2020. Evaluating China’s Automated Essay Scoring
System iWrite. Journal of Educational Computing Research, 58(4), pp. 771-790.
https://doi.org/10.1177/0735633119881472.
Qin, Y., Xu, Z., Wang, X. and Skare, M., 2023. Artificial Intelligence and Economic
Development: An Evolutionary Investigation and Systematic Review. Journal of the
Knowledge Economy. https://doi.org/10.1007/s13132-023-01183-2.
Regan, P.M. and Jesse, J., 2019. Ethical challenges of edtech, big data and personalized
learning: twenty-first century student sorting and tracking. Ethics and Information
Technology, 21(3), pp. 167-179. https://doi.org/10.1007/s10676-018-9492-2.
Sadiku, M., Ashaolu, T., Ajayi-Majebi, A. and Musa, S., 2021. Artificial Intelligence in
Education. International Journal of Scientific Advances, 2(1), pp. 5-11.
Sanusi, I.T., Olaleye, S.A., Agbo, F.J. and Chiu, T.K.F., 2022a. The role of learners’
competencies in artificial intelligence education. Computers and Education: Artificial
Sanusi, I.T., Olaleye, S.A., Oyelere, S.S. and Dixon, R.A., 2022b. Investigating learners’
competencies for artificial intelligence education in an African K-12 setting. Computers
and Education Open, 3, article no.100083. https://doi.org/10.1016/j.caeo.2022.100083.
Vol. 26 • No. 65 • February 2024 199

Schiff, D., 2021. Out of the laboratory and into the classroom: the future of artificial
intelligence in education. AI & Society: Knowledge, Culture and Communication, 36(1),
pp. 331-348. https://doi.org/10.1007/s00146-020-01033-8.
Schwarz, B.B., Prusak, N., Swidan, O., Livny, A., Gal, K. and Segal, A., 2018. Orchestrating
the emergence of conceptual learning: a case study in a geometry class. International
Journal of Computer-Supported Collaborative Learning, 13(2), pp. 189-211.
https://doi.org/10.1007/s11412-018-9276-z.
Sobolu, R., Stanca, L. and Bodog, S.A., 2023. Automated Recognition Systems: Theoretical
and Practical Implementation of Active Learning for Extracting Knowledge in Image-
based Transfer Learning of Living Organisms. International Journal of Computers
Communications & Control, 18(6), article no. 5728. https://doi.org/10.15837/ijccc.
2023.6.5728.
Talan, T., 2021. Artificial Intelligence in Education: A Bibliometric Study. International
Journal of Research in Education and Science, 7(3), pp. 822-837.
https://doi.org/10.46328/ijres.2409.
Tondeur, J., Scherer, R., Siddiq, F. and Baran, E., 2020. Enhancing pre-service teachers’
technological pedagogical content knowledge (TPACK): a mixed-method
study. Educational Technology Research and Development, 68(1), pp. 319-343.
https://doi.org/10.1007/s11423-019-09692-1.
Touretzky, D., Gardner‐ McCune, C., Breazeal, C., Martin, F. and Seehorn, D., 2019. A Year
in K–12 AI Education. AI Magazine, 40(4), pp. 88-90. https://doi.org/10.1609/
aimag.v40i4.5289.
UNESCO, 2019. Artificial Intelligence in Education: Challenges and Opportunities for
Sustainable Development. Paris: UNESCO Education Sector.
Vinichenko, M.V., Melnichuk, A.V. and Karácsony, P., 2020. Technologies of improving
the university efficiency by using artificial intelligence: motivational aspect.
Entrepreneurship and Sustainability Issues, 7(4), pp. 2696-2714. https://doi.org/10.9770/
jesi.2020.7.4(9).
Vlasova, E. Z., Avksentieva, E. Y., Goncharova, S. V. and Aksyutin, P.A., 2019. Artificial
intelligence-The space for the new possibilities to train teachers. Espacios, 40(9), article
no. 17.
Xia, Q., Chiu, T.K.F., Lee, M., Sanusi, I.T., Dai, Y. and Chai, C.S., 2022. A self-
determination theory (SDT) design approach for inclusive and diverse artificial
intelligence (AI) education. Computers & Education, 189, article no. 104582.
https://doi.org/10.1016/j.compedu.2022.104582.
Zhao, X., Guo, Z. and Liu, S., 2021. Exploring Key Competencies and Professional
Development of Music Teachers in Primary Schools in the Era of Artificial Intelligence.
Scientific Programming, 2021, pp. 1-9. https://doi.org/10.1155/2021/ 5097003.

AE
A SYSTEMATIC ANALYSIS OF NEW APPROACHES TO DIGITAL

ECONOMIC EDUCATION BASED ON THE USE OF AI TECHNOLOGIES
Laurențiu Mihai1 , Leonardo-Geo Mănescu2 , Laura Vasilescu3 ,
Anca Băndoi4 and Cătălina Sitnikov5
1)2)3)4)5)
University of Craiova, Romania

Mihai, L., Mănescu, L.G., Vasilescu, L., Băndoi, A. Received: 29 September 2023
and Sitnikov, C., 2024. A Systematic Analysis of New Revised: 19 November 2023
Approaches to Digital Economic Education Based on Accepted: 20 December 2023
the Use of AI Technologies. Amfiteatru Economic,
26(65), pp. 201-219.
Abstract
Since the start of the global COVID-19 pandemic, most higher education institutions have
been forced to exchange the traditional teaching environment for online education, and many
chose to continue to use digital education platforms after its end, especially through the use of
artificial intelligence (AI) applications and technologies. Our research represents a systematic
literature review of a number of 60 scientific papers, aiming to study how the concept of digital
economic education based on artificial intelligence is approached in the scientific literature,
how artificial intelligence applications are used in digital economic education, and which are
the critical success factors and the challenges that this domain is facing. Our findings have
shown that most researchers define digital education as the use of technology to support
educational activities, while highlighting artificial intelligence and its different applications as
an essential element of current digital education, which has the potential to fundamentally
transform the economic processes. The large-scale adoption of e-learning systems based on
artificial intelligence is influenced by technology, by their superior capabilities in terms of
coherently correlating the learning and studying processes, by the teachers’ trust in the results
generated by these technologies and by cultural factors, while facing several challenges related
to the users’ resistance to change, digital competences, the systems’ accessibility, as well as
financial issues. Furthermore, based on this research endeavour, a model system of
correlations and elements has been developed, specifically for the digital economic education
based on artificial intelligence. This model includes both the success factors and the unique
challenges inherent in the particular application areas.
Keywords: Artificial Intelligence (AI), digital economic education, systemic model
regarding digital economic education based on AI
JEL Classification: D83, I21, I23, O33

Corresponding author, Laurențiu Mihai – e-mail: [email protected]
Vol. 26 • No. 65 • February 2024 201

AE A Systematic Analysis of New Approaches to Digital Economic Education
Based on the Use of AI Technologies
Introduction based on the current state of knowledge

Artificial intelligence represents a critical element of the contemporary surge of emerging
technologies, which, lately, have significantly transformed the landscape of economics
education. Artificial intelligence (AI) refers to the ability of a software application to execute
tasks commonly attributed to intelligent beings (Salas Pilco and Yang, 2022), comprising
several branches, including digital communication, machine learning, big data analysis and
processing, and natural language processing (Chen et al., 2020). The rapid expansion of this
technology is progressively reshaping the way individuals interact, communicate, reside,
acquire knowledge, and engage in professional activities.
Artificial intelligence has various applications within the realm of digital economic
education. For example, AI has been included in several educational technologies, including
chatbots (Kasneci et al., 2023), intelligent tutoring systems, and automated assessment
software (Hwang and Tu, 2021). According to Chen et al. (2020), AI-based systems offer
increased opportunity for all individuals involved in the process of economics learning and
training. Previous studies have demonstrated the benefits of incorporating AI into economics
education, such as enhanced collaboration among students, customised learning experiences,
efficient scheduling of learning activities (Chatterjee and Bhattarcharjee, 2020), adaptive
feedback on learning processes (Salas Pilco and Yang, 2022), reduced administrative burden
for teachers (Xu and Ouyang, 2022), and automation of exam evaluation and other
educational tasks specific to economics education (Maslova et al., 2020).
This paper focuses on AI in Education (AIED), which encompasses the use of artificial
intelligence technologies, including intelligent tutoring systems, chatbots, digital assistants,
and other digital automation components that contribute to the advancement and
enhancement of digital economic education (Micu et al., 2021; Bearman et al., 2022). AIED
possesses significant potential for enhancing learning, teaching, evaluation, and
administration in economics higher learning institutions, offers the opportunity to customise
and adapt the economics educational content to the specific requirements of students, while
supporting academics in the educational process, making it a significant emerging technology
in the field of digital economic education (Felix, 2021; Ooi et al., 2023).
The current research addresses AIED as a central element in optimising teaching-learning
processes, yielding significant benefits in personalising the educational experience (Felix,
2021). AIED applications demonstrate the capacity to analyse substantial volumes of data
generated during the learning process, discerning patterns and adapting content based on
individual learners' needs. This adaptability and customisation of instruction represent a notable
advantage of employing AIED, contributing to the optimisation of levels of knowledge
assimilation (Chen et al., 2022). Thus, the use of AI in digital education opens substantial
perspectives for enhancing the efficiency and relevance of the educational process in the current
context characterised by accelerated technological advancements (Celik et al., 2022).
Thus, by conducting a systematic review of the relevant literature published in the last four
years (2020-2023), our paper aims to contribute to the ongoing research on AIED. Our
research endeavour focused on answering four research questions: (1) How is the concept of
digital economic education based on AI presented in the recent literature; (2) How is AI used
in digital economic education? (3) What are the critical success factors influencing the use
of artificial intelligence applications? (4) What are the challenges that the use of AI
applications in digital economic education is facing?

AE
In order to answer the research questions, this study will discuss on the general features of
AIED, the AI’s impact on the economics learning process, the factors which affect the
adoption of these technologies, the most used AIED applications as well as the challenges
that these technologies have to overcome in order to be adopted on a larger scale.
Consequently, the paper is structured into two main sections: Research Methodology which
refers to search query used in the Web of Science database, the inclusion and exclusion
criteria, the qualitative assessment and the coding process, and Results and Discussion which
includes: the structure on categories of the papers included in the analysis, the concept of
economics digital education based on AI, the use of AI in economics education, critical
success factors affecting the use of AI applications, and the challenges which economics
digital education based on AI is facing and finally, the conclusions and recommendations
formulated considering the research results are presented.
1. Research Methodology
1.1. Search query used in the Web of Science database
In order to fulfil our research objective and to answer the four research questions, we have
conducted a comprehensive search of papers that were related to AIED, using the Web of
Science Database. The search query used the following string ((TS=“e-learning” OR
TS=“digital economics education” OR TS=“online economics learning”) AND (TS=“digital
tools” OR TS=“AI” OR TS = Artificial Intelligence”)), which yielded 12,962 results.
In order to conduct a first stage analysis, we have used VOSViewer software, which has
generated several bibliometric maps, based on which we concluded that the journals which
have published the most papers on the subject of AIED are Education and Information
Technologies (416 papers), followed by Education Sciences (370 papers) and International
Journal of Emerging Technologies in Learning (175 papers).
At the same time, the most cited journal (as of July 2023) is Education and Information
Technologies (2926 citations), followed by Computer and Education (2245 citations),
Education Sciences (1980 citations), International Journal of Emerging Technologies in
Learning (1209 citations), and Interactive Learning Environments (1128 citations).
Lastly, the most cited papers that address the topic of digital economic education based on
artificial intelligence until July 2023 (Figure no.1) are Radianti et al. (2020) (419 citations),
Köenig et al. (2020) (244 citations), Watermeyer et al. (2021) (310 citations), and Almaiah
et al. (2020) (308 citations).
Vol. 26 • No. 65 • February 2024 203

Figure no. 1. Bibliometric analysis regarding the most cited papers

on the topic of AIED
1.2. Inclusion and exclusion criteria and quality assessment

In order to refine our search, only research articles, conference proceedings, and review paper
(12,843 papers), published after 2020 (6,212 papers) were included, from three Web of
Science categories: Education Educational Research, Education Scientific Disciplines and
Social Issues or Education Special (2,379 papers), that have been written in English (2,604
papers). Moreover, we have excluded all the papers that have been published in medical
journals, resulting in 1,887 papers. Finally, we limited our search only to open access papers,
which yielded a final number of 915 articles. The decision to limit the search to open access
articles was made based on considerations of accessibility, open access articles being
available to all individuals interested in a specific thematic area, enabling them to read and
analyse the content without restrictions.
Moving forward, 183 papers were allocated to each of the five authors. After reading the
titles and the abstracts, those articles which were considered irrelevant to our research were
excluded. In order to be considered relevant, the papers should have focused on artificial
intelligence used as a digital education tool. There were excluded papers which were purely
technical, focused on programming, or focused on technical, humanities, or healthcare.
During this first quality assessment stage, 143 articles were identified that met the selection
criteria.
Among them, each of the five authors thoroughly read 30-31 papers, conducting a critical
analysis of each article, using their experience as reviewers for various scientific journals.
During this second stage of quality assessment, we eliminated the papers which were found
inconclusive, poorly written, employed questionable methods or in which the findings were
not thoroughly presented and discussed, reaching a final number of 59 relevant articles.

AE
Finally, we have included one more paper (Zawacki-Richter et al., 2019) in our analysis,
although it was published before 2020, because it was being cited in many of the articles
included in the final portfolio by July 2023 (409 citations). Therefore, 60 papers were used
in this current research (Figure no. 2).
Figure no. 2. PRISMA sample design diagram for systematic evaluation

Source: adapted after Bond et al., 2020, p.9
1.3. The coding process

This stage aimed to code the 60 articles into four categories: digital education based on AI,
the use of AI in digital education, and critical success factors the challenges this area is facing.
The authors independently coded the selected articles, initially coding the first 10 papers
based on the four categories. After establishing and testing the coding methodology, relevant
information was extracted for further analysis. Other variables included the journal, the
publication year, the paper type, and the digital tool type (Table no.1). All authors
participated in the final coding process, coding 12-13 papers, and reached a consensus
regarding the ambiguous findings.

2.1. The structure of the papers included in the study
Our systematic review encompasses a total of 60 research papers, which have been classified
into four general categories (Table no. 1). Thus, our sample consist of 22 papers categorised
as “The concept digital economics education based on AI”, 24 papers included in the category
“The use of artificial intelligence in digital economic education”, 5 papers categorised as
“Critical success factors affecting the use of Artificial Intelligence applications in digital
economic education” and 7 papers included in the category “Challenges facing the use of
artificial intelligence applications in economic education”. Moreover, one paper was found
to be relevant to both the “The concept of e-learning” and “Critical influence factors”
categories, while another paper was included in both the “Critical influence factors” and
“Challenges” categories.
Vol. 26 • No. 65 • February 2024 205

Table no.1. Paper classification

Variable Percent Frequency Variable Percent Frequency
Publication year AI applications used in the paper
2019 1.67% 1 AI 31.66% 19
2020 41.67% 25 MOOCs, LMSs 46.67% 28
2021 25% 15 Gamification 1.67% 2
2022 23.33% 14 Blockchain 1.67% 1
2023 8.33% 5 VR 1.67% 1
Journal Others 16.66% 10
Int. J. Emerg. 18.33% 11
Subject Category
Technol. Learn.
Educ. Inf. Technol. 18.33% 11 Digital education 38.33% 23
based on AI
Educ. Sci. 8.33% 5 AI in digital 40% 24
educaiton
Comput. Educ. 6.67% 4 Critical success 10% 6
factors
Comput. Educ.: Artif. 5% 3 Challenges 15% 9
Intell.
Others 43.34% 2
2.2. The concept digital economic education based on AI

Digital education, also referred to as online education, e-learning, distance education, or
distance learning, encompasses teaching and learning processes conducted within an online
setting (El-Sabagh, 2021). O'Neill (2023) defines e-learning as the use of technology to aid
instructional activities, with the aim of achieving specified educational goals without the
need for both the learner and the instructor to be physically present in the same area, and
includes various elements, including technologies, educational systems, surroundings, and
the primary users, namely students and teachers.
In the last decade, the educational environment has undergone a significant shift from the
traditional structure to a more engaging and stimulating learning environment (Butler-Henderson
and Crawford, 2020). In this context, the global adoption of digital technologies and tools is
having a profound impact on the methods and practises employed in education. Among the many
digital tools that can be used in online education, artificial intelligence represents an innovative
and revolutionary approach that facilitates the customisation of both educators' and students'
experiences. Consequently, the use of artificial intelligence in education has the potential to
enhance and adapt pedagogical practises by different forms of personalised education, intelligent
content delivery, automation of educational tasks, tutoring, and ensuring inclusive educational
opportunities for students with special requirements.
In order to properly implement an AIED-based platform, several elements need to be considered
such as the students’ involvement (Bond et al., 2020; Theresiawati et al., 2020; El-Sabagh, 2021;
Chiu et al., 2021; Ng et al., 2023), digital literacy (Falloon, 2020; Köenig et al., 2020; Muammar
et al., 2023), satisfaction (Giray, 2021; Shams et al., 2021; Zhao et al., 2021), overall students s’
performance (Butler-Henderson and Crawford, 2020; Rakic et al., 2020; Rajabalee and Santally,

AE
2021), and local factors such as technological resources and internet connectivity (Anthony et
al., 2020; Khalil Awan et al., 2021; Lloret-Irles et al., 2022).
Several authors argue that the concept of e-learning is closely associated with student
engagement (Bond et al., 2020; Theresiawati et al., 2020; Chiu et al., 2021; El-Sabagh, 2021;
Ng et al., 2023), which is driven by motivation, which is, in turn, influenced by the students’
autonomy, competency, and relatedness. Thus, there is a need to enhance student
involvement in terms of behaviour, affect, and cognition, which should be a primary focus
for educators (Ng et al., 2023). Bond et al. (2020) revealed that the most often identified
variables are behavioural, affective, and cognitive engagement. El Sabagh (2021) highlighted
that students’ engagement is closely related to the personalisation of the learning process
according to their individual learning styles, which also positively influences their academic
performance. AI applications have an essential role in this adaptive approach to e-learning,
which aims to promote direct learning, facilitate knowledge development, and enhance the
overall learning experience (Theresiawati et al., 2020; El Sabagh, 2021).
Another element under discussion is the level of student satisfaction which is influenced by
the platforms’ quality (Dangaiso et al., 2020; Shams et al., 2022), design (Giray, 2021) and
accessibility (Shams et al., 2020), as well as the way they are implemented (Almusharraf and
Kharo, 2020). Theresiawati et al. (2020) and Giray (2021) suggest three fundamental
elements which influence the quality of an AI based e-learning system: quality of teaching
staff, the learning management system’s quality, and content quality. Students’ engagement
and satisfaction have been shown to be closely related, as highlighted by Rajabalee et al.
(2020), while Rakic et al. (2020) indicate a noteworthy correlation between students'
academic performance and the utilisation of AI as a digital education tool.
Several papers examined the educators’ experience with AI-based e-learning platforms, as
well as their perception regarding their efficiency as a learning instrument, highlighting
several drawbacks, such as limited time available for teachers to enhance their digital literacy
(Köenig et al., 2020; Khalil Awan et al., 2021), inadequate training (Zhao et al., 2021; Ng et
al., 2023), and improper management of technical issues (Dhillon and Murray, 2021;
Cranfield et al., 2021).
The accelerated adoption of e-learning systems based on artificial intelligence has led to the
recognition of the importance of educators’ digital competence (Falloon, 2020; Köenig et al.,
2021; Zhao et al., 2021). In their works, Ng et al. (2023) and Muammar et al. (2023) have
discussed the DigCompEdu framework, a guiding principle for educators in effectively
integrating resources and designing AI-based learning programmes. The model includes
several elements such as interaction, digital resources (including AI applications), teaching
and learning, assessment, empowering students, and enabling learners’ digital proficiency
(Ng et al., 2023), while aiming to evaluate the academic community’s ICT proficiency and
their current digital and AI related competences (Muammar et al., 2023).
2.3. The use of AI in digital economics education

Our research has shown that several types of digital tools can be used in order to properly
implement an e-learning system, among which AI is highlighted as having the potential to
transform the educational process. Hillmayr et al. (2020) identified five categories of
AI-based digital education tools: exercise and practice programmes, tutoring systems,
Vol. 26 • No. 65 • February 2024 207

hypermedia systems, intelligent tutoring systems and simulations, the last two being the most
effective. Bonfield et al. (2020), discusses other four types of AI applications: smart
campuses, digital assistants, massive open online courses (MOOCs), and Learning
Management Systems (LMSs).
Lizcano et al. (2020) highlighted the existence of a number of other blockchain-based
applications, which enable the decentralised validation of students' acquisition of economics
competencies, ensuring that their training aligns with the prevailing job landscape and the
market demands. Radianti et al. (2020) focuses on virtual reality economics educational
applications, which are shown to be an excellent educational instrument for higher education
but are still in an experimental stage. Abdulaziz Alsuhbhi et al. (2020) dealt with the topic of
introduction of game elements in economics learning systems, which includes the use of
levels, experience points, badges, dashboards, progress bars, content unlocking, and leader
boards. Futhermore, Mhlanga and Moloi (2020) aimed to study the use of AI technologies in
the economics educational sector and their findings suggest that the COVID-19 pandemic
determined the development of virtual learning platforms, the use of educational applications
and websites, the establishment of STEM digital schools, but also a widespread transition to
distance learning.
Bearman et al. (2022) defined artificial intelligence used as a digital education tool (AIED)
as a digital technology capable of revolutionising conventional education, offering more
dynamic and enhanced educational approaches using highly personalised, scalable, and cost-
effective alternative solutions. Xu and Ouyang (2022) defined AIED as an emerging
interdisciplinary domain that uses AI applications to transform educational design and
enhance student learning. According to Lameras and Arnab (2022), AIED covers the
development, implementation, and assessment of tools, pedagogical models, instructional
frameworks, ethical considerations, and teaching staff competencies. Chen et al. (2020) states
that the development and execution of AIED technologies involves the collaboration of
system designers, data scientists, product designers, statisticians, linguists, cognitive
scientists, psychologists, education experts, and numerous other professionals. Felix (2021)
highlights cost and time saving as two main significant advantages of AIED, which is
particularly relevant in the current economic climate characterised by limited financial
resources and the need for effective time management. Using AI to assume some aspects of
the educational process has the potential to minimise salary costs, while allowing faculty
members to focus on other areas, such as teacher-student interpersonal relationship or
academic research which contributes to their professional reputation.
Kasneci et al. (2023) as well as Chen et al. (2020) discussed several AIED systems, such as
natural language processors, collaborative robots, large language models, and chatbots,
designed to either assist teachers in their tasks or operate autonomously, carrying out
activities similar to those of an educator. Xu and Ouyang (2022) argue that AI has the
potential to serve as an educational tool, functioning as a tutor or instructor, thus altering the
dynamics of the instructor-student relationship, shifting the paradigm from an instructor-
centred to a student-centred approach.
Huang (2021) identified three key competencies that students need to develop to properly
benefit from AIED: knowledge competence, team competence, and learning competence.
Huang’s (2021) findings show a negative correlation between teamwork competence and
human-tool collaboration competence and AI course contents, however, being contradicted

AE
by Sanusi et al. (2022), findings which have proven the significance of teamwork and human-
tool collaboration in AIED literacy.
AI applications have been shown to support educators in performing their administrative
tasks much faster, through exam automation (Celik et al., 2022), task evaluation and grading
(Chen et al., 2020), student performance assessment (Zawacki-Richter, 2019; Chiu et al.,
2023), feedback and design support (Zawacki-Richter, 2019; Salas-Pilco and Yang, 2022)
and customisation (Chen et al., 2020). AI-enhanced grading systems have shown superior
objectivity, accuracy, efficiency, and security in evaluating language writing, speaking, and
mathematical assessments compared to human instructors (Chiu et al., 2023), as well as
providing constructive feedback (Salas-Pilco and Yang, 2022; Southworth et al., 2023).
Our research has shown that AI technologies can support the economics learning and
instruction processes through providing adaptive teaching strategies and customised content
(Celik et al., 2022; Lameras and Arnab, 2022), analysis of learner engagement,
understanding, and integrity (Zawacki-Richter, 2019), diagnosing learning difficulties and
student specific problems and addressing them in a timely manner (Hwang and Tu, 2021;
Celik et al., 2022), calculating the odds of students dropping out of school (Salas-Pilco and
Yang, 2022), analysing the course material and providing feedback regarding its
development (Chen et al., 2020; Chiu et al., 2023). Chiu et al. (2023) and Salas-Pilco and
Yang (2022) also highlight the use of AI in the teachers’ own professional development,
through real-time evaluation of their teaching behaviour during regular classes. Moreover,
the implementation of AIED offers a potential means to overcome the physical limitations,
by making educational resources accessible through the Internet, (Chen et al., 2020; Felix,
2021; Salas-Pilco and Yang, 2022), facilitating communication (Lameras and Arnab, 2022)
and virtual collaboration and cooperation between students, as well as between students and
teachers (Zawacki-Richter, 2019), regardless of their geographical location.
Chen et al. (2020) presented various examples of AIED, such as Interactive Learning
Environments (ILEs), adaptive learning systems (ALSs), intelligent tutoring systems (ITSs),
which can facilitate administrative tasks, instructions, and educational activities. These ILEs
and ALSs are used to oversee performance, offer feedback, and facilitate interactions
between teachers and students. Additionally, ITSs such as ACTIVE Math, MATHia,
Why2Atlas, Comet, Viper, TurnItIn, Ecree, and Knewton are used to enhance the accuracy
and objectivity of task and exam grading (Celik et al., 2022) and to detect plagiarism and
ethical issues (Chen et al., 2020). Zawacki-Richter (2019) as well as Huang (2021) describe
four models that are often integrated in intelligent tutoring systems: the student model, the
teacher model, the domain model, and the diagnosis model, which evaluates errors and
defects based on the domain model.
Furthermore, chatbots, large language models (LLMs), and interactive books are used as
tools for facilitating student-machine dialogues (Huang, 2021; Lameras and Arnab, 2022;
Chiu et al., 2023; Kasneci et al., 2023) and have been shown to positively impact low-
achieving students by enhancing their confidence and sense of usefulness while reducing
feelings of embarrassment. One of the most popular LLMs is OpenAI’s GPT-4, which has
been used to generate tests and quizzes as well as an educational agent meant to foster
curiosity and critical thinking among students (Bhavani, 2020; Lameras și Arnab, 2022;
Kasneci et al., 2023).
Vol. 26 • No. 65 • February 2024 209

Nevertheless, AIED education has its shortcomings, such as limited reliability and technical
capacities, lack of adequate infrastructure in schools and universities, limited efficiency in
evaluation of text structure, content logic and coherence, a lack of understanding and interest
of teachers in the use of AI, as well as security and privay risks (Chatterjee and Bhattacharjee,
2020; Celik et al., 2022; Kasneci et al., 2023). According to Felix (2021), students may be
sceptical toward AI or at the least possess a limited comprehension of its capabilities, which
might lead to a broader sense of mistrust towards instructors or the educational system. On
the other hand, too much trust placed in inappropriate contexts may result in the student
preferring the AI above that of their educators or classmates. Many educators possess
insufficient knowledge about the operational mechanisms of these systems (Chiu et al.,
2023), expressing a sense of reduced control (Chatterjee and Bhattacharjee, 2020), leading
them to see their job as operating inside an opaque system.
AIED, like most innovative technologies used in a traditional setting, may be seen as a
disruptive force and thus, users need to adopt new behaviours to be able to efficiently use it.
Thus, Bearman et al. (2022) found that students need to acquire and enhance their critical
abilities in the use and assessment of information while Chatterjee and Bhattacharjee (2020)
found that the users’ attitude plays a key role in their intention to adopt and use AI and that
their attitude is influenced by the perceived utility of AI technologies as well as its
accessibility. This implies that, in order to implement the large-scale adoption of AIED, it is
essential for designers, developers, and system administrators in higher education institutions
to prioritise the functionality and practicality of the system. Thus, it is essential that the
authorities demonstrate a high level of commitment in correctly communicating the final
users’ needs to the developers and the choice of technology should align more consistently
with the requirements and preferences of the users. At the same time, the users should be
informed about the functionalities and capabilities of the system through product
documentation, training, and live demonstrations (Chatterjee and Bhattacharjee, 2020).
Even though the research on AIED is extensive, not many studies focus on the ethical aspects
of this phenomenon, which presents a basic concern about how the educational technology
community, including developers, designers, policy makers, and educators, can behave
ethically in order to mitigate or limit ethical drawbacks that may affect students' learning
experiences (Lameras and Arnab, 2022). Among these ethical issues we must mention
copyright and plagiarism issues with AI generated text, students and teachers relying too
much on AI, difficulty in distinguishing AI generated text, difficulty in verifying the validity
of AI generated text (Holmes et al., 2022; Kasneci et al., 2023).
2.4. Critical factors influencing the successful use of artificial intelligence applications
in digital business education
Alqahtani and Rajkhan (2020, p.4) define the critical success factors as “characteristics,
conditions, or variables that, when properly sustained, maintained, or managed, can have a
significant impact” on the students’ and teachers’ intent to adopt AI technologies in the
competency-based educational process. Our research has revealed five categories of factors
that affect the large-scale adoption of AI based digital economics education: technological
factors, quality factors, trust factors, self-efficacy factors, and cultural aspects.
Technological factors include technical skills among students and instructors (Almaiah et al.,
2020; Alqahtani and Rajkhan, 2020) and appropriate hardware and software requirements,

AE
as well as adequate internet connectivity, which are needed for the correct functioning of
AI- based online education systems. Additionally, it is crucial for universities to supply the
required technical resources to ensure consistent maintenance and upgrading.
When talking about quality factors, Almaiah et al. (2020) and Obeng and Coleman (2020)
found that many users agree that an efficient AIED application needs to be intuitive, reliable,
flexible and accessible for all types of users, especially those with less technical skills.
According to Almaiah et al. (2020), accessibility is particularly important, as there is a direct
correlation between the system’s accessibility and utility and the students’ intention to adopt
it. In this regard, AIED applications designers and developers must focus on a user-friendly,
accessible, and straightforward system, tailored to its’ users’ needs and expectations (Nguyen
et al., 2020; Obeng and Coleman, 2020).
Almaiah et al. (2020) highlight trust as one of the most important factors that influence the
adoption of AIED systems, which refers to the system’s security, privacy, and general
reliability, as well as the provision of efficient, effective, and transparent mechanisms for
engaging in e-learning activities within the online education framework (Nguyen et al.,
2020).
Almaiah et al. (2020) and Nguyen et al. (2020) argue that users need to be self-efficient to be
able to properly use an AIED application, in order to be able to set challenging goals,
persevere in the face of obstacles, and bounce back from failures. In order to develop these
capabilities, several authors (Almaiah et al., 2020; Alqahtani and Rajkhan, 2020; Nguyen et
al., 2020; Zhao, 2021) recommended that universities establish training programmes aimed
at enhancing their digital skills, properly presenting the e-learning systems’ and the AI
applications’ features as well as clearly communicating the rules and regulations that govern
them.
Several other authors have mentioned culture as a critical factor that may increase an AI
educational applications’ adoption rate. Thus, Almaiah et al. (2020) argues about the
importance of connecting with students through various social media applications. Social
media serves as an easy and effective way of engaging students (Elumanai et al., 2020), and,
at the same time, can facilitate communication between faculty members and students
(Obeng and Coleman, 2020).
2.5. Challenges facing the use of artificial intelligence applications in skills-based

economic education
Several authors highlight resistance to change as one of the biggest challenges regarding the
large-scale adoption of AIED applications in the competency-based economics education.
Humans are inherently resisting to change, not willing to leave their comfort zone, and thus,
hindering the transition from traditional to digital education (Liu and Yu, 2022). This
resistance can be attributed to trust issues regarding assignment submission and online
examinations (Almaiah et al., 2020; Mercader and Gairin, 2020), as well as lack of motivation
(Mercader and Gairin, 2020), awareness (Almazova et al., 2020) and trust regarding the
AIED applications’ features (Watermeyer et al., 2021).
Vol. 26 • No. 65 • February 2024 211

In order to mitigate these challenges, economics universities need to set up and communicate
clear policies and procedures (Nuere and de Miguel, 2021) to properly train its users on how
to efficiently use these AI based systems (Watermeyer et al., 2021). At the same time, it is
important to note that the methodological work employed by educators in a digital
educational setting differ greatly from traditional modes of instruction (Nuere and de Miguel,
2021), teachers needing to employ active collaborative strategies and to assist students in
developing their own unique learning approaches (Almazova et al., 2020) and overcome
challenges associated with electronic communication (Liu and Yu, 2022. They need to
prioritise students’ engagement (Nuere and de Miguel, 2020), support the development of
critical thinking abilities and self-reflection and online collaboration (Tomczyk and Walker,
2021). Moreover, Turnbull et al. (2021) talk about the issue of cheating and plagiarism, which
is more prevalent in AI based online learning than in traditional education.
Other challenges are related to the fact that users have different technical skills (Tomczyk
and Walker, 2021) and may have different perception regarding the AIED applications’
characteristics (Maatuk et al., 2022), aspects which forces universities to provide additional
support for those users which are facing technical difficulties (Turnbull et al., 2021;
Watermeyer et al., 2021). Almazova’s et al. (2020) and Mercader and Gairin’s (2020)
findings show that academic staff older than 55 years old are facing more difficulties in
properly using the e-learning systems than their younger colleagues. Moreover, it is
imperative for educators to acquire proficiency in online pedagogy, which relates to
designing and delivering AI-assisted educational courses (Turnbull et al., 2021), as well as
being able to properly communicate and collaborate with their peers (Nuere and de Miguel,
2021).
Another barrier that hinders the large-scale adoption of AIED applications is related to
adequate financial support. A professional custom-made AI software is not cheap (Maatuk et
al., 2022), and some universities may have trouble securing the proper funding needed to
implement it (Almazova et al., 2020). This might also lead to improper infrastructure
(Tomczyk and Walker, 2021), which might be able to withstand heavy network traffic,
leading to latency issues which disrupt the learning process (Maatuk et al., 2022). At the
same time, some students might not have the adequate hardware equipment needed for these
applications (Liu and Yu, 2022) or Internet connection (Turnbull et al., 2021; Almazova et
al., 2020).
Considering the aforementioned factors, the present study has successfully developed a
model for a system of correlations and elements in the realm of digital economic education
based on AI. This model incorporates the use of artificial intelligence and effectively
integrates key success factors and unique challenges within a cohesive framework, built on
specific domains of application (Figure no. 3).

AE
Figure no. 3. A systemic model for digital economics education based on AI
Conclusions
This study aimed to contribute to the broadening existing literature regarding to digital
economic education based on the AI platforms and applications, through a systematic review
of 60 papers published between 2019 and 2023. Thus, we tried to highlight how the concept
of digital economics education based on AI (AIED) presented in recent literature is, the
specific fields of application of AIED, and what are the factors and challenges which
influence the large-scale adoption of AIED applications. Most of the authors defined AIED
as an emerging interdisciplinary area that uses AI applications to transform education and
develop the learning process.
Our findings have shown that AI can support teachers in carrying out administrative tasks,
enhance the instruction and learning process through a task grading, exam automation,
performance evaluation and feedback, creating personalised content and facilitating
communication.
Moreover, our study has revealed five categories of factors which can influence AIED
applications: technology, quality, trust, self-efficacy, and culture. Most of the authors stress
the importance of digital literacy as a key element of artificial intelligence applications, as
well as proper training and setting up adequate policies and procedures.
The papers included in our research identified several challenges that educational AI
applications are facing, such as the users’ resistance to change, barriers related to the
accessibility, availability, usability and quality, different levels of digital literacy among its
users, and adequate financial support and funding.
Among the limitations of our study, we should mention the narrow period in which we have
focused (2020-2023), which restrained the number of papers included in the review (60), and
thus, directing them toward the use of AI educational applications after the start of the global
COVID-19 pandemic, the concept of digital economics education in general, as well as the
use of AI and other digital tools in economics education being a topic that has been addressed
both in practice and in research well before the start of the pandemic, which, however,
accelerated their adoption and development. A recommendation is addressed to researchers
Vol. 26 • No. 65 • February 2024 213

who want to study the concept of digital economics education to address the evolution of this
topic in the decades before the COVID-19 pandemic, as well. At the same time, we
recommend that the topic of digital economic education based on AI should be further
studied, considering its numerous areas of application and the fact that it is still in the
development phase and significant progress is achieving constantly.
Acknowledgements
The research within this paper was conducted within and with the support of the
Interdisciplinary Research Center for Economic and Social Sciences, INCESA (Research
Infrastructure in Applied Sciences), University of Craiova.
References
Abdulaziz Alsuhbhi, M., Sahari, N. and Tengku Wook, T.S.M., 2020. A Conceptual
Engagement Framework for Gamified E-Learning Platform Activities. International
Journal of Emerging Technologies in Learning, [e-journal] 15(22), pp. 4-23.
https://doi.org/10.3991/ijet.v15i22.15443.
Almaiah, A.M., Al-Khasawneh, A. and Althunibat, A., 2020. Exploring the critical
challenges and factors influencing the E-learning system usage during COVID-19
pandemic. Education and Information Technologies, 25, pp. 5261-5280.
Almazova, N., Krylova, E., Rubstova, A. and Odinokaya, M., 2020. Challenges and
Opportunities for Russian Higher Education amid COVID-19: Teachers’ Perspective.
Education Sciences, [e-journal] 10(12), article no. 368. https://doi.org/10.3390/
educsci10120368.
Almusharraf, N.M and Khahro, S.H., 2020. Students’ Satisfaction with Online Learning
Experiences During the COVID-19 Pandemic. International Journal of Emerging
Technologies in Learning, [e-journal] 15(21), pp. 246-267. https://doi.org/10.3991/
ijet.v15i21.15647.
Alqahtani, A.Y and Rajkhan, A.A., 2020. E-learning Critical Success Factors during the
COVID-19 Pandemic: A comprehensive Analysis of E-Learning Managerial
Perspectives. Education Sciences, [e-journal] 10(9), article no. 216.
https://doi.org/10.3390/educsci10090216.
Anthony, B., Kamaludin, A., Romli, A., Raffei, A.F.M, Eh Phon, D.N.A.L., Abdullah, A.
and Ming, G.L., 2022. Blended Learning Adoption and Implementation in Higher
Education: A Theoretical and Systematic Review. Technology, Knowledge and Learning,
[e-journal] 27, pp. 531-578. https://doi.org/10.1007/s10758-020-09477-z.
Bearman, M., Ryan, J. and Ajjawi, R., 2023. Discourses of artificial intelligence in higher
education: a critical literature review. Higher Education, [e-journal] 86, pp. 369-385.
https://doi.org/10.1007/s10734-022-00937-2.
Bhavani, G., 2020. Artificial Intelligence: Simulations in Audit Education. In: s.n., 11th
International Conference on Computing, Communication and Networking Technologies
(ICCCNT). N.d. Kharagpur, India. S.l.: s.n.

AE
Bond, M., Buntins, K., Bedenlier, S., Zawacki-Richter, O. and Kerres, M., 2020. Mapping
research in student engagement and educational technology in higher education: a
systematic evidence map. International Journal of Educational Technology in Higher
Education, [e-journal] 17, article no. 2. https://doi.org/10.1186/s41239-019-0176-8.
Bonfield, C.A., Salter, M., Longmuir, A., Benson, M. and Adachi, C., 2020. Transformation
or evolution?: Education 4.0, teaching and learning in the digital age. Higher Education
Pedagogies, [e-journal] 5(1), pp. 223-246. https://doi.org/10.1080/23752696.2020.
1816847.
Butler-Henderson, K. and Crawford, J., 2020. A systematic review of online examinations:
A pedagogical innovation for scalable authentication and integrity. Computers &
Education, [e-journal] 159, article no. 104024. https://doi.org/10.1016/j.compedu.
2020.104024.
Celik, I., Dindar, M., Muukkonen, H. and Järvelä, S., 2022. The Promises and Challenges of
Artificial Intelligence for Teachers: a Systematic Review of Research. TechTrends,
[e-journal] 66, pp. 616-630. https://doi.org/10.1007/s11528-022-00715-y.
Information Technologies, [e-journal] 25, pp. 3443-3463. https://doi.org/10.1007/
s10639-020-10159-7.
Chen, L., Chen, P. and Lin, Z., 2020. Artificial Intelligence in Education: A Review. In: IEEE
Access, [e-journal] 8, pp. 75264-75278. https://doi.org/10.1109/ACCESS.2020.2988510.
Chiu, T.K.F., Xia, Q., Zhou, X., Chai, S.C., and Cheng, M., 2023. Systematic literature
review on opportunities, challenges, and future research recommendations of artificial
intelligence in education. Computers and Education: Artificial Intelligence, [e-journal] 4,
article no. 100118. https://doi.org/10.1016/j.caeai.2022.100118.
Chiu, T.K.F., 2022. Applying the self-determination theory (SDT) to explain student
engagement in online learning during the COVID-19 pandemic. Journal of Research on
Technology in Education, [e-journal] 54(Sup1), pp. S14-S30. https://doi.org/10.1080/
15391523.2021.1891998.
Cranfield, D.J., Tick, A., Venter, I.M., Blignaut, R.J. and Renaud, K., 2021. Higher
Education Students’ Perceptions of Online Learning during COVID-19 – A Comparative
Study. Education Sciences, [e-journal] 11, article no. 403. https://doi.org/10.3390/
educsci11080403.
Dangaiso, P., Makudza, F. and Hogo, H., 2022. Modelling perceived e-learning service
quality, student satisfaction and loyalty. A higher education perspective. Cogent
Education, [e-journal] 9(1), article no. 2145805. https://doi.org/10.1080/2331186
X.2022.2145805.
Dhillon, S. and Murray, N., 2021. An Investigation of EAP Teachers’ Views and Experiences
of E-Learning Technology. Education Sciences, [e-journal] 11(2), article no. 54.
https://doi.org/10.3390/educsci11020054.
El-Sabagh, H.A, 2021. Adaptive e-learning environment based on learning styles and its
impact on development students' engagement. International Journal of Educational
Technology in Higher Education, [e-journal] 18, article no. 53. https://doi.org/10.1186/
s41239-021-00289-4.
Vol. 26 • No. 65 • February 2024 215

Elumalai, K.V., Sankar, J.P., Kalaichelvi, R., John, J.A., Menon, N., Alqahtani, M.S.N. and
Abumelha, M.A., 2020. Factors affecting the quality of e-learning during the COVID-19
pandemic from the perspective of higher education students. Journal of Information
Technology Education: Research, [e-journal] 19, pp. 731-753. https://doi.org/
10.28945/4628.
Falloon, G., 2020. From digital literacy to digital competence: the teacher digital competency
(TDC) framework. Education Technology Research and Development, [e-journal] 68,
pp. 2449-2472. https://doi.org/10.1007/s11423-020-09767-4.
Felix, C.V., 2020. The Role of the Teacher and AI in Education. In: E. Sengupta, P.
Blessinger and M.S. Makhanya eds., 2021. International Perspectives on the Role of
Technology in Humanizing Higher Education (Innovations in Higher Education
Teaching and Learning. Bingley: Emerald Publishing Limited, pp. 33-48.
https://doi.org/10.1108/S2055-364120200000033003.
Giray, G., 2021. An assessment of student satisfaction with e-learning: An empirical study
with computer and software engineering undergraduate students in Turkey under
pandemic conditions. Education and Information Technologies, [e-journal] 26,
pp. 6651-6673. https://doi.org/10.1007/s10639-021-10454-x.
Hillmayr, D., Ziernwald, L., Reinhold, F., Hofer, S.I. and Reiss, K.M, 2020. The potential of
digital tools to enhance mathematics and science learning in secondary schools:
A context-specific meta-analysis. Computers & Education, [e-journal] 153, article no.
103897. https://doi.org/10.1016/j.compedu.2020.103897.
Holmes, W., Porayska-Pomsta, K., Holstein, K., Sutherland, E., Baker, T., Buckingham
Shum, S., Santos, O.C., Rodrigo, M.T., Cukurova, M., Bittencourt, I.I. and Koedinger,
K.R., 2022. Ethics of AI in Education: Towards a Community-Wide Framework.
International Journal of Artificial Intelligence in Education, [e-journal] 32, pp. 504-526,
https://doi.org/10.1007/s40593-021-00239-1.
Huang, X., 2021. Aims for cultivating students’ key competencies based on artificial
intelligence education in China. Education and Information Technologies, [e-journal] 26,
pp. 5127-5147, https://doi.org/10.1007/s10639-021-10530-2.
Hwang, G.J. and Tu, Y.F., 2021. Roles and Research Trends of Artificial Intelligence in
Mathematics Education: A Bibliometric Mapping Analysis and Systematic Review.
Mathematics, [e-journal] 9, article no. 584. https:// doi.org/10.3390/math9060584.
Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser,
U., Groh, G., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T.,
Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T., Stadler, M., Weller,
J., Kuhn, J. and Kasneci, G., 2023. ChatGPT for good? On opportunities and challenges
of large language models for education. Learning and Individual Differences, [e-journal]
103. https://doi.org/10.1016/j.lindif.2023.102274.\.
Khalil Awan, R., Afshan, G. and Bano Memon, A., 2021. Adoption of E-Learning at Higher
Education Institutions: A Systematic Literature Review. Multidisciplinary Journal for
Education, Social and Technological Sciences, [e-journal] 8(2), pp. 74-91.
https://doi.org/10.4995/muse.2021.15813.

AE
Köenig, J., Jäger-Biela, D.J. and Glutsch, N., 2020. Adapting to online teaching during
COVID-19 school closure: teacher education and teacher competence effects among early
career teachers in Germany. European Journal of Teacher Education, [e-journal] 43(4),
pp. 608-622. https://doi.org/10.1080/02619768.2020.1809650.
Lameras, P. and Arnab, S., 2022. Power to the Teachers: An Exploratory Review on Artificial
Intelligence in Education. Information, [e-journal] 13, article no. 14. https://doi.org/
10.3390/info13010014.
Liu, M. and Yu, D., 2023. Towards intelligent E-learning systems. Educational and
s10639-022-11479-6.
Lizcano, D., Lara, J.A., White, B. and Aljawarneh, S., 2020. Blockchain-based approach to
create a model of trust in open and ubiquitous higher education. Journal of Computing in
Higher Education, [e-journal] 32, pp. 109-134. https://doi.org/10.1007/s12528-019-
09209-y.
Lloret-Irles, D., Sefura-Heras, J.V., Nogues-Pedregal, A.M., Saumell- Castelló, C and Akita-
Udrea, M., 2022. Digital skills of Valencian university students. In: Universitat
Politecnica de Valencia, 8th International Conference on Higher Education Advances
(HEAd’22). N.d., Valencia, Spain. S.l.: Universitat Politecnica de Valencia.
Maatuk, A.M., Elberkawi, E.K., Aljawarneh, S., Rashaideh, H. and Alharbi, H., 2022.The
COVID-19 pandemic and E-learning: challenges and opportunities from the perspective
of students and instructors. Journal of Computing in Higher Education, [e-journal] 34,
pp. 21-38. https://doi.org/10.1007/s12528-021-09274-2.
Maslova, I., Burdina, G., and Krapotkina, I. 2020. The Use of Electronic Educational
Resources and Innovative Educational Technologies in University Education.
International Journal of Emerging Technologies in Learning, [e-journal] 15(16),
pp. 68-79. https://doi.org/10.3991/ijet.v15i16.14909.
Mercader, C. and Gairín, J., 2020. University teachers’ perception of barriers to the use of
digital technologies: the importance of the academic discipline. International Journal of
Educational Technology in Higher Education, [e-journal] 17(4). https://doi.org/10.1186/
s41239-020-0182-x.
Mhlanga, D. and Moloi, T., 2020. COVID-19 and the Digital Transformation of Education:
What Are We Learning on 4IR in South Africa? Education Sciences, [e-journal] 10(7),
article no. 180. https://doi.org/10.3390/educsci10070180.
Micu, A., Micu, A.E., Geru, M., Căpățînă, A., Muntean, M.C., 2021. The Impact of Artificial
Intelligence Use on the E-Commerce in Romania. Amfiteatru Economic, [e-journal]
23(56), pp. 137-154. http://dx.doi.org/10.24818/EA/2021/56/137.
Muammar, S., Hashim, K.F.B. and Panthakkan, A., 2023. Evaluation of digital competence
level among educators in UAE Higher Education Institutions using Digital Competence
of Educators (DigComEdu) framework. Education and Information Technologies,
[e-journal] 28, pp. 2485-2508. https://doi.org/10.1007/s10639-022-11296-x.
Ng, D.T.K., Leung, J.K.L., Su, J., NG, R.C.W. and Chu, S.K.W., 2023. Teachers’ AI digital
competencies and twenty-first century skills in the post-pandemic world. Education
Technology Research and Development, [e-journal] 71, pp. 137-161.
https://doi.org/10.1007/s11423-023-10203-6.
Vol. 26 • No. 65 • February 2024 217

Nguyen, H.T.H., Pham, H.V., Vu, N.H. and Hoang, H.T., 2020. Factors Influencing
Students’ Intention to Use E-learning System: A Case Study Conducted in Vietnam.
International Journal of Emerging Technologies in Learning, [e-journal] 15(18),
pp. 165-182. https://doi.org/10.3991/ijet.v15i18.15441.
Nuere, S. and de Miguel, L., 2021. The Digital/Technological Connection with COVID-19:
An Unprecedented Challenge in University Teaching. Technology, Knowledge and
Learning, [e-journal] 26, pp. 931-943. https://doi.org/10.1007/s10758-020-09454-6.
O’Neill, A., 2023. Was humpty dumpty right?: Towards a functional definition of e-learning.
Education and Information Technologies, [e-journal]. https://doi.org/10.1007/s10639-
023-11900-8
Obeng, A.Y and Coleman, A., 2020. Evaluating the effects and outcome of technological
innovation on a web-based e-learning system. Cogent Education, [e-journal] 7(1).
https://doi.org/10.1080/2331186X.2020.1836729.
Ooi, K.B, Wei-Han Tan, G., Al-Emran, M., Al-Sharafi, M.A., Capatina, A., Chakraborty, A.,
Dwivedi, Y.K., Huang, T.L., Kar, A.K., Lee, V.H., Loh, X.M., Micu, A., Mikalef, P.,
Mogaji, E., Pandey, N., Raman, R., Rana, N.P., Sarker, P., Sharma, A., Teng, C.I.,
Wamba, S.F. and Wong, L.W., 2023. The Potential of Generative Artificial Intelligence
Across Disciplines: Perspectives and Future Directions. Journal of Computer Information
Systems, [e-journal]. https://doi.org/10.1080/08874417.2023.2261010.
Radianti, J., Majchrzak, T.A., Fromm, J. and Wohlgenannt, I., 2020. A systematic review of
immersive virtual reality applications for higher education: Design elements, lessons
learned, and research agenda. Computers & Education, [e-journal] 147, article no.
103778. https://doi.org/10.1016/j.compedu.2019.103778.
Rajabalee, Y.B. and Santally, M.I., 2021. Learner satisfaction, engagement and performances
in an online module: Implications for institutional e-learning policy. Education and
s10639-020-10375-1.
Rakic, S., Tasic, N., Marjanovic, U., Softic, S., Lüftenegger, E., and Turcin, I., 2020. Student
Performance on an E-Learning Platform: Mixed Method Approach. International Journal
of Emerging Technologies in Learning, [e-journal] 15(02), pp. 187-203.
Salas-Pilco, S.Z. and Yang, Y., 2022. Artificial intelligence applications in Latin American
higher education: a systematic review. International Journal of Educational Technology
in Higher Education, [e-journal] 19(21). https://doi.org/10.1186/s41239-022-00326-w.
Sanusi, I.T., Olaleye, S.A Agbo, F.J. and Chiu, T.K.F., 2022. The role of learners’
competencies in artificial intelligence education. Computers and Education: Artificial
Intelligence, [e-journal] 3, article no. 100098. https://doi.org/10.1016/
j.caeai.2022.100098.
Shams, M.S., Niazi, M.M., Gul, H., Mei, T.S. and Khan, K.U., 2022. E-Learning Adoption
in Higher Education Institutions During the COVID-19 Pandemic: A Multigroup
Analysis. Frontiers in Education, [e-journal] 6, article no. 783087.
Https://doi.org/10.3389/ feduc.2021.783087.

AE
Southworth, J., Migliaccio, K., Glover, J., Glover, J.’N., Reed, D., McCarty, C., Brendemuhl,
J. and Thomas, A., 2023. Developing a model for AI Across the curriculum:
Transforming the higher education landscape via innovation in AI literacy, Computers
and Education: Artificial Intelligence, [e-journal] 4, article no. 100127.
Theresiawati, T., Seta, H.B., Hidayanto, A.N. and Abidin, Z., 2020. Variables affecting
e-learning services quality in Indonesian higher education: Students’ perspectives.
Journal of Information Technology Education: Research, [e-journal] 19, pp. 259-286.
https://doi.org/10.28945/4489.
Tomczyk, Ł. and Walker, C., 2021. The emergency (crisis) e-learning as a challenge for
teachers in Poland. Education and Information Technology, [e-journal] 26, pp. 6847-
6877. https://doi.org/10.1007/s10639-021-10539-7.
Turnbull, D., Chugh, R. and Luck, J., 2021. Transitioning to E-Learning during the COVID-
19 pandemic: How have Higher Education Institutions responded to the challenge?.
Education and Information Technology, [e-journal] 26, pp. 6401-6419.
https://doi.org/10.1007/s10639-021-10633-w.
Watermeyer, R., Crick, T., Knight, C. and Goodall, J., 2021. COVID-19 and digital
disruption in UK universities: afflictions and affordances of emergency online migration.
Higher Education, [e-journal] 81, pp. 623-641. https://doi.org/10.1007/s10734-020-
00561-y.
Xu, W. and Ouyang, F., 2022. A systematic review of AI role in the educational system based
on a proposed conceptual framework. Education and Information Technology, [e-journal]
27, pp. 4195-4223. https://doi.org/10.1007/s10639-021-10774-y.
Zawacki-Richter, O., Marín, V.I., Bond, M. and Gouverneur, F., 2019. Systematic review of
research on artificial intelligence applications in higher education – where are the
educators?. International Journal of Educational Technology in Higher Education,
[e-journal] 16(39). https://doi.org/10.1186/s41239-019-0171-0.
Zhao, Y., Pinto Llorente, A.M. and Sánchez Gómez, M.C., 2021. Digital competence in
higher education research: A systematic literature review. Computers & Education,
[e-journal] 168, article no. 104212. https://doi.org/10.1016/j.compedu.2021.104212.
Zhou, L., 2021. Effect Evaluation and Influencing Factors of E-Learning Training in
Colleges. International Journal of Emerging Technologies in Learning, [e-journal]
16(22), pp. 73-86. https://doi.org/10.3991/ijet.v16i22.26877.
Vol. 26 • No. 65 • February 2024 219

AE Artificial Intelligence and Competency-Based Education: A Bibliometric Analysis
ARTIFICIAL INTELLIGENCE AND COMPETENCY-BASED EDUCATION:

A BIBLIOMETRIC ANALYSIS
Cătălina Radu1 , Carmen Nadia Ciocoiu2 , Cristina Veith3
and Răzvan Cătălin Dobrea4
1)2)4)
3)
University of Bucharest, Bucharest, Romania.

Radu, C., Ciocoiu, C.N., Veith, C. and Dobrea, R.C., Received: 20 September 2023
2024. Artificial Intelligence and Competency-Based Revised: 18 November 2023
Education: A Bibliometric Analysis. Amfiteatru Accepted: 15 Decemeber 2023
Economic, 26(65), pp. 220-240.
Abstract
In the context of the educational transition toward a competency-based approach, this study
aimed to identify trends, challenges, and emerging opportunities generated by the
intersection of Artificial Intelligence (AI) and Competency-Based Education (CBE). The
research was carried out using a bibliometric analysis of 1,028 articles included in the Web
of Science database and based on reports provided by the biblioshiny application, the
graphical interface of the bibliometrix R package. The results included a quantitative
analysis of scientific production, collaborations, and cocitations, as well as the evolution
and thematic map of the field. These revealed an annual increase of 8.43% in publications
with acceleration after 2017 and global involvement, with the United States and China in
leading positions. Thematic analyses have shown the field's evolution from technological
foundations to an interdisciplinary approach, highlighting the influences of global events,
such as COVID-19. The research confirmed the profound interaction between AI and CBE,
demonstrating its potential, complexity, and the need for collaborative and interdisciplinary
approaches. The bibliometric analysis performed can serve as a guide for future research
directions and for identifying strategic directions in the implementation of AI in education.
Keywords: artificial intelligence (AI), competency-based education (CBE), bibliometric

analysis, thematic map, Web of Science (WoS).
JEL Classification: I20, I21

Corresponding author, Cătălina Radu – e-mail: [email protected]

AE
Introduction
In recent decades, education has undergone significant transformations, moving from a
traditional model centred on memorisation and standardised assessments to a more flexible
and personalised educational system. Competency-Based Education (CBE) is a
contemporary approach that emphasises the development of practical skills, pertinent
knowledge, and essential cognitive abilities vital for success in our current world
(Johnstone and Soares, 2014). A key shift in this transition has been the integration of
artificial intelligence (AI) into the educational process by providing advanced solutions and
technologies that improve the efficiency, effectiveness, and relevance of the learning
experience. AI refers to the capability of a machine to make decisions in a manner akin to
human intelligence (Winkler-Schwartz et al., 2019). Indeed, artificial intelligence has
emerged as a rapidly expanding field of research and application, leading to a true
revolution in various sectors of modern society (Fosso Wamba et al., 2021). Within the
educational realm, AI has become an instrumental tool with extraordinary potential to
improve teaching and learning processes, transforming traditional methods into innovative
and personalised approaches (Mhlanga, 2021).
This paper offers a bibliometric analysis that aims to investigate and highlight the benefits and
challenges posed by AI applications in CBE. The motivation for this bibliometric scrutiny
stems from a need to comprehend the evolution of research in the field, key themes and pivotal
moments, identification of leading authors and primary contributions, research sources,
collaborations, emerging research directions, and potential gaps or understudied areas.
Recently, various bibliometric analysis studies have been conducted on the relationship
between artificial intelligence and education. However, these studies either have a general
focus on the AI use in an educational setting (Baek and Doleck, 2020; Talan, 2021; Prahani
et al., 2022) or concentrate on applying AI for specific educational levels or types, such as
higher education (Hinojo-Lucena et al., 2019), e-learning (Jia et al., 2022), mathematics
teaching-learning process (Hwang et al., 2020), or leadership education (Harto et al., 2022).
A detailed bibliometric analysis, focused exclusively on AI's use within the CBE context, is
still absent. In this context, the present study aims to answer the following research
question: How has the specialised literature evolved, and what are the main research trends
regarding the intersection of AI and CBE? The results of the analysis can contribute to
developing both theoretical and practical domains. In this way, the authors can provide
clarity on the current research landscape, identifying gaps and the predominant directions
of investigation. From a theoretical perspective, the aim of bibliometric analysis is to
highlight areas where researchers have primarily focused their efforts and to present the
dynamic development of the theoretical landscape. However, from a practical applied point
of view, the study aims to contribute to the advancement of knowledge and the
improvement of educational processes, directly impacting the development of key
competencies in students.
To explore the crucial role that AI plays within the context of competency-based education,
this paper is structured into four distinct sections. The first part presents a brief yet
comprehensive review of the literature on CBE and AI applications in education,
identifying the main trends and challenges in the field, as well as how AI can facilitate the
development of key abilities and competencies in education. The second part delves into
the research methodology and the results of the bibliometric analysis, a method allowing us
to trace the research evolution, and pinpoint significant influences and connections between
Vol. 26 • No. 65 • February 2024 221

various topics and concepts related to AI and CBE. The third section is dedicated to
discussions, where a detailed analysis of research findings and their implications in the
context of previous studies and existing theories is conducted, also addressing the
limitations of the study and prospective research directions. In the final part, conclusions
are presented, emphasising the primary findings of the research and highlighting the
contributions to the domain of education and artificial intelligence, alongside its theoretical
and practical implications.

1.1. Competency-Based Education: A Holistic and Balanced Approach to Learning
Competency-Based Education (CBE) has emerged as a major topic of interest in
educational research and pedagogical practice in recent times (Paek, Um, and Kim, 2021).
A competent graduate is one who can apply what they have learnt, thereby operating
independently in various real-world practical contexts (Yip and Smales, 2000). At present,
CBE is still in the early stages of implementation in certain disciplines, while in others,
such as medicine, it already has an extensive history and is supported by numerous studies
validating its efficacy (Brammer and Goodrich, 2021).
CBE emphasises the development and evaluation of specific competencies rather than
focussing solely on theoretical knowledge. The primary goal of CBE is to prepare learners
to successfully tackle real-world challenges by focussing on developing skills and aptitudes
applicable in various contexts (Johnstone and Soares, 2014). This educational approach is
tailored and applicable (Brammer and Goodrich, 2021), fostering the growth of practical
skills and competencies essential for individuals to achieve their personal and professional
objectives.
Unlike traditional education, CBE does not adhere to a standard time framework for class
attendance and homework. As a result, students often complete courses or even entire
programmes in shorter time frames, which can reduce costs and make education more
accessible (Lindsay, 2018; Mehall, 2019; Brammer and Goodrich, 2021). Learning takes
place in modules and, given the student-centred approach, learners progress to the next
topic or skill once they have sufficiently mastered the associated competencies (Evans,
Landl and Thompson, 2020; Brammer and Goodrich, 2021). Research indicates that CBE
can increase motivation and engagement in the learning process (Evans, Landl, and
Thompson, 2020), and foster critical and analytical thinking skills (Chen, Zhang, and Li,
2022). These benefits suggest that the implementation of a competency-based approach in
education can contribute significantly to improving the learning experience and the holistic
development of students.
The CBE has played a pivotal role in the reform of global education systems. Reform
advocates, policymakers, educators, and experts have defined “key competencies” for
compulsory education in various ways (Anderson-Levitt and Gardinier, 2021). Numerous
countries have adopted or piloted competency-based learning in their curricula, attempting
to align education with the demands of the labour market and the requisites of modern
society. The implementation of CBE is not without challenges; such an approach can be
complex and requires significant resources for course design and the technology used
(Burnette, 2016). Competency evaluations can occasionally be subjective and challenging

AE
to administer objectively. For instance, Gielissen et al. (2022) highlight the issue of the
proficiency attained by each learner (typically measured) and the pace of individual
progress in learning, both being predictors of future success. A further challenge arises
from the time-based structures of tertiary institutions (courses, semesters), and abandoning
these is a hard-to-implement innovation (Gruppen et al., 2016). Lastly, it is necessary to
emphasise that identifying relevant and necessary competencies in a rapidly changing
society is a challenge in itself.
1.2. The Role of Artificial Intelligence in Competency-Based Education
Artificial intelligence (AI) is defined as a system's capability to correctly interpret external
data, learn from these data, and use this knowledge to accomplish specific tasks and
objectives through flexible adaptation (Kaplan and Haenlein, 2019). Consequently, these
systems can think and act human-like, employing advanced algorithms and intricate
mathematical models to achieve specific aims (Akgun and Greenhow, 2022). For example,
ChatGPT represents an artificial intelligence model, specifically, a language generation
model developed by OpenAI, employing machine learning techniques to produce text that
closely resembles human language. The advent of ChatGPT marks a transformative shift in
the landscape of education (Kasneci et al., 2023; Peters et al., 2023).
One of AI's remarkable facets in education is its ability to customise individual learning
experiences (Chen, Chen, and Lin, 2020). By analysing students' behaviour and progress,
AI systems can adjust the curriculum and teaching methodologies to align with individual
learning paces and needs (Kuleto et al., 2021; Mhlanga, 2021). This leads to more effective
and efficient learning, and students feel more motivated and engaged. Furthermore, Huang
(2021) examined how artificial intelligence curriculum design is correlated with the
cultivation of key competencies in students. This is particularly important, since academia
should carefully assess its capabilities and make a concerted effort to provide the latest
knowledge and skills to empower students to face the challenges of future realities (Ellahi,
Khan and Shah, 2019).
Although sometimes criticised for being less empathetic, AI can recognise emotions, such
as a student's frustration (Poria et al., 2019). In CBE, this capability can be pivotal, with
recognition unlocking student potential. Adaptive learning systems generate precise student
profiles and models based on their emotional states, knowledge levels, individual
personality traits, and held competencies (Almohammadi et al., 2017). In addition, AI
offers real-time assistance and continuous feedback. Virtual tutors and AI assistants can
answer student questions, provide supplementary explanations, and give detailed feedback
on academic progress (Hwang et al., 2020). This ensures that students have access to
personalised support anytime, helping to consolidate knowledge and develop skills.
In addition to student benefits, AI also significantly impacts educators and academic
institutions. Teachers can perform various administrative functions, such as grading student
assignments through an automated process (Chen, Chen, and Lin, 2020), allowing them to
focus more on direct educational activities, human interaction, innovative content creation,
and fostering a collaborative learning environment (Akgun and Greenhow, 2022), while
evaluations are conducted more efficiently. Additionally, AI assists researchers in analysing
large data volumes, discerning trends and patterns, thus facilitating the development of new
theories and addressing intricate challenges in various study areas (Kuleto et al., 2021).
Vol. 26 • No. 65 • February 2024 223

All these aspects emphasise the significant potential of artificial intelligence to transform
education and research.
Beyond all these advantages, there are a series of risks and challenges associated with AI-
based education. Although AI systems may be more cost effective in the long run than
well-paid human resources, their selection by university managers is not necessarily a good
choice, as it compromises the personalisation of education (Kaplan and Haenlein, 2019).
Ethical concerns, such as data protection and ensuring a secure and fair environment for all
students, are ongoing debates (Nguyen et al., 2022). Using AI in education typically
involves collecting and analysing extensive personal data, ranging from facial recognition
(Borenstein and Howard, 2021) to monitoring systems with detailed information about
students' and educators' actions and preferences (Akgun and Greenhow, 2022). AI, being
human-created, can inherently possess biases that may disadvantage certain students (Baker
and Hawn, 2022). Other potential issues include an overreliance on technology and the
possible replacement of teachers with robots (Kanungo et al., 2022). While fostering
growth in labour productivity, AI technologies contribute to partial job displacement
(Lazaroiu and Rogalska, 2023). Lastly, even though AI boasts accessibility, not all students
have access to AI technology (Holmes et al., 2022), potentially widening the gap between
the haves and have-nots. However, with ethical and responsible management, AI has the
potential to bring about significant enhancements in the educational process and student
learning experiences around the world.
The bibliometric study aimed to identify and analyse the evolution of the specialised
literature addressing the intersection between artificial intelligence (AI) and education, with
a particular focus on competencies, skills, and efficiency. Specifically, we aim to achieve
the following objectives:
 Identify the main research trends in the literature addressing the relationship between
AI and CBE in terms of the most influential authors, publications, institutions involved, and
collaboration networks;
 Analyse the evolution of the specialised literature on AI and CBE in terms of both the
number of publications per year and the topics addressed.
The chosen research method is bibliometric analysis because it allows the synthesis of a
large amount of information and the identification of essential research characteristics (Aria
and Cuccurullo, 2017).
The research process followed the steps below according to the PRISMA approach:
Data Collection: A systematic search was conducted in the Web of Science (WoS)
database by Topic, using the following keywords: (“artificial intelligence”) AND
(“education*”) AND (“competenc*” OR “skill*” OR “proficienc*” OR “capabilit*” OR
“abilit*”). The selection of keywords was based on a careful analysis of articles addressing
aspects of AI and CBE, which necessitated the use of synonyms for the term
“competences” alongside the primary keyword. Other words, such as “robot”, “machine
learning”, “chatGPT”, etc., were not included because testing showed that they either led to

AE
finding works covered by the term “artificial intelligence” or to papers deviating from the
research subject.
The search was carried out in June 2023 and produced n=1,853 results, to which a series of
exclusion and inclusion criteria were applied. The exclusion criterion was the year 2023 to
obtain a complete and comparable selection across years, resulting in a total of n=1,606
articles. The inclusion criteria applied were document type (article or review article or early
access) (n=1076) and English language (n=1,028). For the year 2023, although excluded
from the analysis, 40 works were retained, represented by those included in the database
since 2022 (early access articles). The database comprising 1,028 articles was reviewed and
corrected for import errors or missing records. Furthermore, all articles were analysed to
confirm their eligibility with respect to the treated subject.
Data Preprocessing: After the data was collected, the search results were imported into the
bibliometrix R-package. This package allows for comprehensive bibliometric analyses
compared to other major software tools (Aria and Cuccurullo, 2017).
Data Analysis: Biblioshiny, a graphical interface for bibliometrix, was used to analyse the
data extracted from WoS.
Interpretation and Data Visualisation: The analysis results were interpreted and
visualised using various tables, graphs, and diagrams generated by biblioshiny. These
helped the study identify trends and synthesise information in an easy-to-understand
manner.
Before starting the research, various research databases were evaluated to determine the
most appropriate data source for the study. Ultimately, exclusive use of the Web of Science
(WoS) database was chosen for the following reasons (Azañedo et al., 2022): its broad
coverage of research literature from various fields of study, including science, technology,
social sciences, and humanities; its high-quality standards ensured the credibility and solid
foundation of the analysed studies; advanced search and filtering features, which allowed
precise topic-based searches and the application of specific filters, such as document type,
language, and year range; compatibility with biblioshiny, the tool utilised for bibliometric
analysis.

Table 1 centrally presents the records retained for bibliometric analysis. The collected data
covers a 32-year period (1991-2022), with 1.028 documents found in 573 different sources.
The annual growth rate of publications is 8.43%, indicating a continuous increase in
interest in the intersection of AI and education. The documents were written by 3.693
different authors, with 192 of them contributing to single-authored docs. This shows a large
diversity of contributions and may indicate an interdisciplinary nature of the field. Of the
collected documents, 195 are single-authored documents, and, on average, each document
has 3.88 coauthors. 23.54% of the documents have international coauthorship, suggesting
significant global collaboration in this research field. Most of the documents are articles
(848), followed by reviews (103) and early access articles (43). There are also a smaller
number of articles classified as book chapters (18 as part of an article and 1 as part of a
review), and articles classified as conference papers (10).
Vol. 26 • No. 65 • February 2024 225

Table no. 1. Centralisation of articles extracted from WoS
Description Results Description Results

Main information about data Authors' Collaboration
Timespan 1991:2022 Single-authored docs 195
Sources (Journals, Books, etc.) 573 Co-Authors per Doc 3,88
Documents 1028 International co-authorships % 23,54
Annual Growth Rate % 8,43 Document types
Average citations per document 10,18 article 848
References 45048 article; book chapter 18
Document contents article; early access 43
Keywords Plus 1375 article; proceedings paper 10
Author's Keywords 2966 review 103
Authors review; book chapter 1
Authors 3693 review; early access 5
Authors of single-authored docs 192
Source: authors, based on data extracted from WoS.
The timeline of publications (Figure 1) shows a clear trend of increasing interest in the
intersection of AI and education, as reflected in the growing number of articles published
over time. In the early period, from the early 1990s to 2016, the number of publications
remained relatively low, with less than 10 articles per year. However, starting in 2017, we
see a rapid and consistent increase in the number of publications per year. This indicates the
growing recognition of the importance and relevance of the intersection between AI and
education in the academic community.
Figure no. 1. Evolution of the number of published articles

Note: The articles for 2022 also include early access ones.
Source: The authors, based on data extracted from WoS
In 2017, the number of published articles increased to 24, almost triple from the previous
year. This upward trend continued, culminating in a surge in research in the last two years
of the analysed period. This rapid growth can be attributed to several factors, including the
increasing importance of AI in education, advancements in AI technologies, increased
research, development, and investments in this field, and a growing need and demand for
AI-related skills and abilities in the educational context. The main research trends indicate

AE
an exponential growth in interest in the intersection between AI and education, with a

notable acceleration of research activity in recent years.
Top Scientific Production by Countries
The United States leads the ranking with a total of 570 publications, indicating a leadership
position in this field (Table 2). This can be attributed to a combination of factors, including
advanced technological development, a strong culture of innovation, and significant
investments in research and development in the field of artificial intelligence and education.
China, with 497 publications, comes very close to the United States, ranking second. This
reflects China's rapid rise in the fields of technology and education, backed by significant
R&D investments and a national strategy focused on promoting artificial intelligence in
education. The United Kingdom and Canada also rank high, with 223 and 164 publications,
respectively. These are countries with a strong tradition in the educational field and
technological innovation, reflected in the high volume of research produced.
Germany, Spain, India, Australia, Italy, France, Saudi Arabia, and Turkey round out the top
10, with publication numbers ranging from 107 in Germany to 50 in Turkey. The presence
of these countries highlights the wide geographic distribution of research in this field, with
significant contributions from Europe, Asia, North America, Australia, and the Middle
East. A wide variety of countries contribute to research in the field of artificial intelligence
and education, with the United States and China leading the way. This indicates the global
relevance of the topic and the broad involvement of the international academic community.
Table no. 2. Top production by country
Region Frequency Region Frequency

USA 570 India 92
China 497 Australia 85
UK 223 Italy 82
Canada 164 France 72
Germany 107 Saudi Arabia 64
Spain 101 Turkey 50
Source: The authors, using biblioshiny, based on data extracted from WoS
The Top Most Relevant Affiliations

The ranking of affiliations highlights the academic institutions that lead research in this
field, providing a view of the diversity and global distribution of these research efforts
(Table 3). The University of Toronto ranks first with 29 articles. It is one of Canada's most
prestigious universities, renowned for its strong programmes in technology and education
fields. Additionally, it is distinguished by significant investments in research and
innovation. McGill University, which ranks second with 25 articles, is also one of the top
Canadian institutions. This underlines the importance Canada attaches to the field of AI and
education. Stanford University, ranked third with 16 articles, is one of the world's most
recognised universities in technology and innovation. Its location in Silicon Valley provides
a close link to the tech industry. King Abdulaziz University in Saudi Arabia, with
14 articles, and the Chinese University of Hong Kong, with 13 articles, reflect the growing
importance given by non-Western countries to the research in this field.
Vol. 26 • No. 65 • February 2024 227

Medical schools, such as Harvard Medical School, with 11 articles, also highlight the
impact of AI and education in the medical field and in the professional training of
physicians. Other leading universities include the University of Valencia, Carnegie Mellon
University, Queen's University, and the University Health Network, each with 10 articles,
indicating a varied and global interest in this research field.
Table no. 3. Most relevant affiliations
Affiliation Articles
University of Toronto 29
Mcgill University 25
Stanford University 16
King Abdulaziz University 14
Chinese University of Hong Kong 13
Harvard Medical School 11
University of Valencia 11
Carnegie Mellon University 10
Queen’s University 10
University Health Network 10
The top of the most relevant sources
This ranking (Table 4) provides a view of the most relevant and active sources of literature
in this research field, ranked by the number of published articles. “Mobile Information
Systems”, with 27 articles, and “Frontiers in Psychology” and “Sustainability”, each with
26 articles, are in the top three positions. These results were predictable, considering the
large number of articles these journals publish annually, many in the thousands for the last
two. Other significant journals include “Wireless Communications & Mobile Computing”,
“Computational Intelligence and Neuroscience”, and “Education and Information
Technologies”, each with a significant number of articles on this topic. These journals show
that there are a variety of research fields that intersect with AI and education.
Table no. 4. Most relevant sources
No. Sources Articles
1 Mobile Information Systems 27
2 Frontiers in Psychology 26
3 Sustainability 26
4 Wireless Communications & Mobile Computing 16
5 Computational Intelligence and Neuroscience 13
6 Education and Information Technologies 13
7 Applied Sciences 12
8 IEEE Access 12
9 International Journal of Artificial Intelligence in Education 11
10 International Journal of Emerging Technologies in Learning 11
11 Education Sciences 10
12 Frontiers in Education 10
13 Journal of Intelligent & Fuzzy Systems 10
14 Interactive Learning Environments 9
15 BMC Medical Education 7
16 Computers & Education 7

AE
No. Sources Articles

17 Engineering 7
18 Mathematical Problems in Engineering 7
19 Scientific Programming 7
20 Sensors 7
The authors' impact is determined based on the number of citations, as well as the Hirsch
index (h-index), and its more recent variants, the G index and the M index (Table 4).
Dautenhahn, K., Kaplan, A., and Haenlein, M. lead the list, having the highest number of
citations. However, the highest Hirsch (h-index), G, and M indices belong to other authors,
namely Del Maestro R.F. (Emeritus Professor Emeritus in Neuro-Oncology at McGill
University), Winkler-Schwartz A. and Yilmaz R., both from the Neurosurgical Simulation
and Artificial Intelligence Learning Centre at McGill University. The high values of these
indices indicate that these authors have a significant number of works in the database that
have been cited by other researchers. On the other hand, in the case of Dautenhahn, there
are two works that meet the search criteria, a large number of citations, but low indices.
This is due to the fact that citation reporting is strictly done in relation to the number of
works included in the database processed with Bibliometrix; as a result, the value of the
indices cannot exceed that of the number of works. Additionally, Table 5 provides other
relevant information about the authors, such as the start year (PY_start) when each author
began publishing in this field.
Table no. 5. Authors’ impact (descending order by the total number of citations)
Total Number of Starting
Autor name h_index g_index m_index citatios publications year
(TC) (NP) (PY_start)
Dautenhahn K. 2 2 0.118 652 2 2007
Kaplan A. 2 2 0.4 644 2 2019
Haenlein M. 2 2 0.4 600 2 2019
Sheridan T.B. 1 1 0.125 245 1 2016
Del Maestro R.F. 5 7 1 193 7 2019
Winkler-Schwartz A. 3 6 0.6 169 6 2019
Yilmaz R. 3 6 0.6 164 6 2019
Ledwos N. 3 5 0.6 163 5 2019
Mirchi N. 3 5 0.6 163 5 2019
Kellnhofer P. 1 1 0.333 157 1 2021
Kim C. 1 1 0.333 157 1 2021
Li B.C. 1 1 0.333 157 1 2021
Matusik W. 1 1 0.333 157 1 2021
Shi L. 1 1 0.333 157 1 2021
Bissonnette V. 3 4 0.6 150 4 2019
The most cited documents
Documents are ranked based on the number of local citations, respectively, based on the
WoS Core Collection and global ones, across all WoS databases (Table 6). The ratio
between local and global citations (LC/GC Ratio) indicates the document's relevance to
researchers in the specific field. Documents with a high LC/GC ratio may indicate that the
Vol. 26 • No. 65 • February 2024 229

research is highly relevant to the specialised community. However, there are some
exceptions to this trend. For instance, Kaplan A.'s 2019 article published in Business
Horizons has 8 local citations but 574 global citations, indicating a broad impact within the
academic community but not necessarily within the specific research community relating to
the intersection between AI and education.
Table no. 6. Top of the most cited articles
Local Global
LC/G GC
Article Year Citations Citations LC (n)
C (%) (n)
(LC) (GC)
Chen et al., 2020. Artificial Intelligence
in Education: A Review. IEEE Access, 8,
pp.75264-75278. 2020 16 136 11.76 21.56 8.26
Sit et al., 2020. Attitudes and perceptions
of UK medical students towards artificial
intelligence and radiology: a multicentre
survey. Insights into imaging, 11, pp.1-6. 2020 12 88 13.64 16.17 5.34
Winkler-Schwartz et al., 2019. Artificial
Intelligence in Medical Education: Best
Practices Using Machine Learning to Assess
Surgical Expertise in Virtual Reality
Simulation. Journal of Surgical Education,
76(6), pp.1681-1690. 2019 11 69 15.94 11.60 2.52
Mirchi et al., 2020. The Virtual Operative
Assistant: An explainable artificial
intelligence tool for simulation-based
training in surgery and medicine. PLOS
ONE, 15(2), e0229596. 2020 11 75 14.67 14.82 4.55
Sapci et al., 2020. Artificial intelligence
education and tools for medical and health
informatics students: systematic review.
JMIR Medical Education, 6(1), e19285. 2020 11 38 28.95 14.82 2.31
Kaplan, A. & Haenlein, M., 2019. Siri, Siri,
in my hand: Who is the fairest in the land?
On the interpretations, illustrations, and
implications of artificial intelligence.
Business Horizons, [online] 62(1), pp.15-25. 2019 8 574 1.39 8.44 20.97
Huang, X., 2021. Aims for cultivating
students’ key competencies based on
artificial intelligence education in China.
Education and Information Technologies,
26, pp.5127-5147. 2021 7 24 29.17 18.17 2.83
Note: LC (n) = normalised local citations; GC (n) = normalised global citations
The most cited references in the WoS Core Collection

Table 7 details the references cited and contains key works on the intersection of AI and
CBE. As can be seen, a significant part of the research focusses on how digital
technologies, including AI, can enhance educational practices. There is also a growing
interest in using AI in medical education and in the training of health professionals.

AE
The choice of specialised journals such as Nature, MIS Quarterly, and IEEE Access
indicates an interdisciplinary approach to research that involves a variety of approaches and
subdomains. In addition, there is a social concern about the impact of technology on
education and work.
Table no. 7. Most local cited references
No. Cited References Citations
1. Zawacki-Richter et al., 2019. Systematic review of research on artificial
intelligence applications in higher education – where are the educators?
International Journal of Educational Technology in Higher Education, 16(1),
pp.1-27. 35
2. Popenici, S.A.D. and Kerr, S., 2017. Exploring the impact of artificial
intelligence on teaching and learning in higher education. Research and Practice
in Technology Enhanced Learning, 12(1), pp.1-13. 28
3. Topol, E. J. , 2019. High-performance medicine: the convergence of human and
artificial intelligence. Nature Medicine, 25(1), pp.44-56. 26
4. Frey, C. B., & Osborne, M. A., 2017. Technological Forecasting and Social
Change, 114, pp.254-280. 25
5. Pinto dos Santos et al. 2019. Medical students’ attitude towards artificial
intelligence: a multicentre survey. European Radiology, 29, pp.1640-1646. 23
6. Wartman, S. A., & Combs, C. D., 2018. Medical Education Must Move From the
Information Age to the Age of Artificial Intelligence. Academic Medicine, 93(8),
pp.1107-1109. 21
7. Luckin, R. & Holmes, W., 2016. Intelligence unleashed: An argument for AI in
education. London: Pearson 20
8. Braun, V. & Clarke, V., 2006. Using thematic analysis in psychology.
Qualitative Research in Psychology, 3(2), pp.77-101. 18
9. Roll, I., & Wylie, R., 2016. Evolution and Revolution in Artificial Intelligence in
Education. International Journal of Artificial Intelligence in Education, 26,
pp.582-599. 18
10. Schwab, K., 2017. The fourth industrial revolution. Currency 17
11. Chen et al., 2020. Artificial Intelligence in Education: A Review. IEEE Access, 8,
pp.75264-75278. 16
12. Esteva et al., 2017. Dermatologist-level classification of skin cancer with deep
neural networks. Nature, 542(7639), pp.115-118. 16
13. Timms, M. J., 2016. Letting Artificial Intelligence in Education Out of the Box:
Educational Cobots and Smart Classrooms. International Journal of Artificial
Intelligence in Education, 26, pp.701-712. 16
Most frequent author keywords
The top keywords are dominated by terms used in the search, as well as others suggesting a
focus on specific AI technologies and how they can be applied in education (table 8).
Table no. 8. Most frequent words
Words Occurrences Words Occurrences

artificial intelligence 437 intelligent tutoring systems 15
education 99 internet of things 15
machine learning 92 curriculum 14
e-learning 48 information technology 14
Vol. 26 • No. 65 • February 2024 231

Words Occurrences Words Occurrences

deep learning 38 natural language processing 14
higher education 30 augmented reality 13
technology 26 learning 13
big data 25 artificial intelligence in education 12
virtual reality 25 computational thinking 11
covid 22 digital health 11
medical education 21 engineering education 11
robotics 21 neural networks 11
assessment 18 simulation 11
automation 16 skills 11
chatbot 16 training 11
industry 4 0 16 competency 10
Cocitation network of sources
The journal co-citation network consists of two clusters coloured red and blue (Figure 2).
Clustering is done based on betweenness centrality, which measures how often a source
acts as a bridge on the shortest path between two other sources, and closeness centrality,
which measures the average distance of a source to all other sources in the network (Ulrik,
2001). The “Science” source has the highest intermediation score, indicating a central role
in connections between other research sources, while the source with the highest closeness
score is “Acad Med”, suggesting that it is closely tied to many other research sources. The
first cluster (red) includes many medical journals like “Acad Med”, “J Med Internet Res”,
“Science”, “Nature”, and “Plos One”, suggesting a lot of research at the intersection of
medicine, education, and AI. The second cluster (blue) includes journals dedicated to
education technology and AI, such as ”Comput Educ”, “Comput Hum Behav”, “Ieee
Access”, and “Lect Notes Comput Sc”. This indicates another significant research direction
that focusses on the broad application of AI in education.
Figure no. 2. Co-citation network of sources


AE
Collaboration network of authors

The author collaborations are divided into 10 clusters, indicating the existence of multiple
research groups working together (Figure 3). In addition, it provides clues about key
authors who might have significant influence in this field, either by promoting
collaboration among other researchers or through their own academic authority.
Figure no. 3. Collaboration network of authors

Collaboration network of countries

The country collaboration network is divided into 4 clusters, representing regional research
groups (Figure 4). For example, countries like the USA, China, the United Kingdom, and
Canada belong to cluster 1 (red), most European countries are in cluster 2 (blue), a few
Asian countries are in cluster 3 (green), and Ecuador is the sole member of cluster 4
(purple).
Figure no. 4. Collaboration network of countries

Vol. 26 • No. 65 • February 2024 233

Thematic evolution
By comparing the two periods (figure 5), the interest areas are clearly evolving. While the
first period focused on fundamental technological developments (AI, machine learning, and
robotics), in the second period, we see a blend of technology (AI and E-learning) with
major global events (COVID). This reflects the adaptability and interconnectedness of
research with global events and developments.
Figure no. 5. Thematic evolution during two periods of time

Thematic Map
The thematic map is divided into four areas and contains eight clusters, positioned
according to density and centrality. Each of these clusters consists of several keywords
(Figure 7). It is important to mention that the type of algorithm used for grouping/clustering
influences the number and grouping method of clusters in the four zones of the thematic
map. In this case, the Fast Greedy algorithm was used. The results show that in the
quadrant with the “main themes”, there are two clusters with well-structured themes
important for the research field, namely “e-learning, training, ethics” and “higher
education, automation, chatbot”. The latter focusses on how technology and AI are
integrated into higher education and their impact on the job market and student training.
Furthermore, a growing interest in adaptive and personalised learning was also shown by
the frequent appearance of terms like “intelligent tutoring systems” and “adaptive
educational system”.
The “basic themes” are the most important for the research field, but are still
underdeveloped. Here, two clusters are included; the largest one (consisting of 31 words) is
dominated by terms such as “artificial intelligence, education, machine learning”, while the
second one, “technology, covid, medical education”, is dedicated to the application of
technology in medical education. Themes associated with keywords like “human-computer
interaction” and “information technology” are emerging themes, poorly developed, and
marginal compared to the AI and CBE fields.

AE
Figure no. 6. Thematic map

The research results provide a detailed insight into how AI intersects with the CBE field,
which has multiple practical and theoretical implications. An interdisciplinary approach is
observed, highlighted by works published in prestigious journals such as Nature, MIS
Quarterly, and IEEE Access. This approach emphasises the complexity and vast thematic
intersections between AI and CBE. Another relevant aspect is the focus of AI research on
medical education, as suggested by the medical journals included in the cocitation network
and references to articles in this field. This aligns with the literature that demonstrated the
effectiveness of CBE in medicine (Brammer and Goodrich, 2021) and, at the same time, the
great potential of AI to revolutionise teaching and learning methods in this field (Chen,
Chen, and Lin, 2020).
The two highlighted periods underscore the adaptability of research to global events.
Therefore, while the first period focused on fundamental technologies, the second placed
more emphasis on education and incorporated major events such as the COVID-19
pandemic, also being a period when CBE was studied more intensively (Paek, Um and
Kim, 2021). As technology continues to evolve, it is essential to stay in touch with trends to
maximise the benefits of AI in education.
The results show a wide range of subdomains within AI that are relevant for CBE, from
"deep learning" and robotics to adaptive educational systems. Thus, a great potential for AI
is observed to contribute to various aspects of the educational process. There is a clear trend
towards digitalisation, with terms like “e-learning“, “virtual reality”, and “information
technology” being the centre of attention. This could suggest a fundamental shift in how
education is delivered and experienced.
Vol. 26 • No. 65 • February 2024 235

Another noteworthy aspect is the growing interest in adaptive and personalised learning,
highlighted by the frequent appearance of terms related to adaptive educational systems.
We observe a movement towards more personalised solutions in education that cater to the
individual needs of students; as we know, this is a central trend in CBE (Chen, Chen, and
Lin, 2020; Hwang et al., 2020; Brammer and Goodrich, 2021; Kuleto et al., 2021; Mhlanga,
2021). The evolution of education in recent decades, especially the transition from a
traditional learning model to a competency-based one, emphasises its adaptability and the
ongoing need for innovation in this domain. With a growing emphasis on developing
practical skills and relevant knowledge, CBE has been proven to efficiently address real-
world challenges.
Lastly, we want to stress the importance of this study; while there are previous bibliometric
analyses that investigated the link between artificial intelligence and education (Hinojo-
Lucena et al., 2019; Baek and Doleck, 2020; Hwang et al., 2020; Talan, 2021; Harto et al.,
2022; Jia et al., 2022; Prahani et al., 2022), our research specifically focusses on the use of
AI within CBE, thus making a significant contribution to fill a knowledge gap in this
domain. By bringing a focused perspective to this specific subject, we hope that the results
obtained will shape new insights into how AI can influence and optimise CBE processes in
an innovative and practical manner.
Conclusions
The bibliometric analysis provided in this article reveals an exponential growth in academic
interest in the intersection between AI and CBE. With a global presence of research and the
involvement of numerous countries and top academic institutions, it is clear that this
intersection represents a frontier zone for innovation in education. The increased emphasis
on this topic in recent years indicates a rapid adaptation of the academic community to
technological advancements and global challenges, such as COVID-19.
Thematic analyses highlight the evolution of AI in education, illustrating the shift from
focussing on technological fundamentals to interdisciplinary approaches that combine
technology with global events and various fields, such as medicine. In addition,
collaboration networks underline the value added by cooperation and international
knowledge exchange.
With respect to the limitations of the research, the Web of Science (WoS) database was
used exclusively. Consequently, there is a possibility that relevant works indexed in other
databases (such as Scopus) or published in regional or niche journals not included in WoS
might have been excluded. Additionally, the exclusive selection of works in the English
language can limit the global representation of research in the field of AI and CBE, and the
search keywords employed can be deemed restrictive. However, we believe that these
limitations had a relatively minor impact and that the conclusions drawn are valid.
As the main direction of future research, the authors aim to conduct a quantitative study
investigating how the practical implementation of AI in CBE influences educational
outcomes and the development of competencies among students. The key competencies
considered will be those related to leadership. Therefore, the study will aim to assess the
level of development of leadership competencies before and after the implementation of AI

AE
in CBE, identify significant differences in student academic results, and analyse how
technology impacts both academic performance and competency development.
From a theoretical standpoint, this work not only fills a gap in the specialised literature but
also provides a solid foundation for future research. The intersection between AI and CBE
is full of opportunities and challenges. The interdisciplinary approach and global
collaboration will be essential to navigate this complex and dynamic research area. From a
practical perspective, there are several significant managerial implications for educational
institutions and leaders in the field of education. Thus, educational managers should
evaluate and adapt educational programmes to include competency-based elements,
reflecting the rapid advancement of artificial intelligence in education. This involves
constantly reviewing and updating the content of the curriculum to meet the dynamic
demands of the market. Additionally, managers must ensure that teachers and educational
staff are prepared to use these technologies and align with technological developments in
the field of education. Last but not least, it is essential for managers to continuously
monitor trends in research and development in the field of artificial intelligence so that they
can adjust educational strategies based on new discoveries and innovations.
References
Akgun, S. and Greenhow, C., 2022. Artificial intelligence in education: Addressing ethical
challenges in K-12 settings. AI and Ethics, 2(3), pp. 431-440. https://doi.org/10.1007/
s43681-021-00096-7.
Almohammadi, K., Hagras, H., Alghazzawi, D. and Aldabbagh, G., 2017. A survey of
artificial intelligence techniques employed for adaptive educational systems within
e-learning platforms. Journal of Artificial Intelligence and Soft Computing Research,
7(1), pp. 47-64. https://doi.org/10.1515/jaiscr-2017-0004.
Anderson-Levitt, K. and Gardinier, M.P., 2021. Introduction contextualising global flows
of competency-based education: polysemy, hybridity and silences. Comparative
Education, 57(1), pp. 1-18. https://doi.org/10.1080/03050068.2020.1852719.
Aria, M. and Cuccurullo, C., 2017. bibliometrix: An R-tool for comprehensive science
mapping analysis. Journal of Informetrics, 11(4), pp. 959-975. https://doi.org/
10.1016/j.joi.2017.08.007.
Azañedo, D., Visconti-Lopez, F.J. and Hernández-Vásquez, A., 2022. A Web of Science-
Based Bibliometric Analysis of Global Noma Publications. Tropical Medicine and
Infectious Disease, 7(8), article no. 198. https://doi.org/10.3390/tropicalmed7080198.
Baek, C. and Doleck, T., 2020. A Bibliometric Analysis of the Papers Published in the
Journal of Artificial Intelligence in Education from 2015-2019. International Journal
of Learning Analytics and Artificial Intelligence for Education (iJAI), 2(1), pp. 67-84.
https://doi.org/10.3991/ijai.v2i1.14481.
Baker, R.S. and Hawn, A., 2022. Algorithmic bias in education. International Journal of
Artificial Intelligence in Education, 32(4), pp. 1052-1092.
Borenstein, J. and Howard, A., 2021. Emerging challenges in AI and the need for AI ethics
education. AI and Ethics, 1(1), pp. 61-65. https://doi.org/10.1007/s43681-020-00002-7.
Vol. 26 • No. 65 • February 2024 237

Brammer, M.K. and Goodrich, K.M., 2021. Competency-based education model: is it

appropriate for counselor education? Social Science Journal, pp. 1-12.
https://doi.org/10.1080/03623319.2021.1883380.
Burnette, D.M., 2016. The Renewal of Competency-Based Education: A Review of the
Literature. Journal of Continuing Higher Education, 64(2), pp. 84-93.
https://doi.org/10.1080/07377363.2016.1177704.
Chen, S., Zhang, C. and Li, W., 2022. The effects of competency-based training model in
the training of new nurses: A meta-analysis and systematic review. PLOS ONE,
17(11), article no. e0277484. https://doi.org/10.1371/journal.pone.0277484.
Ellahi, R.M., Khan, M.U.A. and Shah, A., 2019. Redesigning Curriculum in line with
Industry 4.0. Procedia Computer Science, 151, pp. 699-708.
Evans, C.M., Landl, E. and Thompson, J., 2020. Making sense of K‐ 12
competency‐ based education: A systematic literature review of implementation and
outcomes research from 2000 to 2019. The Journal of Competency-Based Education,
5(4), pp. 1-28. https://doi.org/10.1002/cbe2.1228.
Fosso Wamba, S., Bawack, R.E., Guthrie, C., Queiroz, M.M. and Carillo, K.D.A., 2021.
Are we preparing for a good AI society? A bibliometric review and research agenda.
Technological Forecasting and Social Change, 164, article no. 120482.
Harto, B., Saymsu, Y.L., Rukmana, A.Y., Komalasari, R. and Dwijayanti, A., 2022.
Bibliometric Analysis of Transforming Leadership Education with Artificial Intelligence.
1st Virtual Workshop on Writing Scientific Article for International Publication Indexed
SCOPUS, pp. 385-390. https://doi.org/10.2478/9788366675827-067.
Hinojo-Lucena, F.-J., Aznar-Díaz, I., Cáceres-Reche, M.-P. and Romero-Rodríguez, J.-M.,
2019. Artificial Intelligence in Higher Education: A Bibliometric Study on its Impact
in the Scientific Literature. Education Sciences, 9(1), article no. 51.
https://doi.org/10.3390/ educsci9010051.
Santos, O.C., Rodrigo, M.T., Cukurova, M., Bittencourt, I.I. and Koedinger, K.R.,
2022. Ethics of AI in Education: Towards a Community-Wide Framework.
International Journal of Artificial Intelligence in Education, 32(3), pp. 504-526.
https://doi.org/10.1007/s40593-021-00239-1.
Hwang, G.J., Xie, H., Wah, B.W. and Gašević, D., 2020. Vision, challenges, roles and
research issues of Artificial Intelligence in Education. Computers and Education:
Artificial Intelligence, 1, article no. 100001. https://doi.org/10.1016/j.caeai.2020.
100001.
Jia, K., Wang, P., Li, Y., Chen, Z., Jiang, X., Lin, C.L. and Chin, T., 2022. Research
Landscape of Artificial Intelligence and e-Learning: A Bibliometric Research.
Frontiers in Psychology, 13, p. 795039. https://doi.org/10.3389/fpsyg.2022.795039.
Johnstone, S.M. and Soares, L., 2014. Principles for Developing Competency-Based
Education Programs. Change: The Magazine of Higher Learning, 46(2), pp. 12-19.
https://doi.org/10.1080/00091383.2014.896705.

AE
Kanungo, R.P., Gupta, S., Patel, P., Prikshat, V. and Liu, R., 2022. Digital consumption and
socio-normative vulnerability. Technological Forecasting and Social Change, 182,
article no. 121808. https://doi.org/10.1016/J.TECHFORE.2022.121808.
Kaplan, A. and Haenlein, M., 2019. Siri, Siri, in my hand: Who’s the fairest in the land? On
the interpretations, illustrations, and implications of artificial intelligence. Business
Horizons, 62(1), pp. 15-25. https://doi.org/10.1016/j.bushor.2018.08.004.
Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser,
U., Groh, G., Günnemann, S., Hüllermeier, E., Krusche, S., Kutyniok, G., Michaeli, T.,
Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T., Stadler, M.,
Weller, J., Kuhn, J. and Kasneci, G., 2023. ChatGPT for good? On opportunities and
challenges of large language models for education. Learning and Individual
differences, 103, article no. 102274. https://doi.org/10.1016/j.lindif.2023.102274
Kuleto, V., Ilić, M., Dumangiu, M., Ranković, M., Martins, O.M.D., Păun, D. and
Mihoreanu, L., 2021. Exploring opportunities and challenges of artificial intelligence
and machine learning in higher education institutions. Sustainability, 13(18), article no.
10424. https://doi.org/10.3390/su131810424.
Lazaroiu, G., and Rogalska, E. 2023. How generative artificial intelligence technologies
shape partial job displacement and labor productivity growth. Oeconomia
Copernicana, 14(3), pp. 703-706. https://doi.org/10.24136/oc.2023.020.
Lindsay, T., 2018. New study: Less expensive competency-based education programs just
as good as traditional programs. [online] Forbes. Available at:
<https://www.forbes.com/sites/tomlindsay/2018/03/27/new-study-less-expensive-
competency-based-education-programs-just-as-good-as-traditional-programs/?sh=
32d24c21674df2> [Accessed 22 July 2023].
Mehall, S., 2019. How online competency‐ based education can enable greater access to
higher education. The Journal of Competency-Based Education, 4(4), pp. 2-5.
https://doi.org/10.1002/cbe2.1201.
Mhlanga, D., 2021. Artificial intelligence in the industry 4.0, and its impact on poverty,
innovation, infrastructure development, and the sustainable development goals:
Lessons from emerging economies? Sustainability, 13(11), article no. 5788.
https://doi.org/10.3390/su13115788.
Nguyen, A., Ngo, H.N., Hong, Y., Dang, B. and Nguyen, B.P.T., 2022. Ethical principles
for artificial intelligence in education. Education and Information Technologies, 28(4),
pp. 4221-4241. https://doi.org/10.1007/s10639-022-11316-w.
Paek, S., Um, T. and Kim, N., 2021. Exploring latent topics and international research
trends in competency-based education using topic modeling. Education Sciences,
11(6), article no. 303. https://doi.org/10.3390/educsci11060303.
Peters M.A, Jackson, L., Papastephanou, M., Jandrić, P., Lazaroiu, G., Evers, C.W., Cope,
B., Kalantzis, M., Araya, D., Marek, T., Mika, C., Sturm, S., Rider, S. and Fuller, S.,
2023. AI and the future of humanity: ChatGPT-4, philosophy and education – Critical
responses. Educational Philosophy and Theory, pp. 1-35. https://doi.org/10.1080/
00131857.2023.2213437.
Poria, S., Majumder, N., Mihalcea, R. and Hovy, E., 2019. Emotion recognition in
conversation: Research challenges, datasets, and recent advances. IEEE Access, 7,
pp. 100943-100953. https://doi.org/10.1109/ACCESS.2019.2929050.
Vol. 26 • No. 65 • February 2024 239

Prahani, B.K., Rizki, I.A., Jatmiko, B., Suprapto, N. and Amelia, T., 2022. Artificial
Intelligence in Education Research During the Last Ten Years: A Review and
Bibliometric Study. International Journal of Emerging Technologies in Learning,
17(8), pp. 169-188. https://doi.org/10.3991/ijet.v17i08.29833.
Talan, T., 2021. Artificial Intelligence in Education: A Bibliometric Study. International
Journal of Research in Education and Science, 7(3), pp. 822-837. https://doi.org/
10.46328/ijres.2409.
Ulrik, B., 2001. A faster algorithm for betweenness centrality. Journal of Mathematical
Sociology, 25(2), pp. 163-177. https://doi.org/10.1080/0022250X.2001.9990249
Winkler-Schwartz, A., Bissonnette, V., Mirchi, N., Ponnudurai, N., Yilmaz, R., Ledwos,
N., Siyar, S., Azarnoush, H., Karlik, B. and Del Maestro, R. F., 2019. Artificial
intelligence in medical education: best practices using machine learning to assess
surgical expertise in virtual reality simulation. Journal of Surgical Education, 76(6),
pp. 1681-1690. https://doi.org/10.1016/j.jsurg.2019.05.015.
Yip, H.K. and Smales, R.J., 2000. Review of competency-based education in dentistry.
British Dental Journal, 189(6), pp. 324-326. https://doi.org/10.1038/sj.bdj.4800758.

AE
ARTIFICIAL INTELLIGENCE IN BUSINESS EDUCATION:

BENEFITS AND TOOLS
Camelia Surugiu1 , Cătălin Grădinaru2* and Marius – Răzvan Surugiu3
1)
University of Bucharest, Romania and National Institute of Research and
Development in Tourism, Bucharest, Romania
2)
University of Bucharest, Romania
3)
Romanian Academy, Institute of National Economy, Bucharest, Romania

Surugiu, C., Grădinaru, C. and Surugiu, M.R., 2024. Received: 28 September 2023
Artificial Intelligence in Business Education: Benefits Revised: 28 November 2023
and Tools. Amfiteatru Economic, 26(65), pp. 241-258. Accepted: 22 December 2023
Abstract
Understanding the impact of artificial intelligence (AI) on education is vital for guiding
teachers in developing educational tools. AI in education (AIEd) comes not only with
opportunities but mostly with challenges for both educators and learners. Finding the proper
tools to integrate AI into the learning framework represents a test for current and future
generations. Even if most students acknowledged AI as a valuable tool, their interaction with
AI in education seems more limited than expected. They mainly concentrated on few tools
with higher awareness.
This paper examines AI’s support for educational activities, key drivers, and tools for
business education. Survey data collected from 254 learners were analysed using multivariate
binary logistic regression. Two research questions were formulated to verify if AI supports
educational activities and what AI tools support business educational activities. Results show
learners appreciate AI for aiding teachers in administrative tasks, personalising learning
plans, and saving time. However, learners are unfamiliar with most benefits of AI tools,
except computer vision, edge computing, and AI chatbots. The paper highlights the need to
increase the use of AI in education to make students more familiar with AI tools and capitalise
on them in business education.
Keywords: artificial intelligence (AI), artificial intelligence in education (AIEd),
instruments, benefits, Romania
JEL Classification: A23, I23, M15, O33
*
Corresponding author, Cătălin Grădinaru – e-mail: [email protected]
Vol. 26 • No. 65 • February 2024 241

AE Artificial Intelligence in Business Education: Benefits and Tools
Introduction
In envisioning future jobs and human resources specialists’ requirements, teachers must
equip students with a proficiency that makes them adaptable to challenges. Teachers
significantly contribute to society by educating generations of upcoming students that will
become future entrepreneurs. Tan (2020) underlines that the teacher exerts a triple role, being
a content expert and creator, a knowledge spreader, and “an ethical-spiritual guide” with
wisdom. The students should be able to chase trends and not just stay current with the events
and environment (Elhajjar, Karam and Borna, 2021). Hence, the demand for innovative
teachers embracing change, integrating new materials, and enabling student-AI interactions
in education is increasing.
AI augments human skills in the workplace (Nuseir, Basheer and Aljumah, 2020) and serves
as an educational partner, enhancing content and competencies (Tan, 2020). Elhajjar, Karam
and Borna (2021) advocate integrating AI into education to equip students with skills
essential for future jobs and digital society’s demands, such as innovation, creativity, and
design thinking. AIEd employs diverse tools, techniques, and systems in educational
activities (McGrath et al., 2023). Investing in human capital to embrace AIEd tools is vital
for societal development, despite widespread distrust and misconceptions about AI's role in
human activities, especially in education (Antonenko and Abramowitz, 2023).
AI technology in education is expected to grow significantly in the coming decades,
presenting new opportunities and challenges (Ouyang and Jiao, 2021; Zhang and Aslan,
2021; Khosravi, 2022). Researchers, policymakers, and practitioners are integrating AIEd to
enhance teaching, personalised learning, assessments, and administrative services (Zhang
and Aslan, 2021; Chiu et al., 2023). AI represents progress in education, offering benefits on
multiple levels, and stimulates the evolution of teaching and learning through technologies
like chatbots, robots, automated assessment, digitised artefacts, and intelligent tutoring
systems, despite occasional organisational challenges (Chiu et al., 2023).
The demand for adaptive digital learning with AI support has surged in the past decade,
driven by challenges like the Covid-19 pandemic and social conflicts in contemporary
societies (Matzavela and Alepis, 2021). Online and blended learning are now prevalent in
modern communities and emerging economies as they strive to integrate these methods into
the educational system. AIEd provides feasible solutions to complex societal problems,
enabling students to engage closely with global challenges and develop real-life problem-
solving skills (Southworth et al., 2023).
In this study, we analyze the AI determining factors and tools that promote its benefits in
educational activities for business.The paper is structured as follows. The next section depicts
the literature review on AI in education and business education. Section two presents the
research methodology. Section three reviews the data analysis and discusses the results. The
last section exposes the conclusions of the paper.
The authors identify a gap in the literature regarding the students’ limited knowledge of the
benefits of AI in business education, even though many of them have previously used AI.
Additionally, it is noted that there is a lack of preparedness among the respondents to adopt
AI in the educational process. The study makes significant contributions by conducting an
empirical investigation that explores the AI determining factors and tools in the context of
business education. Additionally, the study addresses the need for continuous improvements
in the implementation of AI in business education. The study holds both theoretical and

AE
practical significance. By identifying existing gaps in students’ knowledge and their low level
of preparedness for adopting AI in business education, it provides a foundation for the further
development of educational strategies. The study’s findings can serve as a guide for
educators, researchers, and the educational community in improving practices related to the
use of AI in the learning process. By highlighting the need for active engagement in AI-
assisted education innovation, the study contributes to shaping future research directions and
actions to maximize the benefits derived from integrating AI into business education.

AI in education requires transdisciplinary skills to enhance the learning experience
(Southworth et al., 2023). AIEd aids teachers and students in educational processes (Hopcan
et al., 2022) and supports administrative tasks, educational services, assessments, and
procedures.
1.1. Artificial intelligence in education for learners

AI systems in education significantly enhance learners’ involvement and performance
(Zhang and Aslan, 2021). Chiu et al. (2023) identified essential AI roles for learners:
competency-based task assignments, learner-machine discussions, feedback, and adaptive
digital environments. Zhang and Aslan (2021) added other AI facilitations, including
engagement, enriched learning resources, and intellectual stimuli. Southworth et al. (2023)
highlight AIEd’s benefits, fostering technical skills, creativity, critical thinking, and problem-
solving abilities for students.
AIEd personalises learning experiences and communication to cater to individual needs and
abilities, enhancing efficiency (Hopcan et al., 2022; Southworth et al., 2023). Customised
learning materials are tailored based on students’ evaluations, addressing their strengths and
weaknesses (Hopcan et al., 2022). Li and Wang (2023) propose using advanced technology
to create a comfortable communication environment, fostering learner networks with
increased information accessibility for future generations.
Adapting learning experiences sustains student progress and engagement in virtual
environments, promoting skill development (Southworth et al., 2023). Interactivity and
participation substantially increase among students and teachers. Khosravi et al. (2022)
emphasise the significance of AI in various learning interfaces. For remote learners, AI offers
a crucial advantage through simulation cases on complex life topics challenging to address
in traditional settings (e.g., welfare system, losses, and violence) (Asakura et al., 2020).
Ouyang and Jiao (2021) identified three AIEd paradigms: “AI-direct, learner-as-recipient”,
where AI leads learning with a defined pathway for the learner, “AI-supported, learner-as-
collaborator”, where AI optimises interaction among learners, information, and technology,
and “AI-empowered, learner-as-leader”, where AI enhances learners’ intelligence through a
complex system. Cope, Kalantzis and Searsmith (2021) acknowledge AI’s intrinsic
limitations, cautioning against reducing students’ results to simple numbers without human
thought and implications beyond AI's cognitive processes and calculations.
Vol. 26 • No. 65 • February 2024 243

1.2. Artificial intelligence in education for teachers

Crompton, Jones and Burke (2022) emphasise that AIEd builds upon previous learning
theories, facilitating teacher adoption of educational tools and integrating best practices to
enhance learning. AIEd benefits teachers and the teaching process through tailored content
for individual learners, at-risk or gifted students, learning predictive models, personalised
educational resources, improved classroom management, enhanced teaching across various
subjects, academic progress facilitation, and qualified development in pedagogical skills,
human behaviour, and interactions (Zhang and Aslan, 2021; Chiu et al., 2023).
AI unlocks new research potential in universities by expanding experimental and
investigative activities and disseminating study results (Li and Wang, 2023). Teachers can
focus on complex tasks as AI provides student support and answers (Southworth et al., 2023).
Khosravi (2022) advocates integrating explainable AI (XAI) in education, prioritising
human-centred design of educational tools, assessing XAI implementation, and enhancing
AIEd systems for reliability and knowledge transfer support. Many teachers avoid AIEd due
to a lack of understanding (Crompton, Jones and Burke, 2022). To effectively integrate AI
tools, teachers should update their educational practices and embrace the challenges of
technology in teaching (Skavronskaya, Hadinejad and Cotterell, 2023). In the future, teachers
can integrate into their pedagogical approach aspects related to learners’ behaviour, such as
emotions, attention, gestures, and movement, with the support of AI technologies (Crescenzi-
Lanna, 2023).
1.3. Artificial intelligence in education in assessment and administration

Chiu et al. (2023) identified two main contributions of AIEd assessment: automatic grading
and learners’ foreseen performance in online classes. The feedback provided to students after
the assessment process is immediate and constructive, and instructors can support students
in improving specific skills (Hopcan et al., 2022; Southworth et al., 2023). Cope, Kalantzis
and Searsmith (2021) underline the role of AI in the assessment process, not in the
conventional form, but especially related to tracking progress and providing “just-in-time
feedback”. AI can improve assessments (Al Braiki et al., 2020), with benefits highlighting
its efficiency and consistency (Chassignol et al., 2018) and presenting the possibility of
automating the assessment process.
The sustenance provided by AI in education refers to supporting and enhancing the
performance of management platforms, analysis of scaled data, personalisation of various
academic services, increasing work efficiency of administrative staff, and consistency in the
decision-making process (Zhang and Aslan, 2021; Chiu et al., 2023). The work AI can
perform in education also refers to completing tasks that consume instructor’s time at a faster
pace, identifying students’ preferences and learning styles to generate personalised learning
plans, and providing them with timely and direct feedback or assisting teachers in their data-
related decisions or work (Chen, Chen and Lin, 2020).
1.4. Artificial intelligence in business education and AI tools

AI-based education enhances entrepreneurial competencies and fosters creativity, benefiting
businesses (Nuseir, Basheer and Aljumah, 2020). While AI offers real support in business

AE
scenarios for students, further conceptualisation of AI’s use in business and education is
needed, covering critical aspects like processes, activities, and actors (Yang et al., 2022).
AI tools have proven firmly and vastly helpful in various fields, in education or business
education. Amongst them, there are computer vision, prediction systems, data mining (Das
et al., 2015), intelligent learning or teaching systems, learning analytics (Ley et al., 2023),
facial recognition systems, voice or speech recognition systems, virtual laboratories,
augmented reality, virtual reality, hearing and sensing technologies, edge computing, virtual
personalised assistants, real-time analysis, AI chatbot, image recognition, personalised
learning approach, academic analytics, and adaptive learning method (al-Zyoud, 2020; Han,
Park and Lee, 2022).
The article uses quantitative research to examine how AI supports educational activities,
significant drivers, and tools for business education. The study aimed to determine whether
respondents thought of AI as a tool that supported educational activities and which AI
technologies supported business academic endeavours.
Students are direct beneficiaries of the educational process, and therefore, the study focuses
on their situation to highlight the immediate impact of AI in educational activities. It is crucial
to understand in detail how AI can influence them, improving the quality of the learning
process and academic outcomes. Based on the analysis of the literature (Das et al., 2015;
Asakura et al., 2020; Al Braiki et al., 2020; Chen, Chen and Lin, 2020; Nuseir, Basheer and
Aljumah, 2020; al-Zyoud, 2020; Ouyang and Jiao, 2021; Cope, Kalantzis and Searsmith,
2021; Zhang and Aslan, 2021; Hopcan et al., 2022; Khosravi et al., 2022; Crompton, Jones
and Burke, 2022; Yang et al., 2022; Han, Park and Lee, 2022; Chiu et al., 2023; Southworth
et al., 2023; Li and Wang, 2023; Skavronskaya, Hadinejad and Cotterell, 2023; Crescenzi-
Lanna, 2023; Ley et al., 2023), two main research questions have been formulated, namely:
Q1: Have you used AI in educational activities?
Q2: Do you consider AI beneficial in business educational activities?
To perform multivariate statistical analysis, binary logistic regression is often recommended
(Toumi, 2020; Rahman et al., 2021). To respond to these questions, the econometric model
deployed for data analysis was binary logistic regression, which can be written as
(Bangdiwala, 2018):
𝑒 𝑋𝛽
P(𝑌 = 1 | 𝑋) = 𝑙𝑜𝑔𝑖𝑡 −1 (𝑋𝛽) = (1)
1+𝑒 𝑋𝛽
In the equation, P(𝑌 = 1 | 𝑋) represents the probability that variable Y would be affected if
the predictor variables X are taken into consideration; Y is the dependent variable with only
one of two outcomes (1 – the outcome that is trying to predict; 0 – the other outcome), 𝑋𝛽 is
the linear predictor function. The inverse of the logit function gives the probability of Y
having the value of 1.
The coefficient β shows if a direct or an inverse association exists between the dependent
variable and its predictors. The odds ratios higher than 1 indicate that as the independent
variable increases, so do the odds of the dependent variable, but the values of less than
Vol. 26 • No. 65 • February 2024 245

1 suggest that if the independent variable increases, the odds of the dependent variable
decrease (Field, 2013). The response variable is binary, and in our research, the first
dependent variable used was AIEd corresponding to the question from the questionnaire:
Have you used AI in educational activities? Starting from the literature review, eighteen
independent variables were used to explain AI usage by learners. The AIEd predictors are
described in table no. 1.
Table no. 1. Artificial intelligence in education (AIEd) and predictors
Predictors Coding
AI performs administrative tasks instead of teachers AI_AD
AI identifies the degree of fulfilment of students’ work tasks AI_DW
AI enables learning outside the classroom AI_CLASS
AI helps teachers develop personalised learning plans for each student AI_P
AI discovers learning gaps in students AI_DIS
AI helps generate ideas for projects / other curricular and extracurricular activities AI_NI
AI reduces human error (e.g., corrections) AI_HE
AI identifies copyright issues AI_HR
AI will eliminate the role of the teacher in the future AI_T
AI will increase tuition costs AI_COST
AI cannot give me emotional support AI_ES
AI, together with teachers, creates a complete approach to the educational act AI_CA
AI allows the connection with the business environment AI_BE
AI can be used to train teachers AI_TT
AI improves school performance AI_SP
AI reduces the time allocated to learning AI_TIME
AI increases the employability of students in the business environment AI_EMPLOY
AI stimulates the desire to learn and assimilate new knowledge AI_LD
The second research question is Q2: Do you consider AI beneficial in business educational
activities? The respondents were asked to select from nineteen AI tools that learners can use
in business education (AIEd_B) (Table no. 2).
Table no. 2. Artificial intelligence in education (AIEd_B) and predictors
Predictors Coding Predictors Coding
Computer-Vision CV Virtual Reality VR
Prediction Systems PS Hearing and Sensing HST
Technologies
Data Mining DM Edge computing EC
Intelligent Learning Systems EIS Virtual Personalized Assistants VPA
Learning Analytics LA Real-Time Analysis RTA
Facial Recognition Systems FAS AI chatbot CBT
Voice Recognition Systems VAS Image recognition IR
Virtual Laboratories VL Personalised Learning Approach PLA
Augmented Reality AR Academic Analytics AA
Adaptive Learning Method ALM
The data were collected through a questionnaire administrated via Google Forms in June –
July 2023, comprising dichotomic questions (2) and close-ended questions with Likert scale
answers. The study had 254 respondents. The gender distribution was 41.7% male and 58.3%
female (Table no. 3). Regarding age, 94.9% were 18-26 years old, while 5.1% were 26 years
old or above. Concerning income, 34.6% had 3,000 lei or less, 25.9% had more than

AE
3,000 lei, and 39.5% did not declare any income. Education-wise, 27.2% had a bachelor’s
degree or higher, and 72.8% had high school, post-secondary, or professional studies. As for
occupation, 15.0% were employees, and 85.0% were students. The convenience sampling
technique was applied to select respondents conveniently (Edgar and Manz, 2017). This
sampling technique involves selecting respondents from a convenient subset of the
population (Baxter et al., 2015). Although this technique may be perceived as the weakest
method of non-probabilistic sampling, it is often used to obtain a range of attitudes and
opinions that can be further tested in future research (Albert et al., 2010).
The questionnaire was administered to assess how AI influences educational activities and
to evaluate AI tools supporting business educational activities. The respondents are students
from business education programs. The provided responses are anonymous, ensuring the
confidentiality of the collected data.
The statistical analysis was conducted using IBM SPSS Statistics version 29.0, employing
binary logistic regression. The variables were assessed for multicollinearity (Field, 2013),
using the variance inflation factor (VIF) to estimate how much the change of a regression
coefficient rises if the independent variables are correlated. A VIF between 5 and 10 indicates
a high correlation between predictors.
1
𝑉𝐼𝐹𝑗 = (2)
1−𝑅𝑗2
In the above equation, 𝑅𝑗2 is the coefficient of determination (R-squared) for linear regression.
Table no. 3. Descriptive statistics of respondents’ profile
Measure Item Frequency (%)
N = 254
Gender Male 106 (41.7%)
Female 148 (58.3%)
Age 18-25 241 (94.9%)
26 or above 12 (5.1%)
Average monthly income 3,000 lei or less 88 (34.6%)
More than 3,000 lei 66 (25.9%)
No income 100 (39.5%)
Education High school / post-secondary studies/ 185 (72.8%)
professional studies
Bachelor’s degree / Postgraduate studies 69 (27.2%)
Occupation Employee 38 (15.0%)
Student 216 (85.0%)
To test the model fit, we employed Omnibus Tests of Model Coefficients to determine if
there was a significant improvement compared to the null model. The Hosmer and Lemeshow
test was also used to assess the difference between the observed and predicted models (Field,
2013).
(𝑂𝑖 −𝐸𝑖 )2
𝐻 = ∑𝐺𝑖=1 𝐸 (3)
𝐸𝑖 ∗(1− 𝑖 )
𝑁
In the above equation, G represents the number of groups created based on the probabilities
predicted by the model, Oi represents the observed frequency in group i, Ei represents the
expected frequency in group i, and N represents the total number of observations.
Vol. 26 • No. 65 • February 2024 247

The model summary used Nagelkerke’s R-square, an adjusted version of the Cox and Snell
R-square ranging from 0 to 1 (Pallant, 2005). Model validation included receiver operating
characteristic curve (ROC curve) analysis, measuring the area under the ROC curve (AUC).
Within the analysis, the issue of endogeneity has not been addressed. Since both independent
and dependent variables are based on opinions expressed by the same individual, there is a
risk that the observed relationship may not only reflect direct causality but also reciprocal
influences or effects of an unobserved variable. This may affect the generalizability of the
results to other populations or contexts. Thus, the findings of this study may be influenced
by the specific characteristics of the subjects and the context in which the data were collected,
and the extrapolation of results to other groups or situations may be limited. It is necessary
for future research to adopt specific methods to manage these aspects.

For multivariate binary logistic regression, we used AIEd as dependent variable, and the
variables AI_AD, AI_DW, AI_CLASS, AI_P, AI_DIS, AI_NI, AI_HE, AI_HR, AI_T,
AI_COST, AI_ES, AI_CA, AI_BE, AI_TA, AI_SP, AI_TIME, AI_EMPLOY, and AI_LD
as independent variables fitted into the model as will appear in table no. 5. The model fitting
effect was tested by ROC (figure no. 1 (a)). The variables AI_AD, AI_DW, AI_CLASS,
AI_P, AI_DIS, AI_NI, AI_HE, AI_HR, AI_T, AI_COST, AI_ES, AI_CA, AI_BE, AI_TA,
AI_SP, AI_TIME, AI_EMPLOY, and AI_LD jointly projected the AIEd with an AUC of
0.723. According to Hosmer, Lemeshow and Sturdivant (2013), we consider this acceptable
discrimination if 0.7≤ROC<0.8. Consequently, this regression model has good sensitivity
and specificity. For the model, Chi-square (df:18) = 43.369 was statistically significant, with
p-value < 0.05 (=0.011) showing that the full model has a considerable prediction
performance. Moreover, the model explained 18.1% of the variance of AIEd (Nagelkerke R-
square), predicted percentage correct at 73.2%. For the model fit, we also evaluated Hosmer
and Lemeshow test, Chi-square (df:8) = 5.143, df=8, p-value > 0.05 (=0.742), which showed
that the deployed model fitted the data. The collinearity statistics suggest that there isn’t any
collinearity relationship between predictors as their VIF values range between 1.433 and
3.298, as is to be seen in table no. 4.
The second dependent variable used in the multivariate binary logistic model was AIEd_B.
The independent variables were: CV, PS, DM, EIS, LA, FAS, VAS, VL, AR, VR, HST, EC,
VPA, RTA, CBT, IR, PLA, AA, and ALM, as to be mentioned in table no. 6. The model
fitting effect tested by ROC and the predictors jointly projected the AIEd_B with an AUC of
0.817 (figure no. 1 (b)). According to Hosmer, Lemeshow and Sturdivant (2013), we consider
this excellent discrimination if 0.8≤ROC<0.9. Therefore, the second regression model has
good sensitivity and specificity. In this model, Chi-square (df:19) = 38.941 was statistically
significant, p-value < 0.01 (=0.004). According to Nagelkerke R-square, the model explained
28.9% of the variance of AIEd_B, predicted percentage correct at 90.2%. Hosmer and
Lemeshow test, Chi-square (df:8) = 12.550, df=8, p-value > 0.05 (=0.128) showed that the
deployed model fitted the data.

AE
(a) (b)
Figure no. 1. ROC curve
The collinearity statistics indicate that there isn’t a collinearity relationship between
mentioned predictors, as the VIF values are greater than 1 and below 5, ranging between
2.092 and 4.339 (table no. 4).
Within the study, the learners’ reasons for using AI in educational activities were identified,
a multivariate binary logistic model being used. In this model, AIEd was considered as the
dependent variable. Thus, considering the first research question (Q1: Have you used AI in
educational activities?), the analysis results show that four factors are statistically significant
at the 95% confidence level (p-value < 0.1) among the predictors, as will be observed in table
no. 5. Those factors were: AI_AD (AI performs administrative tasks instead of teachers),
AI_P (AI helps teachers develop personalised learning plans for each student), AI_SP (AI
improves school performance), and AI_TIME (AI reduces the time allocated to learning).
Table no. 4. Collinearity statistics
Factor Tolerance VIF Factor Tolerance VIF
AI_AD 0.406 2.461 CV 0.478 2.092
AI_DW 0.397 2.516 PS 0.379 2.639
AI_CLASS 0.449 2.229 DM 0.472 2.119
AI_P 0.315 3.173 EIS 0.437 2.288
AI_DIS 0.409 2.447 LA 0.380 2.634
AI_NI 0.460 2.175 FAS 0.244 4.102
AI_HE 0.534 1.871 VAS 0.264 3.788
AI_HR 0.570 1.755 VL 0.359 2.786
AI_T 0.557 1.794 AR 0.393 2.544
AI_COST 0.560 1.785 VR 0.375 2.665
AI_ES 0.698 1.433 HST 0.393 2.546
AI_CA 0.368 2.714 EC 0.336 2.976
AI_BE 0.382 2.619 VPA 0.354 2.826
AI_TA 0.374 2.671 RTA 0.343 2.912
AI_SP 0.303 3.298 CBT 0.437 2.286
AI_TIME 0.512 1.951 IR 0.365 2.743
AI_EMPLOY 0.378 2.643 PLA 0.302 3.315
AI_LD 0.404 2.474 AA 0.247 4.041
ALM 0.230 4.339
The variable AI_AD was found to have a significant relationship with AIEd (p-value < 0.05).
The results in the model indicated that the probability of using AI in educational activities
Vol. 26 • No. 65 • February 2024 249

increases by 1.488 times when considering that AI performs administrative tasks instead of
teachers, considering all the other variables being constant. Such tasks may include analysing
students' work, providing feedback, grading, or detecting plagiarism (Chen, Chen and Lin,
2020; Pokrivcakova, 2019). These results are consistent with the previous findings (Zhang
and Aslan, 2021; Chiu et al., 2023).
The variable AI_P was significant at a 95% confidence level in the model (p-value < 0.05)
(table no. 5). The results indicate that learners’ perception of AI helping teachers develop
personalised learning plans increases the likelihood of AIEd by 1.651 times, holding other
factors constant. This finding aligns with previous research highlighting AI’s value in
personalising and adjusting individual learning plans (Hopcan et al., 2022; Southworth et al.,
2023).
The variable AI_SP (AI improves school performance) was a significant factor in the model
(p-value < 0.1). The probability of learners considering using AI tools in educational
activities is ambiguous as the value of the odds ratio of AI_SP is 0.668 (< 1), and the lower
limit is 0.416. In contrast, the upper limit is 1.074 (Field, 2013). These findings are in line
with previous studies that show how instrumental AI tools can positively influence the
current school performance or satisfaction of students (Steenbergen-Hu and Cooper, 2014;
Ouyang, Zheng and Jiao, 2022) and identify some gaps that have an impact on their learning
performance (Chassignol et al., 2018). Better performance can also be achieved by linking
AI to innovative assessment practices (Yang et al., 2021).
Table no. 5. Binary logistic regression results, AIEd as dependent variable
95% C.I.for
Factor B S.E. Wald df Sig. Exp(B) EXP(B)
AI_AD 0.397 0.182 4.741 1 0.029** 1.488 1.040 - 2.128
AI_DW -0.214 0.195 1.204 1 0.272 0.807 0.551 - 1.183
AI_CLASS 0.065 0.192 0.114 1 0.736 1.067 0.732 - 1.556
AI_P 0.501 0.222 5.112 1 0.024** 1.651 1.069 - 2.549
AI_DIS -0.247 0.202 1.498 1 0.221 0.781 0.526 - 1.160
AI_NI 0.157 0.179 0.768 1 0.381 1.169 0.824 - 1.660
AI_HE 0.039 0.177 0.048 1 0.827 1.039 0.735 - 1.470
AI_HR -0.261 0.189 1.903 1 0.168 0.770 0.531 - 1.116
AI_T -0.137 0.149 0.852 1 0.356 0.872 0.651 - 1.167
AI_COST -0.123 0.157 0.615 1 0.433 0.884 0.650 - 1.202
AI_ES -0.061 0.150 0.165 1 0.684 0.941 0.700 - 1.263
AI_CA 0.290 0.210 1.913 1 0.167 1.337 0.886 - 2.016
AI_BE -0.319 0.221 2.082 1 0.149 0.727 0.472 - 1.121
AI_TA -0.059 0.222 0.070 1 0.791 0.943 0.611 - 1.456
AI_SP -0.403 0.242 2.769 1 0.096* 0.668 0.416 - 1.074
AI_TIME 0.461 0.187 6.104 1 0.013** 1.586 1.100 - 2.288
AI_EMPLOY 0.238 0.215 1.226 1 0.268 1.269 0.832 - 1.934
AI_LD -0.129 0.214 0.365 1 0.546 0.879 0.578 - 1.337
Constant 0.040 0.635 0.004 1 0.949 1.041
Model summary: n=254; -2Log likelihood=270.334, Cox and Snell R-square=0.127, Nagelkerke R-Square=0.181;
Step 0: Predicted Percentage Correct 71.3%; Step 1: Predicted Percentage Correct: 73.2%; Omnibus Test: Chi-
square = 43.369, df=18, p-value=0.011; Hosmer and Lemeshow test: Chi-square = 5.143, df=8, p-value=0.742;
***Significance at the 1% level, ** at the 5% level, and *at the 10% level.

AE
The results of the model indicated that the variable AI_TIME (AI reduces the time allocated
to learning) had a significant relationship with AIEd at a 95% confidence level (p-value <
0.05). The estimate of the odds ratio for AI_TIME is 1.586, which indicates that the odds of
learners using AIEd, because they consider that AI reduces the time allocated to learning, is
1.586, more significant than the odds of not considering AI_TIME. A study conducted by
Cen, Koedinger and Junker (2007) has proven that using intelligent tutoring systems reduces
students’ learning time and, nowadays, AI intelligent tutoring systems can provide a
relatively well-rounded human-like learning experience. Furthermore, Kong (2020)
mentioned that AI shortens the learning time, and, thus, students have the capability of
learning more content within the same time frame, an affirmation in line with other findings
(Dahotre et al., 2011; Munir, Vogel and Jacobsson, 2022) showing that when compared to
using traditional training materials, students use less time for training whilst improving their
performance. AIEd requires continuous improvement to address learners’ diverse needs and
preferences (Zhang and Aslan, 2021). To achieve this, educators, researchers, and the
education community must actively participate in the AIEd innovation process, integrating
theoretical, practical, conceptual, and empirical dimensions.
Little evidence exists for educational data mining, particularly in offering speculative and
predictive analyses (Cope, Kalantzis and Searsmith, 2021). AI in education has drawbacks,
including higher costs, scalability challenges, ethical concerns, privacy issues, and limited
teacher expertise (Zhang and Aslan, 2021). Integrating AI into education depends on
government support for educational institutions, research, training, and private sector efforts
in developing AI applications (Knox, 2020) for teaching.
The second multivariate binary logistic model was developed to identify the predictors for
AI in business educational activities (the key question in the questionnaire Q2: Do you
consider AI beneficial in business educational activities?). The binary logistic regression on
AIEd_B and its predictors shows that the variables CV, EC, CBT, and ALM are the best
predictors.
The variable CV (computer vision) was found to be a significant factor in the binary
regression model at a 95% confidence level (p-value<0.05). The results indicated that the
odd ratio of CV increased the probability of AIEd_B by 2.169 times, which was in line with
the results of the previous studies (Bebis, Egbert and Shah, 2003). CV has various
applications in education activities with multiple purposes (He et al., 2017; Kusumota, Aroca
and Martins, 2018; Savov, Terzieva and Todorova, 2018) that show various benefits such as
creating an interactive educational environment (Sophocleus et al., 2021).
The variable EC (edge computing) had a significant influence on AIEd_B at a 90%
confidence level (p-value<0.10). The results showed that the probability of EC familiarity
for learners increased by 2.028 in the likelihood of AIEd_B. These findings are consistent
with the motivations of Hua et al. (2023), highlighting that, on the one hand, AI algorithms
can optimise EC, and, on the other hand, EC is an enabler for AI to bring faster response
speeds for AI applications in various other fields. Hwang and Nurtantyana (2022) emphasise
that using AI and EC can extend the education of students.
The variable CBT significantly influenced AIEd_B at a 95% confidence level (p-value
<0.05). The increase in the odds ratio of CBT is a 0.443 decrease in the odds of considering
AI beneficial for business educational activities, and the decrease could be as much as
Vol. 26 • No. 65 • February 2024 251

0.227-fold or a 0.865-fold drop. This result follows similar conclusions as Chen, Chen and
Lin (2022), showing favourable student perceptions.
The variable ALM (adaptive learning method) had a significant influence on AIEd_B at a
90% confidence level (p-value<0.1). Still, the effect of ALM on AIEd_B is not apparent, as
the value of the odds ratio is 0.528 (< 1), with the lower limit of 0.256 and the upper limit is
1.088 (Hosmer, Lemeshow and Sturdivant, 2013). The findings are in line with Muñoz et al.
(2022) that adaptive learning technology enables students to obtain instruction and practice
at a faster pace and is also suited to their capabilities. A possible negative impact may be seen
in the challenges of adaptive learning in education, such as “the lack of cognition of brain
and technology, the bottleneck of the model of emotion domain, the separation of education
and technology, the security of data management and the risk of privacy leakage” (Li, He
and Xue, 2021). Also, implementing adaptive technologies in the educational process
depends highly on the teachers’ role (Morze et al., 2021). Other studies also show the
effectiveness of adaptive learning in enabling students to become proficient in specific
content (Shelle et al., 2018).
The results of the binary logistic regression specific to IAEd_B and the dependent variables
mentioned above (table no. 6.) indicate that the significant impact is exerted by CV, EC, CBT
and ALM.
Table no. 6 Binary logistic regression results, AIEd_B as dependent variable
95% C.I.for
Factor B S.E. Wald df Sig. Exp(B) EXP(B)
CV 0.774 0.308 6.327 1 0.012** 2.169 1.186 – 3.965
PS -0.325 0.333 0.957 1 0.328 0.722 0.376 – 1.386
DM -0.375 0.302 1.534 1 0.216 0.688 0.380 – 1.244
EIS 0.195 0.322 0.366 1 0.545 1.215 0.646 - 2.285
LA 0.629 0.382 2.709 1 0.100 1.875 0.887 – 3.966
FAS -0.375 0.413 0.825 1 0.364 0.687 0.306 – 1.544
VAS 0.063 0.367 0.030 1 0.863 1.066 0.519 – 2.189
VL 0.047 0.375 0.016 1 0.900 1.048 0.503 – 2.187
AR 0.133 0.277 0.232 1 0.630 1.143 0.664 – 1.966
VR 0.398 0.360 1.224 1 0.269 1.488 0.736 – 3.011
HST -0.158 0.313 0.253 1 0.615 0.854 0.462 – 1.578
EC 0.707 0.395 3.198 1 0.074* 2.028 0.934 – 4.402
VPA -0.508 0.324 2.450 1 0.117 0.602 0.319 – 1.136
RTA -0.027 0.337 0.006 1 0.936 0.973 0.503 – 1.883
CBT -0.814 0.341 5.695 1 0.017** 0.443 0.227 – 0.865
IR 0.502 0.313 2.573 1 0.109 1.652 0.895 – 3.049
PLA 0.538 0.360 2.234 1 0.135 1.712 0.846 – 3.467
AA 0.017 0.367 0.002 1 0.962 1.018 0.496 – 2.089
ALM -0.639 0.369 2.997 1 0.083* 0.528 0.256 – 1.088
Constant -0.221 0.814 0.074 1 0.786 0.801
Model summary: n=254; -2Log likelihood=133.122, Cox and Snell R-square=0.142, Nagelkerke R-
Square=0.289; Step 0: Predicted Percentage Correct 89.4%; Step 1: Predicted Percentage Correct:
90.2%; Omnibus Test: Chi-square = 38.941, df=19, p-value=0.004; Hosmer and Lemeshow test: Chi-
square = 12.550, df=8, p-value=0.128; ***Significance at the 1% level, ** at the 5% level, and *at the
10% level.

AE
Table no. 5 and table no. 6 present the Wald test values, which are used to assess the statistical
significance of each coefficient in the regression models. The obtained value for the Wald
test is used to test the null hypothesis that the associated coefficient is zero. A higher value
of the Wald test indicates a significant influence of the independent variable on the dependent
variable. Additionally, for each coefficient, the associated probability value with the Wald
test is important, as lower probability values suggest rejecting the null hypothesis and
considering the coefficient as different from zero.
Conclusions
This study empirically investigated the triggers and AI tools fostering AI benefits in business
educational activities. Research among students revealed their limited familiarity with AI's
benefits for education, despite 87.8% having used AI previously. Only 71.3% of respondents
used AIEd, indicating little awareness of AI tools and their benefits. Additionally, 46.8% of
learners declared being unprepared to embrace AI in business education, although they were
more prepared than teachers for AIEd_B (only 27.2% of the teachers were prepared). The
authors employed multivariate binary logistic regressions to respond to two research
questions. For the first question, if AI supports educational activities, the results of the study
revealed that the respondents used AI mainly because it performs administrative tasks instead
of teachers; AI helps teachers develop personalised learning plans for each student; AI
improves school performance, and AI reduces the time allocated to learning. The second
research question focused on AI tools used in AIEd, revealing limited learner knowledge and
practice regarding CV, PS, DM, EIS, LA, FAS, VAS, VL, AR, VR, HST, EC, VPA, RTA,
CBT, IR, PLA, AA, and ALM in business education.
The questionnaire used in the study is an original tool, and the analysis focuses on a specific
group of subjects and a unique context that has not been extensively covered in the existing
literature. This research contributes to existing knowledge by highlighting unique aspects
that enrich the understanding in the analyzed field.
The paper highlights the limited degree of awareness regarding the benefits of using AI in
business education, even though respondents have used AI in various activities. Additionally,
the lack of preparedness among respondents to effectively adopt AI in the educational process
has been emphasized. The study contributes to existing literature through empirical
investigation of AI determining factors and tools in the context of business education,
underscoring the need for continuous improvements in the implementation of AI in this
domain.
A limitation of this study refers to the sample, its size and level of knowledge. It considers
the students’ opinions from Romania without the possibility of expanding the analysis to
other situations encountered in different universities. Future research should use larger and
more diverse samples to enhance the generalizability of results to a significant population.
Additionally, it is important to consider the impact of AI on teachers and administrative
components in education, and future research should focus on exploring these aspects.
Throughout this investigation, binary logistic regression models were employed,
acknowledging their limitations in capturing the entire complexity of the studied
phenomenon. Looking ahead to future research, we consider transitioning to a Generalized
Ordered Logit model beneficial, as it allows for a more detailed and precise approach to the
relationships between variables. Through this model, we intend to thoroughly explore the
Vol. 26 • No. 65 • February 2024 253

influence of each explanatory variable on different levels of the dependent variable,

highlighting consistent contributions across various value tiers. The study may appear timely
based, viewing the students' understanding of the field and the limited use of AI in the
Romanian educational system for business students. It may fail to present a holistic and
deepened view of the topic.
References
Albert, B., Tullis, T. and Tedesco, D., 2010. Beyond the Usability Lab: Conducting Large-
scale Online User Experience Studies. Burlington: Morgan Kaufmann.
Al Braiki, B., Harous, S., Zaki, N. and Alnajjar, F., 2020. Artificial intelligence in education
and assessment methods. Bulletin of Electrical Engineering and Informatics, [e-journal]
9(5), pp. 1998-2007. https://doi.org/10.11591/eei.v9i5.1984.
al-Zyoud, H.M.M., 2020. The Role of Artificial Intelligence in Teacher Professional
Development. Universal Journal of Educational Research, [e-journal] 8(11B),
pp. 6263-6272. https://doi.org/10.13189/ujer.2020.082265.
Antonenko, P. and Abramowitz, B., 2023. In-service teachers’ (mis)conceptions of artificial
intelligence in K-12 science education. Journal of Research on Technology in Education,
[e-journal] 55(1), pp. 64-78. https://doi.org/10.1080/15391523.2022.2119450.
Asakura, K., Occhiuto, K., Todd, S., Leithead, C. and Clapperton, R., 2020. A Call to Action
on Artificial Intelligence and Social Work Education: Lessons Learned from A
Simulation Project Using Natural Language Processing. Journal of Teaching in Social
Work, [e-journal] 40(5), pp. 501-518. https://doi.org/10.1080/08841233.2020.1813234.
Bangdiwala, S., 2018. Regression: binary logistic. International Journal of Injury Control
and Safety Promotion, [e-journal] 25(3), pp. 336-338. https://doi.org/10.1080/
17457300.2018.1486503.
Baxter, K., Courage, C. and Caine, K., 2015. Choosing a User Experience Research Activity.
In: K. Baxter, C. Courage and K. Caine, eds. 2015. Interactive Technologies,
Understanding your Users. 2nd ed. S.l.: Morgan Kaufmann, pp. 96-112.
https://doi.org/10.1016/B978-0-12-800232-2.00005-5.
Bebis, G., Egbert, D. and Mubarak, S., 2003. Review of computer vision education. IEEE
Transactions on Education, [e-journal] 46(1), pp. 2-21. https://doi.org/10.1109/
te.2002.808280.
Cen, H., Koedinger, K.R. and Junker, B. 2007. Is Over Practice Necessary? - Improving
Learning Efficiency with the Cognitive Tutor through Educational Data Mining. In: R.
Luckin, K.R. Koedinger and J.E. Greer, eds. 2007. Frontiers in Artificial Intelligence and
Applications. Amsterdam: IOS Press, pp. 511-518.
Intelligence trends in education: a narrative overview. Procedia Computer Science,
[e-journal] 136, pp. 16-24. https://doi.org/10.1016/j.procs.2018.08.233.
Access, [e-journal] 8, pp. 75264-75278. https://doi.org/10.1109/ACCESS.2020.2988510.

AE
Chiu, T., Xia, Q., Zhou, X., Chai, C.S. and Cheng, M., 2023. Systematic literature review on
opportunities, challenges, and future research recommendations of artificial intelligence
in education. Computers and Education: Artificial Intelligence, [e-journal] 4, article no.
Cope, B., Kalantzis, M. and Searsmith, D., 2021. Artificial intelligence for education:
Knowledge and its assessment in AI-enabled learning ecologies. Educational Philosophy
and Theory, [e-journal] 53(12), pp. 1229-1245. https://doi.org/10.1080/00131857.
2020.1728732.
Crescenzi-Lanna, L., 2023. Literature review of the reciprocal value of artificial and human
intelligence in early childhood education. Journal of Research on Technology in
Education, [e-journal] 55(1), pp. 21-33. https://doi.org/10.1080/15391523.
2022.2128480.
Crompton, H., Jones, M. and Burke, D., 2022. Affordances and challenges of artificial
intelligence in K-12 education: a systematic review. Journal of Research on Technology
in Education, [e-journal]. https://doi.org/10.1080/15391523.2022.2121344.
Dahotre, A., Krishnamoorthy, V., Corley, M. and Scaffidi, C., 2011. Using Intelligent Tutors
to Enhance Student Learning of Application Programming Interfaces. Journal of
Computing Sciences in Colleges, [e-journal] 27, pp. 195-201. https://dl.acm.org/doi/abs/
10.5555/2037151.2037190.
Das, S., Dey, A., Pal, A. and Roy, N., 2015. Applications of artificial intelligence in machine
learning: review and prospect. International Journal of Computer Applications,
[e-journal] 115(9), pp. 31-41. https://doi.org/10.5120/20182-2402.
Edgar, T. and Manz, D., 2017. Exploratory Study. In: T.W. Edgar and D.O. Manz, eds. 2017.
Research Methods for Cyber Security. S.l.: Syngress, pp. 95-130.
https://doi.org/10.1016/B978-0-12-805349-2.00004-2.
Elhajjar, X., Karam, S. and Borna, S., 2021. Artificial Intelligence in Marketing Education
Programs. Marketing Education Review, [e-journal] 31(1), pp. 2-13.
https://doi.org/10.1080/10528008.2020.1835492.
Field, A., 2013. Discovering Statistics Using IBM SPSS Statistics: (and Sex and Drugs and
Rock ’n’ Roll). 4th ed. Los Angeles, London, New Delhi: Sage.
Han, J.W., Park, J. and Lee, H., 2022. Analysis of the effect of an artificial intelligence
chatbot educational program on non-face-to-face classes: a quasi-experimental study.
BMC Medical Education, [e-journal] 22, article no. 830, pp. 1-10.
https://doi.org/10.1186/s12909-022-03898-3.
He, B., Xia, M., Yu, X., Jian, P., Meng, H. and Chen, Z., 2017. An educational robot system
of visual question answering for preschoolers. In: s.n., The 2017 2nd International
Conference on Robotics and Automation Engineering (ICRAE). 29-31 December 2017,
Shanghai, China. S.l.: s.n.
Hopcan, S., Polat, E., Ozturk, M.E. and Ozturk, L., 2022. Artificial intelligence in special
education: a systematic review. Interactive Learning Environments, [e-journal].
https://doi.org/10.1080/10494820.2022.2067186.
Hosmer, D.W., Lemeshow, S. and Sturdivant, R., 2013. Applied logistic regression. 3rd ed.
New York, NY: Wiley.
Vol. 26 • No. 65 • February 2024 255

Hua, H., Li, Y., Wang, T., Dong, N., Li, W. and Cao, J., 2023. Edge Computing with
Artificial Intelligence: A Machine Learning Perspective. ACM Computing Surveys,
Hwang, W.-Y. and Nurtantyana, R., 2022. X-Education: Education of All Things with AI
and Edge Computing - One Case Study for EFL Learning. Sustainability, 14(19), article
no. 12533, pp. 1-23. https://doi.org/10.3390/su141912533.
Knox, J., 2020. Artificial intelligence and education in China. Learning, Media and
Technology, [e-journal] 45(3), pp. 298-311. https://doi.org/10.1080/17439884.2020.
1754236.
Khosravi, H., Shum, S.B., Chen, G., Conati, C., Tsai, Y.S., Kay, J., Knight, S., Martinez-
Maldonado, R., Sadiq, S. and Gasevic, D., 2022. Explainable Artificial Intelligence in
education. Computers and Education: Artificial Intelligence, 3, article no. 100074,
pp. 1-22. https://doi.org/10.1016/j.caeai.2022.100074.
Kong, F., 2020. Application of Artificial Intelligence in Modern Art Teaching. International
Journal of Emerging Technologies in Learning, [e-journal] 15(13), pp. 238-251.
Kusumota, V., Aroca, R. and Martins, F., 2018. An Open Source Framework for Educational
Applications Using Cozmo Mobile Robot. In: s.n. The 2018 Latin American Robotic
Symposium, 2018 Brazilian Symposium on Robotics (SBR) and 2018 Workshop on
Robotics in Education (WRE). 6-10 November 2018, João Pessoa, Brazil. S.l.: s.n.
Ley, T., Tammets, K., Pishtari, G., Chejara, P., Kasepalu, R., Khalil, M., Saar, M., Tuvi, I.,
Väljataga, T. and Wasson, B., 2023. Towards a partnership of teachers and intelligent
learning technology: A systematic literature review of model‐ based learning analytics.
Journal of Computer Assisted Learning, [e-journal] pp. 1397-1417. https://doi.org/
10.1111/jcal.12844.
Li, F., He, Y. and Xue, Q., 2021. Progress, Challenges and Countermeasures of Adaptive
Learning: A Systematic Review. Educational Technology and Society, [e-journal] 24(3),
pp. 238-255. https://www.jstor.org/stable/27032868.
Li, P. and Wang, B., 2023. Artificial Intelligence in Music Education. International Journal
of Human-Computer Interaction, [e-journal] https://doi.org/10.1080/10447318.2023.
2209984.
Matzavela, V. and Alepis, E., 2021. Decision tree learning through a Predictive Model for
Student Academic Performance in Intelligent M-Learning Environments. Computers and
Education: Artificial Intelligence, [e-journal] 2, article no. 100035, pp. 1-12.
McGrath, C., Pargman, T.C., Juth, N. and Palmgren, P., 2023. University teachers’
perceptions of responsibility and artificial intelligence in higher education – An
experimental philosophical study. Computers and Education: Artificial Intelligence,
[e-journal] 4, article no. 100139, pp. 1-9. https://doi.org/10.1016/j.caeai.2023.100139.
Morze, N., Varchenko-Trotsenko, L., Terletska, T. and Smyrnova-Trybulska, E., 2021.
Implementation of adaptive learning at higher education institutions by means of Moodle
LMS. Journal of Physics: Conference Series, [e-journal] 1840, article no. 012062,
pp. 1-13. https://doi.org/10.1088/1742-6596/1840/1/012062.

AE
Munir, H., Vogel, B. and Jacobsson, A., 2022. Artificial Intelligence and Machine Learning
Approaches in Digital Education: A Systematic Revision. Information, [e-journal] 13(4),
article no. 203, pp. 1-26. https://doi.org/10.3390/info13040203.
Muñoz, J.L.R., Ojeda, F.M., Jurado, D.L.A., Peña, P.F.P., Carranza, C.P.M., Berríos, H.Q.,
Molina, S.U., Farfan, A.R.M., Arias-Gonzáles, J.J. and Vasquez-Pauca, M.J., 2022.
Systematic review of adaptive learning technology for learning in higher education.
Eurasian Journal of Educational Research, [e-journal] 98(98), pp. 221-233.
https://doi.org/10.14689/ejer.2022.98.014.
Nuseir, M., Basheer, M.F. and Aljumah, A., 2020. Antecedents of entrepreneurial intentions
in smart city of Neom Saudi Arabia: Does the entrepreneurial education on artificial
intelligence matter?. Cogent Business and Management, [e-journal] 7(1), pp. 1-16.
https://doi.org/10.1080/23311975.2020.1825041.
Ouyang, F. and Jiao, P., 2021. Artificial intelligence in education: The three paradigms.
Computers and Education: Artificial Intelligence, [e-journal] 2, article no. 100020,
Ouyang, F., Zheng, L. and Jiao, P., 2022. Artificial intelligence in online higher education: A
systematic review of empirical research from 2011 to 2020. Education and Information
Technologies, [e-journal] 27, pp. 7893-7925. https://doi.org/10.1007/s10639-022-10925-9.
Pallant, J., 2005. SPSS Survival Manual Second-Edition: A Step by Step Guide to Data
Analysis Using SPSS. Sydney: Allen and Unwin.
Pokrivcakova, S., 2019. Preparing teachers for the application of AI-powered technologies
in foreign language education. Journal of Language and Cultural Education, [e-journal]
7(3), pp.135-153. https://doi.org/10.2478/jolace-2019-0025.
Rahman, H., Zafri, N.M., Akter, T. and Pervaz, S., 2021. Identification of factors influencing
severity of motorcycle crashes in Dhaka, Bangladesh using binary logistic regression
model. International Journal of Injury Control and Safety Promotion, [e-journal]
pp. 1-12. https://doi.org/10.1080/17457300.2021.1878230.
Savov, T., Terzieva, V. and Todorova, K., 2018. Computer vision and Internet of Things:
Attention System in educational context. In: s.n, The 19th International Conference on
Computer Systems and Technologies - CompSysTech’18. September 2018, s.l. [online]
s.l.: s.n. Available at: <https://doi.org/10.1145/3274005.3274014> [Accessed 21
December 2023].
Shelle, G., Earnesty, D., Pilkenton, A. and Powell, E., 2018. Adaptive Learning: An
Innovative Method for Online Teaching and Learning. The Journal of Extension,
[e-journal] 56(5), pp. 1-8. https://doi.org/10.34068/joe.56.05.17.
Skavronskaya, L., Hadinejad, A. and Cotterell, D., 2023. Reversing the threat of artificial
intelligence to opportunity: a discussion of ChatGPT in tourism education. Journal of
Teaching in Travel and Tourism, [e-journal] 23(2), pp. 253-258.
https://doi.org/10.1080/15313220.2023.2196658.
Sophokleous, A, Christodoulou, P, Doitsidis, L. and Chatzichristofis, S.A., 2021. Computer
Vision Meets Educational Robotics. Electronics, [e-journal] 10(6), pp. 1-24.
https://doi.org/10.3390/electronics10060730.
Vol. 26 • No. 65 • February 2024 257

Southworth, J., Migliaccio, K., Glover, J., Glover, J., Reed, D., McCarty, C., Brendemuhl, J.
and Thomas, A., 2023. Developing a model for AI Across the curriculum: Transforming
the higher education landscape via innovation in AI literacy. Computers and Education:
Artificial Intelligence, [e-journal] 4, article no. 100127, pp. 1-10. https://doi.org/10.1016/
j.caeai.2023.100127.
Steenbergen-Hu, S. and Cooper, H., 2014. A Meta-Analysis of the Effectiveness of
Intelligent Tutoring Systems on College Students’ Academic Learning. Journal of
Educational Psychology, [e-journal] 106(2), pp. 331-347,
https://doi.org/10.1037/a0034752331.
Tan, C., 2020. Digital Confucius? Exploring the implications of artificial intelligence in
spiritual education. Connection Science, [e-journal] 32(3), pp. 280-291.
https://doi.org/10.1080/09540091.2019.1709045.
Toumi, K., 2020. Islamic ethics, capital structure and profitability of banks; what makes
Islamic banks different?. International Journal of Islamic and Middle Eastern Finance
and Management, [e-journal] 13(1), pp. 116-134. https://doi.org/10.1108/IMEFM-05-
2016-0061.
Xu, W., Zhao, Y., Nian, S., Feng, L., Bai, X., Luo, X. and Luo, F., 2018. Differential analysis
of disease risk assessment using binary logistic regression with different analysis
strategies. Journal of International Medical Research, [e-journal] 46(9), pp. 3656-3664.
https://doi.org/10.1177/0300060518777173.
Yang, S.J.H., Ogata, H., Matsui, T. and Chen, N.S., 2021. Human-centered artificial
intelligence in education: Seeing the invisible through the visible. Computers and
Education: Artificial Intelligence, [e-journal] 2, article no. 100008, pp. 1-5.
Yang, D., Zhao, W., Du, J. and Yang, Y., 2022. Approaching Artificial Intelligence in
business and economics research: a bibliometric panorama (1966–2020). Technology
Analysis and Strategic Management, [e-journal] pp. 1-16. https://doi.org/10.1080/
09537325.2022.2043268.
Zhang, K. and Aslan, A.B., 2021. AI technologies for education: Recent research and future
directions. Computers and Education: Artificial Intelligence, 2, article no. 100025,

AE
QUANTITATIVE EVALUATION OF WILLINGNESS

TO USE ARTIFICIAL INTELLIGENCE WITHIN BUSINESS
AND ECONOMIC ACADEMIC ENVIRONMENT
Daniela Șerban1 , Silvia Elena Cristache2 , Narcisa Georgeta Ciobotar3 ,
Laurențiu Gabriel Frâncu4 and Jiries Mansour5
1)2)3)4)5)

Șerban, D., Cristache, S.E., Ciobotar, N.G., Frâncu, Received: 29 September 2023
L.G., and Mansour, J., 2024. Quantitative Evaluation Revised: 25 November 2023
of Willingness to Use Artificial Intelligence within Accepted: 15 December 2023
Business and Economic Academic Environment.
Abstract
The article analyses the state of awareness, use, and willingness to use artificial intelligence
(AI) in business and economic universities, according to Romanian academics. It is also
highlighting the main consequence of AI use in economic and business university education,
with the aim of identifying an appropriate framework for the regulated implementation of AI
systems in economic universities in Romania. The study aims to identify the advantages,
disadvantages, and the willingness to use AI on the teachers’ personal initiative in research,
teaching, and evaluation activities. The method of analysis used is quantitative, by managing
an online questionnaire to which Romanian academic teachers familiar with AI in education
responded. Data processing is carried out using Smart PLS, which allows the identification
of statistical relationships guiding the perspectives of the use of AI in economic education in
Romania. The sample represents a normal volume pilot sample. The results of the study are
useful because they identify aspects that can optimise the research and education processes,
as well as teaching, evaluation, and learning, to meet the increasing dynamics of AI use in
the economic academic environment in Romania. The academics’ views on the advantages
associated with the use of AI systems and their proposed solutions to maximise the
advantages of AI use in research, teaching, and evaluation activities are also highlighted.
All of this contributes to the development of a framework for the implementation of AI
systems in the economic and business education in Romania. Results indicate an early stage
of AI use and integration in the activities of academics from the analysed universities: AI is
predominantly used for the evaluation of students, which can be done automatically.
The availability of academics to use AI in teaching and research is low.

Corresponding author, Narcisa Ciobotar – e-mail: [email protected]
Vol. 26 • No. 65 • February 2024 259

AE Quantitative Evaluation of Willingness to Use Artificial Intelligence
within Business and Economic Academic Environment
Keywords: Artificial intelligence, AI advantages, AI disadvantages, education, university,

research, teaching, evaluation
JEL Classification: F02, F63, M21, M38, O33
Introduction
The concept of artificial intelligence (AI) has evolved over time from an abstract idea to an
ubiquitous reality in our society. At each development stage, researchers and developers have
contributed to the development of AI technologies (Rosenblatt, 1958; Rumelhart et al., 1994),
transforming the way we interact with technology and the way we approach complex
problems. The use of artificial intelligence and its impact on society have become topics of
increasing discussion and research in recent years (Brynjolfsson and McAfee, 2017).
Baker and Smith (2019) provide a general definition of the artificial intelligence: “Computers
that perform cognitive tasks usually associated with human minds, especially learning and
problem solving”.
Artificial intelligence involves automated communications with students, automated
handwriting evaluation, machine performing statistics, and a common application of
unsupervised learning (Gillani et al., 2023).
Today, AI is used in a variety of fields, from medicine (Topol, 2019) to transportation with
electric cars such as Waymo and Tesla. AI is also found in virtual assistance that relates to
Learning Management Systems (LMS) such as Alexa or Siri and Big Data analytics (Marr,
2015). In education, the main AI systems and applications that generate value and improve
the professional lives of users in the academic environment relate to LMSs such as Moodle,
Canvas or Blackboard, learning content recommendation systems, automated assessment
systems, virtual tutorials, and AI assistants. As far as the influence of AI in education is
concerned, it has two aspects: the educational process, support and changes in pedagogy, and
the routine function of the teacher and the educational sphere and content. AI applications in
education identify the potential for major advantages that it brings to the field, such as
improvements in the learning process, overall access to quality education, improvements in
learning outcomes for students, especially for disadvantaged students, and facilitating
teacher-student collaboration through analytics (Kamalov et al., 2023).
AI can bring advantages for student testing and evaluation by creating tests and assessments
at levels of difficulty based on each student’s knowledge. AI also helps in administrative
activities that are carried out regularly in education, such as plagiarism checking, scheduling
of courses and seminars, preparation of curricula, management of students and teachers’ data,
and evaluation of teachers’ research work (Khosravi et al., 2022).
This paper analyses the state of awareness, use, and willingness of artificial intelligence use
in the Romanian business and economic university environment, from the perspective of
teachers. The conflicting issue is the challenges associated with the use of AI systems and
the proposed solutions to maximise the benefits of AI in research, teaching, and evaluation
activities.

AE
Given the interest in maximising the advantages and minimising the risks of using AI systems
and applications in academic research, teaching, and evaluation, a survey of academics in
Romanian business and economics universities was conducted.
The purpose of the survey was to evaluate the effectiveness of teachers’ use of artificial
intelligence, of their own initiative, in academic research, teaching, and assessment. The
objectives of the survey were:
 To assess teachers’ awareness of AI;
 To assess awareness of the advantages and disadvantages of using AI;
 To assess the willingness and availability of university teachers to use AI;
 To determine which academic areas can benefit from the use of AI;
 To identify ways of using AI in research, evaluation, and teaching.
Based on the assumption that an individual’s decision to use or not to use AI is the result of
an analysis that balances the reasons, the perceived advantages and disadvantages associated
with AI, the willingness of teachers to use AI in education and its effects on the degree of
acceptance of AI among respondents were examined.
The paper is structured into four sections. The first section presents the literature review,
presenting relevant results obtained over time in the study of the benefits, advantages, and
contributions that determine the interest in using and/ or not using AI. The second section
presents the research methodology with information on the objectives of the study, the
operationalisation of the data collection tool, the sample used, and the registered data analysis
methods. The third section presents the results of the analysis of the responses and
discussions to identify possibilities for further research. The last section presents limitations
of the research and conclusions.

1.1. University business and economics education in the context of the use of artificial
intelligence. Advantages and disadvantages
Artificial intelligence is a driven force in the evolution of contemporary society. This
technological and information revolution has been underpinned by significant developments
in computing and machine learning algorithms. Virtual assistants have become an integral
part of our lives (Jordan, 2018). In academic economic and business education, AI can
accelerate economic research, used as a tool to rapidly identify, and analyse information from
extensive resources (Bessen et al., 2020), with positive impact on innovation and economic
development.
The main advantage of using AI in the university environment is related to the customisation
of learning content and teaching methods in economic education, adapting curricula to the
level of knowledge and learning interest of students. AI contributes to the understanding of
complex economic concepts through the analysis of economic and financial data, ensuring
the development of forecasts and predictive models (Makridakis et al., 2018). Alongside the
research, teaching, and learning process, AI ensures that the evaluation of students’ process
is improved. AI can collect and analyse student performance data to identify weaknesses and
Vol. 26 • No. 65 • February 2024 261

provide recommendations for improvement of student performance. In order to promote

academic integrity, AI systems can identify possible cases of plagiarism in students’ work by
comparing with existing sources (Kakkonen et al., 2019). Rodríguez and Brito (2017) present
a study on the opportunities and threats of artificial intelligence in higher education: teaching-
learning, spaces, objects, methodologies, assessment are some of the aspects transformed by
AI. According to Rodríguez and Brito (2017), the advantages and disadvantages associated
with the use of AI in economic and business education, especially related to data security,
privacy, etc., should not be ignored.
The disadvantages of the safe use of artificial intelligence systems in business and economics
education can affect not only the data, but also the integrity of the educational process and
educational institutions. In this context, AI integrity requires a functional and trusted
framework in Romanian universities that supports infrastructure changes and a significant
amount of digital equipment (Pisica et al., 2023).
The use of AI often involves the collection and storage of significant amounts of sensitive
data, such as financial or personal information from students and teachers. There is a risk that
this data may be exposed to privacy breaches, putting sensitive information at risk (Perna et
al., 2020). A significant example is the COVID-19 pandemic, which had a strong negative
impact on the Romanian university system. The shift to online education has also generated
serious privacy and security issues (Sharma et al., 2022; Pop et al., 2022).
In the context of business and economics education, AI can be used for the rapid generation
and analysis of data and information. The high costs and limited accessibility of AI in
business and economics education can lead to digital exclusion of some students or
professionals, exacerbating inequalities in access to education and opportunities (Selwyn,
2020).
1.2. Regulation of the use of artificial intelligence in academia. Opinions supporting or

not supporting the regulation of artificial intelligence in academia
One issue of global importance is the regulation on how to use the AI technology. The debate
highlights both pro and con views on regulating the use of AI in general and, specifically, in
university business and economics education. Setting ethical and accountability standards for
educational institutions can prevent, in the opinion of academics, the misuse or
discriminatory use of AI in student assessment or learning (Johnson, 2020).
Over-regulation of AI systems can bring with it a myriad of technical and bureaucratic
requirements, which can negatively affect the student experience and the performance of
educational institutions, (Johnson, 2020). It is important to strike a balance between
protecting the rights and safety of students, teachers, and other users and promoting
technological innovation and development in higher education in Romania. In order to
maximise the benefits of AI and minimise its negative impact, it is essential to continue
responsible research, development, and regulation of this technology (Smith, 2019).
Humankind’s efforts are predicted to continue to develop technology, automation and
artificial intelligence more strongly in the future (Khogali and Mekid, 2023).
The European Union has adopted a number of regulations and documents that apply to the
use of AI in various fields, including education. The most representative documents refer to
the Report on Data Protection and Privacy in AI (European Parliament, 2023b), Regulation

AE
on Artificial Intelligence (European Parliament, 2023a), Directive on Copyright in the Digital

Single Market (European Commission, 2023c). The European Union is developing a
regulatory framework for AI in education, including ethical and technical standards for the
assessment and certification of AI technologies (European Commission, 2023a).
Governments have the role of establishing the legal and regulatory framework for the use of
AI in education. This can ensure transparency, accountability, and respect for the rights of
students and teachers (European Commission, 2023b). Organisations and institutions
contribute to the regulation of AI, such as the IEEE Standards Association developing
Technical Standards for Artificial Intelligence and AI robots (IEEE Standards Association,
2023). European authorities have developed documents on data protection and fundamental
rights of AI users to assess the positive and negative implications of new technologies,
including AI and big data, on fundamental rights (European Union Agency for Fundamental
Rights, 2021).
Users AI also have a key role to play in developing the regulatory framework for artificial
intelligence. Through their actions on accepting or rejecting the use of AI systems and
providing feedback, users can help ensure that AI is developed and used responsibly,
ethically, and for the benefit of society (Brynjolfsson and McAfee, 2018). Teacher users can
contribute to improving AI systems by providing feedback on their experience and reporting
issues such as bias or discrimination (Pisica et al., 2023; Chan, 2023). Users of AI, including
students and teachers, can also request access to information on how AI influences the
educational process and demand transparency in decisions made by AI systems (Pisica et al.,
2023; Chan, 2023).
The main objective of this quantitative research is to determine how the advantages and
disadvantages of using artificial intelligence affect teachers’ willingness to use AI. There are
analysed relationships between:
 The advantages and disadvantages of AI as seen by teachers and
 The use of AI in operational, day-to-day research, teaching and evaluation activities
carried out by academics in business and economic universities.
As described in the literature review, artificial intelligence has a wide range of uses in
universities (Chan, 2023; Pisica et al., 2023), with consequences that may lead to favourable
opinions of academics due to its advantages and unfavourable opinions due to the current
stage of development of artificial intelligence (Kamalov et al., 2023). Based on the
advantages and disadvantages identified in the literature as associated with the use of AI in
the academic environment, we formulated the following research hypotheses:
H1: There is an incipient level of AI use in the Romanian business and economic academic
environment.
H2: The advantages perceived by academics as a result of using AI negatively influence their
willingness to use AI in teaching (H2a), research (H2b), and student assessment (H2c).
H3: Disadvantages perceived by academics as a result of using AI negatively influence their
willingness to use AI in teaching (H3a), research (H3b), and student assessment (H3c).
Vol. 26 • No. 65 • February 2024 263

H4: Disadvantages associated with the current stage of AI development negatively influence
the willingness to use AI in teaching (H4a), research (H4b), and student assessment (H4c).
H5: Academics see AI disadvantages as constraints on its use for assessment in academic
settings.
To empirically validate the hypotheses, quantitative research was carried out in which
teachers were invited to answer questions on their perceived usage habits, advantages, and
disadvantages of using AI on their own initiative in a non-formal academic setting. Data were
collected using a questionnaire-based survey, conducted online between August and
September 2023.
The questionnaire was completed by 101 university teachers who had used AI on their own
initiative, representing two-thirds of the total questionnaire population. There is an equal
distribution by age group: the 35 - 54 age segment represents 70% of the total sample. The
structure of the sample by age group corresponds to the structure published by the Ministry
of Education in 2022 for the academic year 2021-2022: 68.29% of teachers aged 35-54
(Ministry of Education, 2022). All types of teaching positions are included, and the gender
distribution is balanced: 58% women and 42% men. The structure of the sample by gender
shows its representativeness (Ministry of Education, 2022), despite the small size, which is
mainly caused by the novelty of the subject. 101 valid responses were obtained, and 23
responses were removed because of the repetitiveness of the scales chosen by the
respondents. The form did not allow the recording of non-responses, nor did it allow moving
to the next question without answering the previous question.
The questionnaire included direct and disguised factual and opinion questions on the state of
knowledge, information, awareness, application of AI software in the economic and business
university environment and the perspective and willingness of its use in research, teaching,
and evaluation activities, in the opinion of the teaching staff of economic universities in
Romania. The questionnaire has more than 100 items, quantitative and qualitative, of which
48 quantitative items are analysed because they respond to the objective of the paper and
allow the validation of stochastic relationships described by the research hypotheses. The
items included in this quantitative analysis are measured using a five-point Likert-type scale
(score 1 given for the opinion “Strongly disagree” and 5 for “Strongly agree”). The second
Likert scale used also had five points (score 1 given for the answer “Not at all” and 5 for the
answer “Very strongly”). The other qualitative items were not included in the modelling.
The reliability of the data was achieved using factor analysis, while relationships between
constructs were empirically verified using structural equation modelling. The choice of
exploratory variables was also based on the non-parametric correlation matrix using
Spearman coefficients. Interpretations were performed based on the outputs of the Smart-
PLS 4.0 software (Ringle et al., 2015), using structural equation models, bootstrapping based
on 5,000 distinct samples, which allows these samples to be generated using a relatively low
volume response for the survey sample (Khan et al., 2019).

AE

3.1. Results of the Factorial Analysis
In order to validate the items included in the questionnaire that are part of the models and to
determine the exploratory and dependent variables, a confirmatory factor analysis based on
principal components was performed. A number of six factors was established to meet the
need to confirm the research hypotheses. Of the six factors, three relate to characteristics of
AI use and three factors relate to respondents’ willingness to use AI for teaching, research,
and evaluation respectively.
If we analyse the first factor (F1), it contains items about the advantages of using artificial
intelligence, as they appear in the opinion of respondents who have used this new
technological product. The advantages measured by the respondents on a scale from 1 to 5
refer to the use out of curiosity, which is an advantage of researchers, curiosity leads them to
try new technologies. This item has a loading factor of 0.760 and is complemented in the
construct by accessibility of information (0.850), ease of use of AI products/services (0.835),
confidence in AI-generated information (0.808), speed of obtaining information (0.858),
documentation (0.872), scientific comparisons (0.872), willingness to contribute to AI
training (0.722) and evaluation of AI solutions in relation to human intelligence solutions
(0.864). The validity of factor 1 is given by the Cronbach-Alpha value of 0.942.
The second factor (F2) refers to disadvantages related to the use of AI in the opinion of the
responding teachers and contains items about the inaccessibility of AI due to the current stage
of development and limitations of AI in certain fields (0.854), the difficulty of using AI in
academic activities (0.921), and the lack of formal training for generating questions to AI
(prompts) (0.858), and its validity is given by the Cronbach-Alpha value of 0.854.
The third factor (F3) contains items that relate to the stage of use of AI and contains six items
about the uncertainty of teachers (0.715), distrust of AI-generated results or different AI-
generated answers for the same question (prompt), if the question is asked differently (prompt
is different) (0.744), protection of data and personal information (0, 724), the willingness not
to contribute to AI training for fear of the risk of displacement or comparison by students of
the level of the teacher with the scientific level of the AI (0.693), lack of legislative regulation
(0.905), and non-authenticity of the information provided by the AI (0.883). The validity of
this factor is given by the Cronbach-Alpha value of 0.891.
The fourth factor (F4), which opens the items proposed by the authors, refers to the
willingness and intention to use AI in teaching on one’s own initiative. This factor contains
four items about the willingness to use AI in academic teaching activities (0.934), intention
to use AI in academic teaching activities in the short term (0.894), intention to use AI in
academic teaching activities in the long term (0.906), and advantages of using AI in academic
teaching activities (0.904) and has a good validity given by the Cronbach-Alpha value of
0.931.
The fifth factor (F5) contains four items, proposed by the authors, and describes the intention
and willingness to use AI in research. The items in F4 are about the intention to use artificial
intelligence in academic research activities in general (0.946), the short-term intention to use
artificial intelligence academic research activities (0.900), the long-term intention to use
artificial intelligence in research activities (0.948) and the advantages of using artificial
intelligence in academic research activities (0.919), activities that involve an extremely high
Vol. 26 • No. 65 • February 2024 265

level of creativity and innovation. The construct validity is given by the Cronbach-Alpha
value of 0.947.
The sixth factor (F6) is also the authors’ original contribution and contains four items relating
to willingness and intention to use AI in student evaluation, lower level of innovation activity,
and higher degree of standardisation and repetitiveness. This factor contains items about the
readiness to use AI in academic assessment activities (0.966), which are automated activities
with a low level of creativity, the short-term intention to use AI in activities of evaluation of
students (0.942), the long-term intention to use AI in evaluation of students’ activities
(0.959), and the advantages of using AI in activities of students’ evaluation, in the opinion of
the respondents (0.966). This factor is also reliable, with a Cronbach-Alpha value of 0.970.
The reliability of the data is given by Cronbach-Alpha values > 0.700 for each construct,
showing the validity of the responses.
According to Table no. 1 that illustrates the confirmatory factor analysis (Joreskog et al.,
2016), the CR reliability coefficient > 0.700 for each construct shows a high degree of
confidence in the recorded responses. The average of the variance extracted, AVE > 0.500
shows that the items explain fewer errors than the variance in the construct. The results were
interpreted using Smart-PLS 4.0 (Ringle et al., 2015) and allowed the identification of factors
influencing the intention to use artificial intelligence in the short and long term by academics,
who are knowledgeable and have already used artificial intelligence in research, teaching and
evaluation activities. The correlations identified by the structural equation model are
consistent with the non-parametric correlations significant for a 1% significance level
(implying a 99% confidence level) analysed to confirm the determinants of academics’
attitudes towards the issue of the willingness of AI use in research, teaching and evaluation.
Table no. 1. Items grouped by factors extracted by Confirmatory Factor Analysis
and the corresponding loadings factors
Item CFA
F1. Advantages of using AI, items adapted from Pisica et al., (2023)
(Cronbach-Alpha = 0.942 > 0.7; CR = 0.951 > 0.7; AVE = 0.686 > 0.5)
Curiosity 0.760
Information accesibility 0.850
AI product/services easiness of use 0.835
Confidence in AI generated information 0.808
Speed of obtaining information 0.858
Research 0.872
Scientific comparisons 0.872
Willingness to contribute to AI training 0.722
Evaluating AI solutions compared with human intelligence solutions 0.864
F2. Disadvantages of using AI, items adapted after Chan (2023)
Inaccessibility 0.854
Dificulty of use 0.921
Lack of training for generating questions addressed to AI (prompts) 0.858
F3. Disadvantages related to the current level of AI development, items adapted from Pisica et al.,
(2023) and Chan (2023)
(Cronbach-Alpha = 0.891 > 0.7; CR = 0.903 > 0.7; AVE = 0. 611 > 0.5)
Uncertainty 0.715

AE
Item CFA
Lack of confidence in AI-generated results/ Different answers for the same question 0.744
(prompt), addressed differently
Personal data and information protection 0.724
Willingness not to contribute to AI training 0.693
Lack of legal regulations 0.905
Lack of authenticity of information provided by AI 0.883
F4. Willingness of teachers to use AI in teaching activities, items created by authors
Using AI in teaching activities 0.934
Short-term intention to use AI in teaching activities 0.894
Long-term intention to use AI in teaching activities 0.906
Advantages of using Ai in teaching activities 0.904
F5. Willingness of teachers to use AI in research activities, items created by authors
Use of AI in research activities 0.946
Short-term teachers’ intention to use AI in research activities 0.900
Long-term teachers’ intention to use AI in research activities 0.948
Advantages of using AI in research activities 0.919
F6. Willingness of teachers to use AI in evaluation activities, items created by authors
(Cronbach-Alpha= 0.970 > 0.7; CR=0.978 > 0.7; AVE = 0.918 > 0.5)
AI use in evaluation activities 0.966
Short-term intention to use AI in evaluation activities 0.942
Long-term intention to use AI in evaluation activities 0.959
Advantages of using AI in evaluation activities 0.966
The confirmatory factor analysis shows a correct construct of the latent variable Teachers’
intention to use AI on their own initiative in research, teaching and assessment, the three
major categories of activities carried out by higher education business and economics
teachers. Interpretation of the identified relationships and their intensity was based on the
outputs of Smart-PLS 4.0 (Ringle et al., 2015; Hair, 2019).
3.2. Modelling academics’ willingness to use artificial intelligence in their main

research, teaching, and student assessment activities
The output in Table no. 2 and Figure no. 1 underlines that not all three components of AI
advantages and disadvantages included in the model have a positive impact on teachers’
intention to use AI in their teaching activity in the short and long term, for a 5% significance
level. The greatest influence is on the advantages of using AI, with a mean correlation
coefficient β = 0.481 (t = 5.422, p = 0.000, CI = [0.281; 0.626]). This shows that the
respondents are teachers open to new, innovative solutions that add value to their teaching
(R = 0.47). Hypothesis H2a is validated with a 5% significance threshold. The disadvantages
of using AI tend towards 0 (out of the 95% confidence interval). Hypothesis H3a is
invalidated with an error of 3%.
Disadvantages of the stage of AI development negatively influence AI use in teaching, with
a low but statistically significant level of intensity due to teachers’ dissatisfaction with the
current stage of AI development. This validates the H4a hypothesis.
Vol. 26 • No. 65 • February 2024 267

Table no. 2. Influence of AI advantages considered on the intention to use AI

in teaching activities (R = 0.466, t = 6.195, p = 0.000)
Relationship Β t p CI
Advantages → Teaching 0.481 5.422 0.000 [0.281; 0.626]
Disadvantages of AI use → Teaching 0.265 2.193 0.028 [-0.066; 0.455]
AI disadvantages → Teaching -0.380 3.698 0.000 [-0.560; -0.210]
The model shows statistical significance and validates the positive influence of teachers’
perceived advantages on the intention to use artificial intelligence in teaching activities. The
model also validates the negative influence of the disadvantages given by the current stage
of development of artificial intelligence on its use in teaching. Advantages of using and
intention to use AI in teaching show the highest correlation coefficient, at a medium intensity
level, statistically significant for a 5% significance level. The model is valid, teachers intend
to use AI in teaching activities because of the current advantages it is offers. Thus, hypothesis
H1 is confirmed with a 95% probability.
Figure no. 1. Model regarding AI’s impact on teaching activities

The second model (in Figure no. 2) identifies the stochastic relationships between the
advantages and disadvantages, in the respondents’ opinion, and those given by the current
level of AI development (exogenous variables) and their use in academic research activities
(endogenous variable), as shown in Table no. 3.
Table no. 3. Influence of AI advantages and disadvantages considered on the intention
to use AI in research activities (R = 0.355, t = 5.670; p = 0.000)
Relationship β t p CI
Advantages → Research 0.454 4.888 0.000 [0.254; 0.612]
Disadvantages of AI use → Research 0.233 1.879 0.060 [-0.109; 0.420]
AI disadvantages → Research -0.276 1.825 0.068 [-0.461; 0.218]
The advantages of using AI on research activity also have a positive, statistically significant
impact on research activity, but at a lower intensity compared to the impact upon the teaching
activity. To measure the influence of the advantages of using AI on the willingness of
teachers to use AI in research, a coefficient β = 0.454 (t = 4.888, p = 0.000, CI = [0.254;
0.612]) is obtained as shown in Table no. 3. Approximately zero influence is caused by the
disadvantages of using AI in research, a factor that has a very small, insignificant, positive
impact on the willingness to use AI in research, for a risk of committing error of degree 1,

AE
alpha’ of 5%, as the β-value = 0.233 and the confidence interval contains the value zero (t =
1.879, p = 0.060 < 0.10, CI = [-0.109; 0.420]).
Both disadvantages of use and those due to the current level of AI development determined
the willingness to use in its confidence interval to tend toward zero. However, in terms of
research activity, the results highlight the reluctance of teachers to use AI in research activity
(R = 0.35). One explanation for the identified correlation may be the risk of plagiarism
attributed to the use of AI in research. Also, academics are capable to carry out research
without the help of AI, as the output shows that this stochastic relationship is insignificant,
research being the most complex academic activity. The model is significant and confirms
hypotheses H2b, H3b, and H4b.
The research activity is positively, statistically significantly influenced, with a low to medium
intensity, by the factor variable advantages of using AI.
Figure no. 2. Model regarding AI’s impact on research activities

The third model (in Figure no. 3) presents the stochastic relationships of constructs upon the
activities of students’ evaluation. All influence relationships on the willingness to use AI in
evaluation activity are statistically significant, although the level of intensity of correlations
is low. The low level of correlation appears due to the specifics of the evaluation activity,
which is repetitive, automatic, and does not involve scientific creativity and innovation.
(Table no. 4)
Table no. 4. The influence of AI on the evaluation activity (R = 0.343, t = 5.129, p = 0.000)
Relationship β t p CI
Advantages → Evaluation 0.445 5.181 0.000 [0.257; 0.592]
Disadvantages of IA use→ Evaluation 0.227 2.318 0.020 [0.007; 0.406]
AI disadvantages → Evaluation -0.290 2.835 0.005 [-0.470; -0.095]
The third model is statistically significant and confirms hypotheses H2c, H3c, and H4c with
95% probability and does not confirm hypothesis H5 with a 5% risk of committing the first-
order error.
Vol. 26 • No. 65 • February 2024 269

Figure no. 3. Model regarding AI’s impact on evaluation activities
3.3. Discussions
The AI influence students’ evaluation activities does not appear to be particularly appreciated
by teachers, as the correlation coefficient shows a stochastic correlation of low intensity (R
= 0.343), but statistically significant (p = 0.000). The low influence is explained by the fact
that students’ evaluation activity cannot be replaced by AI, but is only influenced by AI, as
it is a complex activity that requires human decisions, as teachers hold the legal and direct
responsibility of evaluating students.
Despite the fact that there are disadvantages of using AI, determined by its level of
development, academic teaching staff show an open mind in using AI in the students’
evaluation process. This can be explained by the intensity of the influence of the advantages
on the willingness to use AI in evaluation activities. Variables restricting the willingness to
use AI are: limitations related to data protection and security, uncertainty, and mistrust
generated by insufficient regulation of AI.
Regarding the teaching activity, it can be seen that the advantages of using AI have a positive
influence on teachers’ willingness to use AI in teaching. The study shows an unwillingness
of teachers to use AI due to the disadvantages of using AI. Inference cannot be made; the
relationship is statistically insignificant, zero occurs in the confidence interval with a 5%
significance level. One explanation for the reluctance to use AI in teaching is due to the
disadvantages of using AI. This is explained by teachers’ fear of being compared to AI. The
disadvantages of the current stage of development of AI in teaching have the greatest negative
influence, showing that there is reluctance to expose the personal knowledge to students,
compared to the level of AI.
In the case of the research activity, it can be seen that the advantages of using AI have a
positive and significant influence on the use of AI. Disadvantages do not have a statistically
significant influence on the willingness to use AI in research. This is explained by the high
level of research potential of the teachers. AI is not necessarily needed as a research tool.

AE
Conclusions
The results of this research may have implications on both how AI is used in academic
education and for future prospects of formal regulations. The use of AI in the educational
process is still at an early stage. This result is confirmed by the novelty of the product, the
small sample size, and the beta correlation coefficients in the models analysed, which show
a low level of intensity of the correlations, despite the benefits generated by the use of AI.
In the opinion of the responding teachers, for automatic processes that require less creativity,
such as repetitive or students’ evaluation activities, there is a greater willingness of teachers
to use AI. At the same time, for activities that require more cognitive functions and creativity,
such as research work, the usefulness of AI is reduced. The results on the disadvantages of
using AI obtained from the modelling are inconclusive. In this sense, further studies may be
a direction to continue the research. This is confirmed by studies on the application of AI in
other fields of activity. Huang and Rust (2019) confirm that AI implementation stages start
from simple, automatic, repetitive processes, and AI will be implemented successively for
more powerful functions. It is likely that AI will be increasingly implemented in academia
as a result of social pressures and at the demand of the business environment, which adapts
faster to this pressure (Pelau et al., 2021). Structural equation modelling shows the
complexity of the constructs that influence the use of AI in academia. It is important for
universities, with a view to future regulation of AI use, to know the behavioural influencing
factors of AI use by academics, in order to identify, through further research, which factors
can be controlled and regulated internally by universities.
The usefulness of this research consists of its impact in the Romanian business and economic
academic environment, because it expands the theories of acceptance of the use of technology
in education (Gursoy et al., 2019). Possibilities for further research are: expanding the pilot
sample, refining the items by introducing qualitative items in the models after quantifying
them, and adding other items in the modelling, referring to the regulation of ethics and
academic integrity issues. The success or failure of the use of AI in academia will depend on
the ability of teachers, regardless of region or country of origin, to adapt to the social and
business pressures of increasing use of new technologies (Gursoy et al., 2019).
The limitations of the research are represented by the relatively small size of the sample and
the impossibility to include qualitative items in the quantitative modelling. Qualitative items
will find their usefulness in the construct developments in subsequent studies of this pilot
survey. The transformation of qualitative items into quantitative items in future research will
be necessary due to the exponential development of artificial intelligence and the social
pressure to use it.
Concluding, AI also has a number of advantages in the administrative field of education, such
as plagiarism checking, scheduling courses and seminars, preparing curricula, managing
student and teacher data, and evaluating teachers’ research activities. All these advantages
offered by AI presented in the paper save time and resources by allowing teachers and
education staff to focus more on teaching-learning, research, and evaluation. Alongside these
multiple advantages of AI, it has created a number of controversies and challenges through
the disadvantages it nevertheless entails such as systemic bias, discrimination, inequality for
marginalised groups of students, and xenophobia (Hao, 2017; Hwang and Tu, 2021; Pisica
et al., 2023) to issues of confidentiality and bias in data collection and processing (Holmes et
al., 2022). In this context, we can state that, in the academic environment, academics agree,
Vol. 26 • No. 65 • February 2024 271

to a certain extent, to implement and use of AI in the educational process, mainly in the
learning and students’ evaluation process and, to a smaller extent, in the research process.
References
Baker, T. and Smith, L., 2019. Educ-AI-tion rebooted? Exploring the future of artificial
intelligence in schools and colleges. Exploring the future of artificial intelligence in
schools and colleges. [online] Available at: <https://media.nesta.org.uk/documents/
Future_of_AI_and_education_v5_WEB> [Accessed 18 September 2023].
Bessen, J. E., Brynjolfsson, E. and Hui, T., 2020. AI and the Economy. Annual Review of
Economics, 12(1), pp. 621-642.
Brynjolfsson, E. and McAfee, A., 2017. The business of artificial intelligence. Harvard
Business Review, 95(1), pp. 20-30.
Brynjolfsson, E. and McAfee, A., 2018. The business of artificial intelligence. Harvard
Business Review, 96(1), pp. 1-13.
Chan, C.K.Y., 2023. A comprehensive AI policy education framework for university
teaching and learning. International Journal of Educational Technology in Higher
Education, [e-journal] 20(1), article no. 38. https://doi.org/10.1186/s41239-023-00408-3.
European Commission, 2023a. Promoting equity and non-discrimination in the development
and use of AI. European Commission Report, 2023/002.
European Commission, 2023b. Demanding transparency and explainability in AI. European
Data Protection Supervisor’s Guide, 2023/003.
European Commission, 2023c. The potential impact of regulation on innovation in AI.
European Union Study, 2023/005.
European Parliament, 2023a. EU AI Act: first regulation on artificial intelligence. [online]
Available at: <https://www.europarl.europa.eu/news/en/headlines/society/20230601STO
93804/eu-ai-act-first-regulation-on-artificial-intelligence> [Accessed 19 September
2023].
European Parliament, 2023b. Data Protection and Privacy in AI. Report of the European
Parliament, 2023/004.
European Union Agency for Fundamental Rights FRA, 2021. Understanding the Future -
Artificial Intelligence and Fundamental Rights report. [pdf] Available at:
<https://fra.europa.eu/sites/default/files/fra_uploads/fra-2021-artificial-intelligence-
summary_ro.pdf> [Accessed 17 September 2023].
Gillani, N., Eynon R., Chiabaut, C. and Kelsey, F., 2023. Unpacking the “Black Box” of AI in
Education. Journal Educational Technology & Society, [e-journal] 26(1), pp. 99-111.
https://doi.org/10.30191/ETS.202301_26(1).0008.
Gursoy, D., Chi, O.H., Lu, L. and Nunkoo, R., 2019. Consumers acceptance of artificially
intelligent (AI) device use in service delivery. International Journal of Information
Management, [e-journal] 49, pp. 157-169. https://doi.org/10.1016/j.ijinfomgt.
2019.03.008.
Hair, J.F., Risher, J.J., Sarstedt, M. and Ringle, C.M., 2019. When to use and how to report
the results of PLS-SEM. European Business Review, [e-journal] 31(1), pp. 2-24.
https://doi.org/10.1108/EBR-11-2018-0203.

AE
Hao, K., 2017. The ethical questions that haunt facial-recognition research. [online] MIT
Technology Review. Available at: <https://www.technologyreview.com/2018/02/13/
145424/the-ethical-questions-that-haunt-facial-recognition-research/> [Accessed 23
October 2023].
Santos, O.C., Rodrigo, M.T., Cukurova, M., Bittencourt, I.I. and Koedinger, K.R., 2022.
Ethics of AI in Education: Towards a Community-Wide Framework. International
Journal of Artificial Intelligence in Education, [e-journal] 32(3), pp. 504-526.
https://doi.org/10.1007/s40593-021-00239-1.
Huang, M.-H. and Rust, R.T., 2018. Artificial Intelligence in Service. Journal of Service
Research, [e-jounal] 21(2), pp. 155-172. https://doi.org/10.1177/1094670517752459.
Hwang, G.-J. and Tu, Y.-F., 2021. Roles and Research Trends of Artificial Intelligence in
Mathematics Education: A Bibliometric Mapping Analysis and Systematic Review.
Mathematics, [e-journal] 9(6), article no. 584. https://doi.org/10.3390/math9060584.
IEEE Standards Association, 2023. IEEE Standards Association. [online] Available at:
<https://standards.ieee.org> [Accessed 18 October 2023].
Johnson, R., 2018. Ethical frameworks for AI in higher education. International Journal of
Artificial Intelligence in Education, 35(2), pp. 65-79.
Johnson, R., 2020. Challenges of data privacy and security in the context of AI in higher
education. International Journal of Artificial Intelligence in Education, 30(3),
pp. 87- 102.
Jordan, M.I., 2019. Artificial Intelligence – The Revolution Hasn’t Happened Yet. Harvard
Data Science Review, [e-journal] 1.1. https://doi.org/10.1162/99608f92.f06c6e61.
Kakkonen, T., Kailanto, M., Tebest, T.H. and Lyytinen, A., 2019. Machine learning approach
to detection of student essays with implausible idiosyncrasies. Journal of Learning
Analytics, 6(3), pp. 91-107.
Kamalov, F., Santandreu Calonge, D. and Gurrib, I., 2023. New Era of Artificial Intelligence
in Education: Towards a Sustainable Multifaceted Revolution. Sustainability, [e-journal]
Khan, G.F., Sarstedt, M., Shiau, W.-L., Hair, J.F., Ringle, C.M. and Fritze, M.P., 2019.
Methodological research on partial least squares structural equation modeling (PLS-
SEM): An analysis based on social network approaches. Internet Research, [e-journal]
29(3), pp. 407-429. https://doi.org/10.1108/IntR-12-2017-0509.
Khogali, H.O. and Mekid, S., 2023. The blended future of automation and AI: Examining
some long-term societal and ethical impact features. Technology in Society, [e-journal]
73, article no. 102232. https://doi.org/10.1016/j.techsoc.2023.102232.
Khosravi, H., Shum, S.B., Chen, G., Conati, C., Tsai, Y.-S., Kay, J., Knight, S., Martinez-
Maldonado, R., Sadiq, S. and Gašević, D., 2022. Explainable Artificial Intelligence in
education. Computers and Education: Artificial Intelligence, [e-journal] 3, article no.
Makridakis, S., 2017. The forthcoming Artificial Intelligence (AI) revolution: Its impact on
society and firms. Futures, [e-journal] 90, pp. 46-60. https://doi.org/
10.1016/j.futures.2017.03.006.
Vol. 26 • No. 65 • February 2024 273

Marr, B., 2015. Big Data: Using SMART big data, analytics and metrics to make better
decisions and improve performance. S.l.: John Wiley & Sons Ltd.
Ministry of Education, 2022. Raport privind starea învățământului superior din România
2021-2022. [online] Available at: <https://www.edu.ro/sites/default/files/
_fi%C8%99iere/Minister/2022/Transparenta/Starea_invatamantului/Raport-Starea-
nvatamantului-superior-2021-2022.pdf> [Accessed 4 November 2023]
Identity and Human Skills. Amfiteatru Economic, [e-journal] 23(56), pp. 33-45.
https://doi.org/10.24818/EA/2021/56/33.
Perna, C., Doherty, N.F. and Grubnic, S., 2020. Artificial Intelligence (AI) in organizations:
Ethical considerations. International Journal of Management Reviews, 22(4),
pp. 419-446.
Societies, [e-journal] 13(5), article no. 118. https://doi.org/10.3390/soc13050118.
Pop, R.A., Dabija, D.C., Pelau, C., and Dinu, V. 2022. Usage intentions, attitudes, and
behaviors towards energy-efficient applications during the COVID-19 pandemic. Journal
of Business Economics and Management, [e-journal] 23(3), pp. 668-689.
https://doi.org/10.3846/jbem.2022.16959.
Ringle, C.M., Wende, S. and Becker, J.M., 2015. SmartPLS 3. Boenningstedt: SmartPLS
GmbH.
Rodríguez, G.L. and Brito, V.S., 2017. Artificial intelligence in higher education.
Opportunities and threats. INNOVA Research Journal, 2(8.1), pp. 412-422.
Rosenblatt, F., 1958. The perceptron: A probabilistic model for information storage and
organization in the brain. Psychological Review, [e-journal] 65(6), pp. 386-408.
https://doi.org/10.1037/h0042519.
Rumelhart, D.E., Widrow, B. and Lehr, M.A., 1994. The basic ideas in neural networks.
Communications of the ACM. [e-journal] 37(3), pp. 87-92. https://doi.org/10.1145/
175247.175256.
Selwyn, N. and Gašević, D., 2020. The datafication of higher education: Discussing the
promises and problems. Teaching in Higher Education, 25(4), pp. 527-540.
Sharma, P.N., Liengaard, B.D., Hair, J.F., Sarstedt, M. and Ringle, C.M., 2023. Predictive
model assessment and selection in composite-based modeling using PLS-SEM:
extensions and guidelines for using CVPAT. European Journal of Marketing, [e-journal]
57(6), pp. 1662-1677. https://doi.org/10.1108/EJM-08-2020-0636.
Smith, A., 2019. Ethical Considerations in Implementing AI in Higher Education. Journal of
Higher Education Management, 44(4), pp. 145-159.
Topol, E., 2019. Medicina profunda. Sanatatea redevine umana cu ajutorul inteligentei
artificiale. București: Corint Future.
Universitatea „Alexandru Ioan Cuza” din Iași, 2016. Asociația Română pentru Inteligență
Artificială - cinci ani de existență. Universitatea „Alexandru Ioan Cuza” din Iași.
[online] Available at: <https://www.uaic.ro/asociatia-romana-pentru-inteligenta-
artificiala-cinci-ani-de-existenta/> [Accessed 17 September 2023].

AE
REDESIGN OF ACCOUNTING EDUCATION TO MEET THE CHALLENGES

OF ARTIFICIAL INTELLIGENCE – A LITERATURE REVIEW
Valeriu Brabete1 , Cătălin Mihail Barbu2 , Daniel Cîrciumaru3 ,
Daniel Goagără4 and Dorel Berceanu5
1)2)3)4)5)
University of Craiova, Craiova, Romania

Brabete, V., Barbu, C.M., Cîrciumaru, D., Goagără, D. Received: 27 September 2023
and Berceanu, D., 2024. Redesign of Accounting Revised: 29 November 2023
Education to Meet the Challenges of Artificial Accepted: 21 December 2023
Intelligence – A Literature Review. Amfiteatru
Economic, 26(65), pp. 275-293.
Abstract
The development of emerging technologies, including artificial intelligence (AI), has a direct
impact on the accounting profession in the sense of redefining the role and place that
accounting specialists occupy within economic entities. This paper is a systematic literature
review and aims to emphasise and synthesise the main challenges regarding the impact that
artificial intelligence and other related technologies exert on accounting professionals and
the competencies they need to develop or acquire. At the same time, the outline of specific
solutions was envisaged, through which accounting professionals can be supported in their
efforts to adapt to the new requirements of the labour market imposed by the evolution of
disruptive technologies. To achieve the proposed objectives, Web of Science (WoS) was used
as an academic database and specific tools for bibliometric analyses, including the
VOSviewer application. The results identify universities as the main pillar in the process of
the redesign of accounting education, and update of the curriculum in the field as the main
measure to be undertaken. The originality of the paper resides in the identification of specific
solutions for curriculum reconfiguration in the field of accounting education, the highlighting
and overcoming of possible challenges for the redesign of the education process in the field
of accounting, under the impact of AI.
Keywords: accounting profession, artificial intelligence, digitalization, professional

competencies, accounting curriculum, accounting education
JEL Classification: I20, J24, M40, N30, O33

Corresponding author, Cătălin Mihail Barbu – e-mail: [email protected]
Vol. 26 • No. 65 • February 2024 275

AE Redesign of Accounting Education to Meet the Challenges
of Artificial Intelligence – A Literature Review
Introduction
The remarkable evolution of technology requires a comprehensive understanding of its
impact not only in the context of industry, education, and society, but also in terms of labour
relations and professions, an impact that comes with a set of difficulties, which must be
overcome, and opportunities, which must be capitalised (Tavares et al., 2023). Among the
factors behind these difficulties or opportunities we mention the transition to a digitalised
and resource-efficient economy, the development of robotics and artificial intelligence, as
well as demographic factors (Asonitou, 2022).
The global digital revolution has significantly and irrevocably transformed human society
and the activities carried out, and in this context the accounting profession is expected to
experience a major change in the near future as a result of technological developments.
Practices will be automated and related positions will become obsolete, so that accounting
graduates as well as practitioners in the field should be educated for newly and different tasks
and positions (Al-Htaybat, von Alberti-Alhtaybat and Alhatabat, 2018). Also, the rapid
evolution of technology and the important changes induced by this process in the business
world, have a direct impact on accounting professionals, with advances in the fields of
automation and machine learning, artificial intelligence, data analysis and blockchain. At the
same time, these are being considered disruptive technological elements in the accounting
domain (Aldredge, Rogers and Smith, 2021). These disruptive technologies, being extremely
dynamic, are significantly reshaping business processes (Moore and Felo, 2022) with direct
implications on the status and on the role played by the accounting professionals.
In the context of the effects produced by technological developments, accounting education
and the qualification of accounting students for the current and especially future requirements
of the labour market is considered one of the notable issues that seem to have been overlooked
in many of the education programs of the majority universities (Al-Hattami, 2021). This issue
is raised as the ongoing digital revolution demands the accounting profession to adapt and
bring forward new accounting services, that involve a wider set of competencies than
commonly recognised (Banasik and Jubb, 2021).
Against this background, we have identified a research opportunity, in that there is still a lack
of holistic understanding of how the adoption of AI forces accounting professionals to
redefine, adapt or develop previously acquired competencies, or even acquire new
competencies. Therefore, we propose three research questions (RQ), which we aim to
answer:
 RQ1: In the context of the development of artificial intelligence systems, which are the
competencies that the professional accountants must acquire or develop, to adapt to the new
requirements of the labour market?
 RQ2: Which are the mechanisms to support the accounting professionals in their effort
to reconfigure these competencies?
 RQ3: In the context of the expansion of artificial intelligence systems, which is the role
of the educational system and what are the directions aimed by the curriculum development
process, to support the students or the accounting professionals in the process of adaptation
or development of their competencies?

AE
We consider it appropriate to choose this topic, because, as other authors have pointed out
(Leitner-Hanetseder et al., 2021; Tavares et al., 2023; Yigitbasioglu, Green and Cheung,
2023; Jackson, Michelson and Munir, 2023), in the literature there are relatively few studies
investigating the influence and future consequences of new technologies on the accounting
profession. Based on a systematic literature review, the purpose of our study is twofold:
 to identify, group, and systematise knowledge regarding the impact that artificial
intelligence and other related technologies have on accounting professionals and the
competencies they need to develop or acquire;
 to outline specific solutions through which accounting professionals might be supported
in their efforts to adapt to the new requirements of the labour market imposed by the evolution
of these disruptive technologies.
The results of our study identify the competencies needed by the accounting professionals in
the context of implementing AI-based solutions, the stakeholders involved in obtaining these
competencies, and the role of the educational system in this endeavour. Analysing the impact
of AI on accounting education, this paper makes significant contributions to both theory and
practice: theoretically, we propose a new perspective on the impact of AI on accounting
education, by integrating dissipated research in a new analysis framework, while from a
practical perspective, we identify a series of challenges and specific actionable measures to
overcome these challenges. Our paper is structured as follows: the first section is dedicated
to the rigorous review of the scientific literature and allows us to highlight the general context
determined by the development of artificial intelligence systems, with direct reference to the
impact on the accounting profession; in the second section we explain in detail the research
methodology we used; in the third part of the paper we analyse the results obtained; in the
fourth part we critically discuss the results; in the fifth section we make proposals for new
research directions; finally, we present the conclusions of the research and its limits.
Rapid technological progress directly influences the way businesses are organised and run
(Almufadda and Almezeini, 2022; Birt, Safari and de Castro, 2023), and dramatically
transform economic activity (Bonzanini et al., 2020). The development and implementation
of artificial intelligence were defining elements for triggering the fourth industrial revolution,
which intensified technological changes with impact in almost all sectors of life (Cunha et
al., 2022).
The accounting profession is also undergoing through an important change and
transformation due, in particular, to technological developments, among which digitalisation
and artificial intelligence have a significant impact (Almeida and Carvalho, 2022). In this
context, the accounting professionals must adapt to the new technological alternatives on the
market, which prove to be superior to the traditional accounting system (Ciurea and Man,
2020). Moreover, Gușe and Mangiuc (2022, p. 264) state that “digitalisation will rewrite the
social contract between people and the business environment, the largest gap occurring not
between developed and emerging economies, but between generations born and educated in
the analogue age and those from the digital age”.
Vol. 26 • No. 65 • February 2024 277

It is considered that all these changes demand an adaptation or even a significant

transformation of the competencies required to accounting professionals, so that they can
successfully fulfil their new tasks (Damasiotis et al., 2015) and effectively use the
technologies, found in a continuous evolution, to provide further valuable support to all
interested parties (Banța et al., 2022).
Against the background of these profound transformations, both those who are preparing to
become professional accountants and those who are already being active in this field feel the
need for support to facilitate their approach aimed to redefine, to adapt or to develop the
competencies required by the new conditions existing on the labour market, in the context of
the development of systems based on artificial intelligence.
The solutions identified by researchers in the scientific literature bring into discussion
different actors and diverse ways to support both accounting students and practitioners in
their efforts to adapt to the new context of the labour market.
Jackson, Michelson and Munir (2023) assign roles in this process to universities, professional
organisations and employers, considering that it is necessary to have a partnership between
all these interested parties, which aims to rethink the curriculum, ensure programs relevant
continuous professional development, provide part-time jobs, coaching, mentoring and on-
the-job training, etc. On the same line of cooperation between several actors, Bonzanini et al.
(2020) appreciate that higher education institutions, businesses, professional associations and
public authorities will have to cooperate in the next framework of Industry 4.0. Fredo et al.
(2023) also sustain the idea of collaboration between academia and professional
organisations and argue in favour of the need to diversify digital, communication and social
knowledge and skills. Kokina and Davenport (2017) argue that academia needs to
reconceptualise the accounting curriculum, regulators need to formulate revolutionary
policies, while professional bodies need to redesign the professional development process.
Gonçalves, da Silva and Ferreira (2022) consider that the main role in the process of
education and continuous professional development must be given to higher education,
which must keep pace with technological development to provide specialists with the
necessary competencies so that they remain relevant in relation to the dynamic requirements
of the labour market. However, national and international professional associations are also
involved in this process and they must play a significant role in defining the competencies
sets that are necessary for the training of professionals in the current and future context. Also,
Kroon, Alves and Martins (2021) state that regulators also have an important role,
considering that they must be careful to adjust standards according to the current digital state,
while Birt, Safari and de Castro (2023) argue that the roles must be divided between
educational institutions, standardisation bodies, and professional ones.
There are points of view that shed light on the accounting professionals, as they have a direct
responsibility, given that it is appreciated that the rapid and unpredictable changes of the
profession bring forward the need for continuous learning, flexibility and permanent
adaptation to new circumstances (Li and Zheng, 2018; Grabińska, Andrzejewski and
Grabiński, 2021; Cunha et al., 2022).
Even if the researchers identify several actors that must play a role in the process of
adaptation of accounting professionals to the new requirements of the labour market,
nevertheless, the study of the specialised literature reveals the primary importance of
educational institutions. The majority of the opinions expressed, and in agreement with our

AE
views, converge towards the need to adapt the curriculum in order to ensure the competencies
required by the current and especially the future technological context (Al-Htaybat, von
Alberti-Alhtaybat and Alhatabat, 2018; Al-Hattami, 2021; Banasik and Jubb, 2021; Cong,
2021; Junger da Silva et. al., 2021; De Villiers, 2021; Kroon, Alves and Martins, 2021;
Almeida and Carvalho, 2022; Gonçalves, da Silva and Ferreira, 2022; Bunea and Guinea,
2023; Elo et al., 2023; Karcioglu and Binici, 2023; Landsberg and van den Berg, 2023;
Tavares et al., 2023).
Researchers express heterogeneous opinions regarding the set of competencies that the
accountants of the future should possess, so that they will continue to fulfil a strategic role
within economic entities:
 Yigitbasioglu, Green and Cheung (2023) consider that professional accountants will
play a significant role in the assessment and assurance of new technologies, and this requires
them to master competencies that will allow their involvement in activities aimed at
monitoring and interpreting big data generated by the use of artificial intelligence, in order
to ensure compliance with legal and ethical requirements;
 Tsiligiris and Bowyer (2021) state that the skills needed by future accountants could be
classified into four categories: ethical skills, digital skills, business skills, and soft skills;
 Pargmann et al. (2023) highlight the need to combine technical and professional
competencies in the different areas of accounting, as well as the growing importance of meta-
competencies such as proactive thinking, self-control, creativity, and interdisciplinary action;
 Pasewark (2021) claims that, at the present, the accountants must master unprecedented
analytical and technological competencies, have practical experience in using business
intelligence software, interact with robots, develop skills in leadership, human resources,
global awareness, diversity awareness, and collaboration;
 Kroon, Alves and Martins (2021) write about additional training in business strategy
and business models; understanding the basics of programming and developing the skills
needed to keep up with the latest technologies; development of business analysis capabilities;
acquiring soft skills such as intuition, creativity and communication;
 Li and Zheng (2018) emphasise interdisciplinary skills that allow accountants to use the
advantages of artificial intelligence. They consider it important for accountants to acquire
skills to analyse financial data, to master relevant knowledge about the business, but also to
combine big data with artificial intelligence in the tasks they perform.
Considering the context presented and the divergent opinions that the specialised literature
highlights, our scientific approach will be directed to identify and to synthesise the main
challenges regarding the impact of AI on accounting education, which will allow us to answer
the research questions.
Our study is based on a systematic review of the specialised literature, including specific
tools of bibliometric analyses considered popular and rigorous methods for exploring and
analysing large volumes of data (Donthu et al., 2021). The specific stages of the research
methodology that we applied to carry out our scientific approach are summarised in figure
Vol. 26 • No. 65 • February 2024 279

no. 1. The data collection stage was very important for the relevance of the obtained results,
which is why we paid considerable attention to it. In our research we chose the Web of
Science (WoS) as an academic database because it is considered one of the most important
research databases in the world (Kamble, Gunasekaran and Gawankar, 2018).
Stage 1 – Research design

 establishing the research theme;
 choosing the research method (systematic analysis of specialized literature);
 establishing the research questions.
Stage 2 - Data collection

 selection of the academic database (Web of Science);
 establishing keywords and using them in the automatic search process (N=2699);
 setting search parameters for automatic filtering of results:
- year of publication: 2015-2023 (N=1995);
- WoS categories: Business Finance, Education Educational Research, Management, Business,
Economics, Computer Science Information Systems, Social Sciences Interdisciplinary,
Education Scientific Disciplines, Sociology (N=1429);
- type of documents: article, proceeding paper, early access, book chapters, book (N=1367);
- language: English (N=1268);
- open access articles (N=411).
 manual filtering of the results according to the relevance of the title and abstract, exporting
data from WoS and forming the intermediate database (N=251);
 reading the 251 downloaded articles, excluding the articles whose content was not considered
relevant in relation to the research objectives, and compiling the final database from the
articles considered relevant for achieving the research objectives (N=82).
Stage 3 - Analysis of the obtained data and answering the research questions
 selection, systematization, analysis and capitalization of the ideas extracted from the 82
articles considered relevant to achieve the research objectives;
 formulating the answers to the research questions.
Stage 4 - Formulating conclusions and establishing future research directions

 drawing up conclusions based on the answers to the research questions;
 establishing new future research directions.
Figure no. 1. Stages of the research methodology

For the query of the scientific database, a set of keywords was used, linked by the Boolean
operators “AND” and “OR”, which targeted the topic, allowing the query to follow the title
of the papers, their summary and the keywords. The keywords used to query the database are
presented in table no. 1.

AE
Table no. 1. The structure of the keywords used for the query
WoS query criteria
TS = (“accounting” AND “artificial intelligence”) OR TS = (“accounting” AND “digitalization”)
OR TS = (“accounting” AND “accounting profession”) OR TS = (“accounting” AND
“competences”) OR TS = (“accounting” AND “curriculum”)
The interval set for the query concerned the period 2015-2023, considering that starting from
2015/2016, publications about the influence of artificial intelligence on industry, education,
or other aspects of society began to appear in the literature (Tavares et al., 2023). An
important search parameter was also represented by WoS scientific categories. We selected
nine categories considered relevant to the field in which the research carried out is framed.
The distribution of the 82 articles in the final sample among the nine categories is as follows:
Business - 10, Business Finance - 35, Computer Science Information Systems - 10,
Economics - 5, Education Educational Research - 9, Education Scientific Disciplines - 6,
Management - 3, Sociology - 2, Social Sciences Interdisciplinary – 2. At the same time, the
automatic filtering also had as a reference element the type of documents queried, the
following types being selected: article, proceeding paper, early access, book chapters, book.
Out of the 82 articles that make up the final sample, 66, representing 80.49%, are represented
by articles published in journals, and 16, representing 19.51%, are articles published in the
volumes of international conferences.
To emphasise the relevance of the final database, we also used VOSviewer, an IT software
product created by van Eck and Waltman (2010), which allowed us to develop some
suggestive bibliometric maps. In order to create the bibliometric maps with the help of
VOSviewer, it was necessary to export the WoS database in tab-delimited file format. Using
this software we were able to perform an analysis of the relevance of the keywords used in
the WoS query process, as can be seen in figure no. 2.
Figure no. 2. Occurrence frequency of keywords

The map generated by VOSviewer highlights the frequency of keywords used by the authors
and, from its analysis, it can be seen that among the most used keywords are also those used
Vol. 26 • No. 65 • February 2024 281

by us in the query process. Moreover, it can be observed that three of the keywords used by
us, “artificial intelligence”, “accounting” and “accounting profession”, are in the top of the
most frequently used ones by the authors of the articles included in the final sample, fact that
highlights the relevance of the keywords used in the query process.
Another analysis we carried out through VOSviewer allowed us to highlight the scientific
value of the articles included in our sample, through the journals in which these articles were
cited. Figure no. 3 presents the co-citation network by cited source and indicates the journals
in which are published the articles that most extensively cite the studies we selected. This
emphasises the fact that our database was made up of valuable articles from the specialised
literature, which were cited in prestigious journals. One can see the existence of a consistent
network of co-citations, as well as the fact that journals from the main stream of publications
are pointed out. The best positioned journals from this point of view are Accounting
Education, Journal of Emerging Technologies in Accounting, Accounting Horizons, and
Issues in Accounting Education.
Figure no. 3. Co-citation of the journals included in the sample

The credibility of the articles included in the final analysis, is also underlined by their citation
rate, as shown in Table no. 2.
Table no. 2. Citation rate of sample articles*
Citations Percent Citations Percent
Papers Articles
(Web of Science) (%) (Google Scholar) (%)
0-9 citations 74,39 61 0-9 citations 47,56 39
> 60 citations 3,66 3 > 60 citations 17,07 14
Total 100 82 Total 100 82
Note: * Results valid on August 1st, 2023
The data in Table no. 2 highlight the fact that the analysed articles have a consistent impact
in the specialised literature, taking into account the relatively short interval since their
publication. Thus, more than 25% of the articles have more than 10 citations on WoS, a fact
that proves the importance of the impact of AI in the field of accounting, as the results of our
study also highlight.

AE
3. Results, discussion and new research directions

In order to answer the first research question (RQ1) and to conclude what are the
competencies that the accountants of the future must master, we considered the three main
pillars on which their professional competencies must be based: technology, data and human
interaction (Birt, Safari and de Castro, 2023; Tavares et al., 2023). Against this background,
the study of the articles from our sample allowed us to group the identified competencies into
four categories: digital competencies, data analysis competencies; transversal competencies,
active participation in strategic management.
Regarding the digital competencies, knowledge of digital technologies generates one of the core
skills needed by modern accounting and finance professionals (Banasik and Jubb, 2021).
Concerning data analysis competencies, artificial intelligence will generate a large volume of
data, and therefore it is necessary for the accounting professional to be able to operate in a
versatile way with this data (Banasik and Jubb, 2021; Yigitbasioglu, Green and Cheung, 2023).
Analytical thinking, the ability to identify, select, manipulate, analyse, and interpret data to aid
decision-making is complemented by competencies in data governance, related to how data is
collected, transferred, stored, accessed, and reported (Aldredge, Rogers and Smith, 2021;
Leitner-Hanetseder et al., 2021; Pasewark, 2021). With regard to transversal competencies, our
study identified four main types of social skills: teamwork skills, communication skills,
leadership skills, and critical thinking skills (Tsiligiris and Bowyer, 2021; Almufadda and
Almezeini, 2022; Tavares et al., 2023; Vanhove, Opdecam and Haerens, 2023). Active
participation in strategic management requires the accounting professional to master advanced
decision-making competencies. The accounting professional must become an adviser to top
management in organisations, being able to create predictive analyses and to develop
alternatives based on data analysis, which contribute to the value creation in organisation
(Al-Htaybat, von Alberti-Alhtaybat and Alhatabat, 2018; Onyshchenko et al., 2022).
The second research question (RQ2) that we intend to answer is aimed at identifying the
measures by which accountants can be helped in the process of acquiring the new skills
induced by the evolution of disruptive technologies, including AI. The study of the reviewed
articles allowed us to identify four main categories of stakeholders involved in this
endeavour: higher education institutions, professional associations, regulatory bodies in the
field of accounting, and the institutional environment in a broad sense.
From the analysis of the specialised literature, we identified several measures that can be
undertaken by the four categories of interested parties. These measures can be synthesised as
follows, in relation to:
 the university environment: adapting the curriculum to the new requirements of the
labour market and improving the infrastructure in universities;
 professional associations: ensuring relevant continuous professional development
programs and collaboration with the university environment in redesigning the curriculum;
 regulatory bodies: updating the regulatory framework to allow the implementation of
new changes in accounting practice and education;
 the institutional environment: collaboration with the academic environment and
professional associations for the redesign of the curriculum and the integration of employees
in professional development programs.
Vol. 26 • No. 65 • February 2024 283

The third research question (RQ3) addresses the role of the educational system and the
directions that the curriculum development process must follow in the context of the
development of artificial intelligence systems. Our analysis highlighted that researchers
consider courses and academic staff as the main pillars of this approach.
First, we found that researchers frequently discuss the need to reconfigure the curriculum by
introducing courses that deal with data analysis, big data, data mining, machine learning,
information security, programming, robotics, expert systems, and data retrieval, machine
learning, AI technology, blockchain, cloud computing. At the same time, most analysed papers
discuss the need to include some disciplines dealing with data analysis. This is considered to be
justified in the context of the implementation of AI-based technologies, because these
technologies generate a large volume of data that must be analysed and interpreted.
Another direction is aimed at academics and considers that they must be proactive (Aldredge,
Rogers and Smith, 2021) in the adaptation of the modern teaching methods, such as audio-video
tools, specialised software, tutorial consultations, simulations, and enterprise games (De
Villiers, 2021). Technical support is needed for academics to become versed in the new
methods, to develop transversal and technological competencies (Almeida and Carvalho,
2022). Moreover, it is considered to be useful for academics and practitioners to cooperate for
the creation of modern didactic materials, adapted to the current context (Asonitou, 2022).
We present the identified solutions to the research problems in a synthetic manner in Table no. 3.
Table no. 3. Synthesis of the results
Research
Results of the research
questions
Competencies Digital Data analysis Transversal Active participation
needed by competencies competencies competencies in strategic
accounting management
professionals
Ways in Universities: Professional Regulatory bodies:Business
which adapting the associations: updating the environment:
accounting curriculum to the ensuring relevant regulatory collaboration with
professionals new demands of professional framework to academia and
can be helped the labour market development enable the professional
to reconfigure and improving the programs and implementation of associations for
these infrastructure in collaboration with new changes in curriculum redesign
competencies universities. the academic accounting and employee
environment in practice and integration into
curriculum education. professional
redesign. development
programs.
The role of Introducing new Professional Adapting the Collaboration with
educational courses, adding development of teaching methods. practitioners for the
system in the learning units to teachers. creation of didactic
context of already existing materials.
implementing courses or the
AI-based blended approach.
systems

AE
4. Discussion
4.1. Competencies needed by professional accountants
The importance of technology’s knowledge is highlighted by the fact that it can support the
formation of other competencies. The Internet of Things, big data, data analysis, cloud
computing, electronic archives, artificial intelligence, business intelligence, RPA, next-
generation ERP, electronic invoices, mobile payments, and blockchain are just some of the
technologies that will need to be mastered (Cong, 2021; Almeida and Carvalho, 2022).
Accountants need to have a systemic thinking and must understand the design principles of a
cyber system powered by artificial intelligence (Sutton, Arnold and Holt, 2018). It is certainly
that are not the accountants who will develop accounting systems based on AI, but it is
necessary for them to have some knowledge in order to validate the data generated by AI
systems, especially when the results may seem contradictory or further away from the reference
benchmark (Pasewark, 2021). Accountants must learn to work with automation robots and
supervise them. While accountants do not need to become data scientists or machine learning
experts, basic knowledge and training in statistics, machine learning, data analysis, and
programming are needed so that they can work with machine learning and AI professionals to
design and adjust specific accounting systems (Sun, 2019). Knowledge of business
management software is also important.
The accountant should be able to gather and interpret data and communicate conclusions rather
than simply report them (Birt, Safari and de Castro, 2023). Data security competencies are also
important (Al-Htaybat, von Alberti-Alhtaybat and Alhatabat, 2018; Kokina et al., 2019).
Specifically, the AACSB 2022 standards require students to demonstrate the ability to master
current technologies and adapt to emerging technologies, including statistical techniques,
clustering, databases, modelling, analytics, text analytics, predictive analytics, learning
systems, or data visualisation. The software packages that can help with data analysis include:
business analytics, Alteryx, Power BI, machine learning, and data mining (Almeida and
Carvalho, 2022). On the other hand, there is a large volume of unstructured data that artificial
intelligence cannot yet process. Thus, accountants will have to develop their competencies in
areas where artificial intelligence cannot yet manifest itself: creativity, understanding the
context, tasks that require the use of experience and intuition (Shi, 2020).
Effective interpersonal communication, through various channels, is an essential competency,
as accountants will be required to analyse and interpret large data sets, convey the story the data
tells and facilitate good decision-making by management (De Villiers, 2021; Elo et al. al.,
2023). The growing importance of meta-competencies such as: proactive thinking, self-control,
creativity and interdisciplinary action is highlighted (Pargmann et al., 2023). A characteristic
of these new technologies is the high speed of change. Therefore, the accounting professional
must demonstrate flexibility, adaptability, responsiveness, and agility throughout its
professional career (Pargmann et al., 2023; Zhang et al., 2023). This can be achieved through
vocational training and lifelong learning (De Villiers, 2021; Jackson, Michelson and Munir,
2023; Landsberg and van den Berg, 2023). In an AI-based accounting environment, the
accountant's understanding and knowledge of business ethics is required to assess whether AI-
based suggestions comply with accounting standards and ethical principles (Leitner-Hanetseder
et al., 2021). Disruptive technological advances will require interaction competencies,
including those related to conflict management, team roles, norms, and ethics (De Villiers,
2021; Gonçalves, da Silva and Ferreira, 2022).
Vol. 26 • No. 65 • February 2024 285

The accountant knows that anticipating the impact of technology and understanding business
processes based on technology are essential to providing accounting services that contribute
to added value (Ciurea and Man, 2020). The accountant must constantly evaluate the AI
solutions on the market and propose to the management the selection of the best solutions for
the organisation (Damasiotis et al., 2015). The development of real-time accounting, which
has to cope with the presence of big data and the new dimension of intellectual capital,
requires an accountant who is not only an expert in the field of accounting, but also has
experience as a business professional (Onyshchenko et al., 2022; Landsberg and van den
Berg, 2023). Accountants are the link between managers and AI applications used in
accounting practice (Richardson and Shan, 2019).
4.2. Supporting the accounting professionals to reconfigure their competencies

Universities are indicated, by the bibliographic references consulted, as being the main actor
involved in the training of accounting professionals in the field of new technologies. A number
of 52 articles refer directly to the modalities to support the accounting professionals, and 26 of
them point to universities as the sole responsible in this regard. Analysing the referenced papers,
the conclusion emerged that adapting the curriculum to the new requirements of the labour
market is indicated by the overwhelming majority of researchers as the main measure that must
be adopted by the universities (Al-Hattami, 2021; Banasik and Jubb, 2021; Gonçalves, da Silva
and Ferreira, 2022; Karcioglu and Binici, 2023). In addition, we consider that it is worth
mentioning the opinions of some authors, who discuss the need for professional development
of teaching staff and the infrastructure development in universities, in relation with the new
technological evolutions (Aldredge, Rogers and Smith, 2021; Birt, Safari and de Castro, 2023).
Professional associations and regulatory authorities are considered important actors in the
process of adapting the skills associated with professional accountants. It starts from the idea
that professional associations must provide relevant continuous professional development
programs, the most effective being considered online courses and webinars (Jackson,
Michelson and Munir, 2023), thus complementing the courses and skills provided by
universities. It is also considered that a close relationship between universities and these
bodies facilitates the adequate training of accountants in the field of new technologies
through the effect of multiplying the effects obtained (Fredo et al., 2023). At the same time,
we consider that it is worth mentioning the ideas which suggest the involvement of
professional associations in curriculum design, alongside universities, and this can facilitate
students’ understanding of contemporary practices, including technological trends (Jackson,
Michelson and Munir, 2023). In this extensive process, some authors (Stoica and Ionescu-
Feleagă, 2021) also suggest the involvement of regulatory bodies, by updating the legislation.
The institutional environment, a concept in which we have framed the companies that hire
accountants, including accounting firms, and other organisations that, together with
universities and professional associations, work together to establish the requirements for the
training and improvement of accounting professionals, must cooperate with the academic
environment and professional associations in the matter of professional training and
development. The forms of collaboration recommended by the literature (Osmani, Hindi and
Weerakkody, 2020) are the joint development of training courses, job placement clubs, career
counselling clubs, and apprenticeships, which would enhance the skills of graduates and
prepare them for the labour market. We also retain the idea of co-designing a curriculum that

AE
enables the development of necessary competencies for the graduates’ future careers
(Osmani, Hindi and Weerakkody, 2020).
4.3. The role of the educational system

Regarding the way in which graduates can acquire the knowledge and skills required in the age
of digital technologies, we identified three optics: the introduction of new courses, the addition
of learning units to already existing courses and the mixed (hybrid) approach (Banasik and
Jubb, 2021). However, in most articles a clear orientation of the authors towards one of these
possibilities was not explicitly delineated, although, where views were expressed, we identified
as the predominant option the update of existing courses with IT elements and, in this way, the
creation of interdisciplinary courses. There are also authors who support more than one variant
of the three listed (Holmes and Douglass, 2022).
Incorporating new courses or learning units into accounting curriculum requires significant
reconfiguration. The accounting education system must be flexible to keep up with changes in
real life, and curriculum changes must be based on the demands of the labour market. It is also
appreciated that the involvement of IT experts and labour market representatives is mandatory
in the curriculum revision process (Al-Htaybat, von Alberti-Alhtaybat and Alhatabat, 2018).
In this context, the researchers also mention problems that could be generated by the
introduction of new courses and learning units in an already crowded curriculum, out of
which the predominant ones are those aimed, on one hand, at the competencies of the
academic staff who have to teach the new courses or learning units (Aldredge, Rogers and
Smith, 2021), and on the other hand, the materials that should be removed from old
accounting courses to make way for IT knowledge (Pasewark, 2021). One of the proposals
encountered was the development of new integrated study programs and the formation of
academic instructors who have a professional competence exceeding the field of accounting
(Butler et al., 2021). In relation to the issue of academic instructor’s skills, when asked who
should teach the newly introduced technology courses, accounting instructors or computer
science instructors, supporters of interdisciplinary courses circulated the idea that computer
science instructors could not teach such courses, because they do not understand the
accounting principles (Cong, 2021), leaving accounting instructors to improve themselves in
the technological field. In the long run, the need for academic staff to acquire new
competencies in digital technologies (Cong, 2021) was noted, as many of the academic staff
completed their studies long time ago, when the emphasis on technology was different.
5. Challenges and future research directions
The process to reconfigure education in the field of accounting is extremely complex. Upon
this process act certain variables that the specialised literature did not provide enough space.
In our opinion, among these variables we can include the accentuated dynamics of
technological progress, the inertia of the normative framework in education, seen in relation
to rapid technological developments, the volatile nature of regulations specific to the
financial-accounting and fiscal field or even the ethical implications of using AI on the
accounting profession.
These variables can trigger a series of chain reactions with direct implications for accounting
professionals. For example, the rapid evolution of technologies combined with the repeated
changes in legislation in the field of accounting and taxation require the acquisition of new
Vol. 26 • No. 65 • February 2024 287

competencies or the updating of knowledge held by accountants. This objective implies the
involvement of accounting professionals in a perpetual process of professional development.
In their turn, the inertia of the normative framework in education and even the implications
of an ethical nature generated by the implementation of new technologies can constitute
brakes in the framework of professional development. In this context, we believe that the
study of professional resilience or the study of the influence exercised by the emotional
intelligence of accountants on their professional performance, under the conditions of
implementing the disruptive technologies, can constitute new challenges for researchers in
the field.
We consider that the study carried out allows us to conclude that researchers unanimously
agree with the central idea that the implementation of disruptive technologies, such as AI,
requires the reconfiguration of the competencies held by accounting specialists and the
adoption of measures to support this approach. However, the specialised literature does not
sufficiently investigate the aspects related to the difficulties that can be encountered within
this extensive process. Therefore, we consider that these topics require further in-depth study,
which might constitute future research directions, as it is important to identify viable
solutions to the predicted problems to be encountered in the process of redesigning
accounting education in the context of AI implementation.
Firstly, there are the collaboration forms between the academic environment and the other
interested parties. A hot field such as the use of artificial intelligence in accounting practice
calls for the involvement of both IT and accounting professionals. Firms developing AI-based
applications can send their lecturers to organisations and universities to train specialists on
how to use those applications. Firms developing AI-based applications may also partner with
universities or other organisations involved in training or professional development to co-
create courses, study programs or professional development programs. There are already
successful models in this regard, such as the SAP University Alliances in the field of ERP
applications. In this context, we consider useful that future research could answer questions
such as: who will train the trainers in the field of AI-based applications? What are the specific
ways in which the academic environment must collaborate with the other interested parties
in the curriculum adaptation process? How should the collaboration between teaching staff
and practitioners be realised for the creation of relevant teaching materials? What teaching
methods are best suited to meet the challenges of implementing AI-based applications?
Secondly, another issue that the specialised literature has not clarified, concerns the costs that
universities will have to bear when redesigning education in the field of accounting in order
to train specialists to master the new competencies required by the market. We consider that
future research must also answer the following questions: what are the costs of implementing
AI-based solutions to support the educational process in universities? Can universities
financially support the implementation of AI-based solutions in the educational process? Is
there government support to sustain academia in implementing curriculum for AI-based
competencies?
Thirdly, we appreciate that the adaptation of the legal framework in the field of education and
accounting will constitute an element that will render difficult this broad effort, to redesign the
education process in the field of accounting. It is important that future research also clarify
issues related to: what legislative changes need to be made to make it possible to implement
AI-based solutions in accounting education and practice? How will professional reasoning be
influenced following the changes made in the regulatory framework in the field of accounting?

AE
Fourthly, but not lastly, we consider that future studies should investigate the ethical
implications, that the adoption of AI-based solutions could have in accounting education and
practice.
Conclusions
The bibliometric analysis carried out enabled us to elaborate the answers to the three research
questions, based on the dominant points of view expressed in the scientific literature, which
addressed issues related to the impact on accounting professionals exerted by the
implementation of AI-based technologies. In conclusion, we consider that our paper
demonstrates a pertinent examination of this topic, which we consider to be important and
timely.
We consider our study to be rigorously grounded, but also original. Based on a systematic
literature review, this article proposes specific solutions for curriculum reconfiguration in the
field of accounting education, it identifies possible difficulties in redesigning the education
process in the field of accounting and outlines possible ways to overcome these difficulties.
Our study is based on WoS indexed research papers, which is a relevant scientific database.
However, it should not be overlooked that our analysis excluded papers from the search query
that could have been collected from other sources that are not indexed in the WoS and that
could have provided us with possible relevant elements for the research undertaken. This is
the reason why we consider this aspect a limitation of our research, which needs to be further
investigated in future research projects in the field.
Our research highlights that the implementation of AI-based systems determines the
transformation of the competencies’ set which the accounting professionals must master, as
well as the need to support these professionals in the process of reconfiguration the
competencies required by the labour market. In this context, we have identified several actors
that can play a role in this process, among them universities being considered the central
pillar. The main measure to support accounting professionals, in order to update their skills,
is represented by curriculum reconfiguration. This task is primarily assigned to academia, but
active collaboration between universities, professional bodies, and business is required for
this curriculum redesign process to be relevant.
Acknowledgment
This work was carried out within and with the support of the Centre for Interdisciplinary
Research in Economic and Social Sciences, INCESA (Research Hub for Applied Sciences),
University of Craiova.
References
AACSB, 2022. Standards for Accounting Accreditation. [online] Available at:
<https://www.aacsb.edu/-/media/documents/accreditation/accounting/standards-and-
tables/accounting2018standards-july-12022.pdf?rev=bf6ac63464c6418c8cecbfc0ebb
9816c&hash=A89D1114B12229AC828BDB363FE2B9E8> [Accesed 2 August 2023].
Vol. 26 • No. 65 • February 2024 289

Aldredge, M., Rogers, C. and Smith, J., 2021. The strategic transformation of accounting into
a learned profession. Industry and Higher Education, [e-journal] 35(2), pp.83-88.
https://doi.org/10.1177/0950422220954319.
Al-Hattami, H.M., 2021. University Accounting Curriculum, IT, and Job Market Demands:
Evidence from Yemen. SAGE Open, [e-journal] 11(2), pp. 1-14.
https://doi.org/10.1177/21582440211007111.
Al-Htaybat, K., von Alberti-Alhtaybat, L. and Alhatabat, Z., 2018. Educating digital natives
for the future: accounting educators’ evaluation of the accounting curriculum. Accounting
Education, [e-journal] 27(4), pp. 333-357. https://doi.org/10.1080/09639284.2018.
1437758.
Almeida, A.C. and Carvalho, C., 2022. The adequacy of academic curricula for digital
transformation in the accounting education. In: s.n., 17th Iberian Conference on
Information Systems and Technologies (CISTI). n.d. Madrid, Spain. [online] Available at:
<https://doi.org/10.23919/CISTI54924.2022.9820286> [Accessed 2 August 2023].
Almufadda, G. and Almezeini, N.A, 2022. Artificial Intelligence Applications in the
Auditing Profession: A Literature Review. Journal of Emerging Technologies in
Accounting, [e-journal] 19(2), pp. 29-42. https://doi.org/10.2308/JETA-2020-083.
Asonitou, S., 2022. Impediments and pressures to incorporate soft skills in Higher Education
accounting studies. Accounting Education, [e-journal] 31(3), pp. 243-272.
https://doi.org/10.1080/09639284.2021.1960871.
Banasik, E. and Jubb C., 2021. Are Accounting Programs Future‐ ready? Employability
Skills. Australian Accounting Review, CPA Australia, [e-journal] 31(3), pp. 256-267.
https://doi.org/10.1111/auar.12337.
Banța, V.C, Rîndașu, S.M, Tănasie, A. and Cojocaru, D., 2022. Artificial Intelligence in the
Accounting of International Businesses: A Perception-Based Approach. Sustainability,
[e-journal] 14(11), article no. 6632. https://doi.org/10.3390/su14116632.
Birt, J., Safari, M. and de Castro, V.B., 2023. Critical analysis of integration of ICT and data
analytics into the accounting curriculum: A multidimensional perspective. Accounting &
Finance, [e-journal] pp. 1-27. https://doi.org/10.1111/acfi.13084.
Bonzanini, O.A., Silva, A., Cokins, G. and Gonçalves, M.J., 2020. The Interaction between
Higher Education Institutions and Professional Bodies in the Context of Digital
Transformation: The Case of Brazilian Accountants. Education Sciences, [e-journal]
10(11), article no. 321. https://doi.org/10.3390/educsci10110321.
Bunea, Ș. and Guinea, F., 2023. A. Stakeholders’ Perceptions of the Vocational Competences
Acquired by Students Enrolled in Accounting Master’s Programmes in Romania.
Sustainability, [e-journal] 15(9), article no. 7406. https://doi.org/10.3390/su15097406.
Butler, M.G., Church Swanson, K., King Hoover, G. and Spencer Wheeler, A., 2021. Do
Your Students Know What They Don't Know? An Accounting Competencies Strategy.
Issues in Accounting Education, [e-journal] 36(4), pp. 207-230.
https://doi.org/10.2308/ISSUES-19-083.
Ciurea, M. and Man, M., 2020. The Accounting Profession from Romania in the Digitized
Economy. In: s.n. The 2nd International Scientific and Practical Conference “Modern
Management Trends and the Digital Economy: from Regional Development to Global
Economic Growth” (MTDE 2020), Advances in Economics, Business and Management

AE
Research. N.d. Yekaterinburg, Russia. [online] Available at: <https://doi.org/10.2991/

aebmr.k.200502.050> [Accessed 2 August 2023].
Cong, X., 2021. Research on Curriculum Construction of Big Data and Accounting under the
Background of Big Data. In: s.n., International Conference on Big Data Engineering and
Education (BDEE). N.d., Guiyang, China. [online] Available at:
<10.1109/BDEE52938.2021.00032> [Accessed 2 August 2023].
Cunha, T., Martins, H., Carvalho, A. and Carmo, C., 2022. Not Practicing What You Preach:
How Is Accounting Higher Education Preparing the Future of Accounting. Education
Sciences, [e-journal] 12(7), article no. 432. https://doi.org/10.3390/educsci12070432.
Damasiotis, V., Panagiotis, T., Santouridis, I., Nikolopoulos, S. and Tsifora, E., 2015. IT
Competences for Professional Accountants. A Review. Procedia - Social and Behavioral
Sciences, [e-journal] 175, pp. 537-545. https://doi.org/10.1016/j.sbspro.2015.01.1234.
De Villiers, R., 2021. Seven principles to ensure future-ready accounting graduates – a model
for future research and practice. Meditari Accountancy Research, [e-journal] 29(6),
pp. 1354-1380. https://doi.org/10.1108/MEDAR-04-2020-0867.
Donthu, N., Satish K., Debmalya, M., Nitesh, P. and Weng, M.L., 2021. How to conduct a
bibliometric analysis: An overview and guidelines. Journal of Business Research, [e-
journal] 133, pp. 285-296. https://doi.org/10.1016/j.jbusres.2021.04.070.
Elo, T., Pätäri, S., Sjögrén, H. and Mättö, M., 2023. Transformation of skills in the accounting
field: the expectation – performance gap perceived by accounting students. Accounting
Education, [e-journal] pp. 1-37. https://doi.org/10.1080/09639284.2023.2191289.
Fredo, A.R., Ventura da Motta. M.E, Camargo, M.E. and Camargo Priesnitz, M., 2023.
Digital transformation: the digitization of accounting. Revista De Gestao E Secretariado,
[e-journal] 14(1), pp. 681-714. https://doi.org/10.7769/gesec.v14i1.1542.
Gonçalves, M.J.A., da Silva, A.C.F. and Ferreira, C.G., 2022. The Future of Accounting:
How Will Digital Transformation Impact the Sector? Informatics, [e-journal] 9(1), article
no. 19. https://doi.org/10.3390/informatics9010019.
Grabińska, B., Andrzejewski, M. and Grabiński, K., 2021. The students’ and graduates’
perception of the potential usefulness of Artificial Intelligence (AI) in the academic
curricula of Finance and Accounting Courses. E-mentor, [e-journal] 5(92), pp. 16-25.
https://doi.org/10.15219/em92.1544.
Gușe, G.R. and Mangiuc, M.D., 2022. Digital Transformation in Romanian Accounting
Practice and Education: Impact and Perspectives. Amfiteatru Economic, [e-journal]
24(59), pp. 252-267. https://doi.org/10.24818/EA/2022/59/252.
Holmes, A.F. and Douglass A., 2022. Artificial Intelligence: Reshaping the Accounting
Profession and the Disruption to Accounting Education. Journal of Emerging
Technologies in Accounting, [e-journal] 19(1), pp. 53-68. https://doi.org/10.2308/JETA-
2020-054.
Jackson, D., Michelson, G. and Munir, R., 2023. Developing accountants for the future: new
technology, skills, and the role of stakeholders. Accounting Education, [e-journal] 32(2),
pp. 150-177. https://doi.org/10.1080/09639284.2022.2057195.
Junger da Silva, R., Tommasetti, R., Zaidan Gomes, M. and da Silva Macedo, M.A., 2021.
Accountants' IT responsibilities and competencies from a student perspective. Higher
Vol. 26 • No. 65 • February 2024 291

Education, Skills and Work-Based Learning, [e-journal] 11(2), pp. 471-486.

https://doi.org/10.1108/HESWBL-02-2020-0028.
Kamble, S.S, Gunasekaran, A. and Gawankar, S.A., 2018. Sustainable Industry 4.0
framework: A systematic literature review identifying the current trends and future
perspectives. Process Safety and Environmental Protection, [e-journal] 117, pp. 408-425.
https://doi.org/10.1016/j.psep.2018.05.009.
Karcioglu, R. and Binici, F.O., 2023. Developing a maturity model to identify digital skills
and abilities of accounting professionals: evidence from Turkey. Access to science,
business, innovation in digital economy, [e-journal] 4(2), pp. 221-247.
https://doi.org/10.46656/access.2023.4.2(6).
Kokina, J. and Davenport, T.H., 2017. The Emergence of Artificial Intelligence: How
Automation Is Changing Auditing. Journal of Emerging Technologies in Accounting,
[e-journal] 14(1), pp. 115-122. https://doi.org/10.2308/jeta-51730.
Kokina, J., Gilleran, R., Blanchette, S. and Stoddard, D., 2019. Accountant as Digital
Innovator: Roles and Competencies in the Age of Automation. [online] Available at:
<http://dx.doi.org/10.2139/ssrn.3449720> [Accessed 2 August 2023].
Kroon, N., Alves, M.d.C. and Martins, I., 2021. The Impacts of Emerging Technologies on
Accountants’ Role and Skills: Connecting to Open Innovation - A Systematic Literature
Review. Journal of Open Innovation: Technology, Market, and Complexity, [e-journal]
7(6), article no. 163. https://doi.org/10.3390/joitmc7030163.
Landsberg, E. and van den Berg, L., 2023. 4th Industrial Revolution skills in the current South
African accountancy curricula: A systematic literature review. South African Journal of
Accounting Research, [e-journal] 37(3) pp. 177-201. https://doi.org/10.1080/
10291954.2023.2172833.
Leitner-Hanetseder, S., Lehner, O.M., Eisl, C. and Forstenlechner, C., 2021. A profession in
transition: actors, tasks and roles in AI-based accounting. Journal of Applied Accounting
Research, [e-journal] 22(3), pp. 539-556. https://doi.org/10.1108/JAAR-10-2020-0201.
Li, Z. and Zheng, L., 2018. The Impact of Artificial Intelligence on Accounting. Proceedings
of the 2018 4th International Conference on Social Science and Higher Education
(ICSSHE 2018), [e-journal] 181, pp. 813-816. https://doi.org/10.2991/icsshe-
18.2018.203.
Moore, W.B. and Felo, A., 2022. The evolution of accounting technology education:
Analytics to STEM. Journal of Education for Business, [e-journal] 97(2), pp. 105-111.
https://doi.org/10.1080/08832323.2021.1895045.
Onyshchenko, O., Shevchuk, K., Shara, Y., Koval, N. and Demchuk, O., 2022. Industry 4.0
And Accounting: Directions, Challenges, Opportunities. Independent Journal of
Management & Production, [e-journal] 13(3), pp. 161-195. https://doi.org/10.14807/
ijmp.v13i3.1993.
Osmani, M., Hindi, N. and Weerakkody, V., 2020. Incorporating Information
Communication Technology Skills in Accounting Education. International Journal of
Information and Communication Technology Education, [e-journal] 16(4), pp. 100-110.
http://doi.org/10.4018/IJICTE.2020100107.
Pargmann, J., Riebenbauer, E., Flick-Holtsch, D. and Berding, F., 2023. Digitalisation in
accounting: a systematic literature review of activities and implications for competences.

AE
Empirical Research in Vocational Education and Training, [e-journal] 15(1), pp. 1-37.
https://doi.org/10.1186/s40461-023-00141-1.
Pasewark, W.R., 2021. Preparing Accountants of the Future: Five Ways Business Schools
Struggle to Meet the Needs of the Profession. Issues in Accounting Education, [e-journal]
36(4), pp. 119-151. https://doi.org/10.2308/ISSUES-19-025.
Richardson, V.J. and Shan, Y., 2019. Data Analytics in the Accounting Curriculum. In: T.G.
Calderon ed., 2019. Advances in Accounting Education: Teaching and Curriculum
Innovations (Advances in Accounting Education). Vol. 23. Leeds: Emerald Publishing
Limited, pp. 67-79. https://doi.org/10.1108/S1085-462220190000023004
Shi, Y., 2020. The Impact of Artificial Intelligence on the Accounting Industry. In: Z. Xu,
K.K. Choo, A. Dehghantanha, R. Parizi and M. Hammoudeh eds., 2020. Cyber Security
Intelligence and Analytics. CSIA 2019. Advances in Intelligent Systems and Computing.
Vol. 928. Cham: Springer, pp. 971-978. https://doi.org/10.1007/978-3-030-15235-2_129.
Stoica, O.C. and Ionescu-Feleagă, L., 2021. Digitalization in Accounting: A Structured
Literature Review. In: Bucharest University of Economic Studies, Resilience and
Economic Intelligence Through Digitalization and Big Data Analytics. N.d., S.l.:
Sciendo. [online] Available at: <https://doi.org/10.2478/9788366675704-045>
Sun, T.S., 2019. Applying Deep Learning to Audit Procedures: An Illustrative Framework.
Accounting Horizons, [e-journal] 33(3), pp. 89-109. https://doi.org/10.2308/acch-52455.
Sutton, S.G., Arnold, V. and Holt, M., 2018. How Much Automation Is Too Much? Keeping
the Human Relevant in Knowledge Work. Journal of Emerging Technologies in
Accounting, [e-journal] 15(2), pp. 15-25. https://doi.org/10.2308/jeta-52311.
Tavares, M.C., Azevedo, G., Marques, R.P. and Bastos, M.A., 2023. Challenges of education
in the accounting profession in the Era 5.0: A systematic review. Cogent Business &
Management, [e-journal] 10(2), article no. 2220198. https://doi.org/10.1080/
23311975.2023.2220198.
Tsiligiris, V. and Bowyer, D., 2021. Exploring the impact of 4IR on skills and personal
qualities for future accountants: a proposed conceptual framework for university
accounting education. Accounting Education, [e-journal] 30(6), pp. 621-649.
https://doi.org/10.1080/09639284.2021.1938616.
Van Eck, N.J. and Waltman, L., 2010. Software survey: VOSviewer, a computer program
for bibliometric mapping. Scientometrics, [e-journal] 84(2), pp. 523-538.
https://doi.org/10.1007/s11192-009-0146-3.
Vanhove, A., Opdecam, E. and Haerens, L., 2023. Fostering social skills in the Flemish
secondary accounting education: perceived challenges, opportunities, and future
directions. Accounting Education, [e-journal] https://doi.org/10.1080/09639284.2023.
2208106.
Yigitbasioglu, O., Green, P. and Cheung, M.Y.D., 2023. Digital transformation and
accountants as advisors. Accounting, Auditing & Accountability Journal, [e-journal]
36(1), pp. 209-237. https://doi.org/10.1108/AAAJ-02-2019-3894.
Zhang, C., Zhu, W., Dai, J., Wu, Y., and Chen, X., 2023. Ethical Impact of Artificial
Intelligence in Managerial Accounting. International Journal of Accounting Information
Systems, [e-journal] 49, article no. 100619. https://doi.org/10.1016/j.accinf.2023.100619.
Vol. 26 • No. 65 • February 2024 293

AE Students’ Perceptions of the Use of Artificial Intelligence
in Educational Services
STUDENTS’ PERCEPTIONS OF THE USE OF ARTIFICIAL INTELLIGENCE

IN EDUCATIONAL SERVICES
Ines Djokic1 , Nikola Milicevic2* , Nenad Djokic3 , Borka Malcic4
and Branimir Kalas5
1)2)3)4)5)
University of Novi Sad, Novi Sad, Republic of Serbia.

Received: 25 September 2023
Djokic, I., Milicevic, N., Djokic, N., Malcic, B. and
Revised: 20 November 2023
Kalas, B., 2024. Students’ Perceptions of the Use of
Accepted: 13 December 2023
Artificial Intelligence in Educational Services.
Abstract
The increased significance of artificial intelligence (AI) and the possibilities to implement it
in higher education can be identified in contemporary conditions. At the same time, one
cannot neglect the importance of student perceptions regarding it, since they represent the
final users of educational services. Therefore, the focus of this study was on the
implementation of the already developed Student Conceptions of AI in Education Scale
(SCAIES) instrument on a sample of Serbian students. It is the first implementation of that
eight-factor instrument modelled as a reflective-formative hierarchical construct which can
be considered as its main contribution from the theoretical aspect. Thus, all eight factors
positively and significantly formed the use of AI in education (UAIEd) from students'
perspective, whereas the largest contribution could be attributed to factors arising from
weaknesses of traditional education, i.e. sentiment analysis in education, personalised
learning and student performance prediction. On the other hand, the lowest contribution
could be attributed to factors that could be related to privacy concerns about AI use and the
domain of control, concretely, classroom monitoring and visual analysis, and students' grade
and evaluation. Some general recommendations for higher education institutions were also
provided.
Keywords: Artificial intelligence, education, students, perceptions, services.
JEL Classification: M31, I20
*
Corresponding author, Nikola Milicevic – e-mail: [email protected]

AE
Introduction
Taking into account the great progress in computing power, increasingly sophisticated
algorithms, and the possibility of processing a large amount of data, artificial intelligence
represents a new level of technological and scientific development that will have a huge
impact on the way the world functions as we know it (Gasmi and Prlja, 2021, p. 324).
According to Pelau, Ene and Pop (2021, p. 33), it can be considered as “one of the main
paradigms of the contemporary society”. Regardless of whether people are aware of AI’s
omnipresence or not, it is already a component of daily lives and, at least, it is changing the
way people live (Jeffrey, 2020).
When it comes to AI history, it should be mentioned that the ideas of automation of human
life have appeared throughout history, starting from the ancient period (for example, the
protection of Crete with bronze robot Talos) (van der Vorst and Jelicic, 2019). However, the
term artificial intelligence was first introduced by John McCarthy, an emeritus professor at
Standford, in 1955, who defined it as “the science and engineering of making intelligent
machines“ (Manning, 2020). From its inception until today, AI technology developed,
although with certain issues in the 1970s (during the Cold War), and the 1990s (due to the
use of hand-crafted rules in the expert systems), which were successfully overcome (van der
Vorst and Jelicic, 2019).
A common definition of AI is one in which it is associated with a computer system capable
of performing tasks usually related to intelligent beings (European Commission Joint
Research Centre, 2018). It is similar to the definition of the European Commission, according
to which AI relates to systems that present intelligent behaviour through the analysis of the
environment and performing an action, with certain autonomy, so that the specific task can
be achieved (Boucher, 2020). For Buabbas et al. (2023, p. 1) AI “simply means making
machines capable of simulating intelligence by giving computer human-like capabilities,
such as understanding, reasoning, and problem solving“.
There are different types of AI technology. Following the European Commission, Joint
Research Centre (2018), AI can be explained through three alternative approaches based on
data, logic, and knowledge. The first approach depends mainly on data and its availability,
and in biological terminology, these systems could be called “datavores”. The other two
approaches are set on a cognitive level, whereas logic-based AI deals with problem-solving
processes, and knowledge-based AI implements simple models of inference, claiming that
knowledge is more required than logic for effective decision making. Additionally, AI can
be classified as narrow, general, and Artificial Superintelligence (Marrone, Taddeo and Hill,
2022). The most common is narrow AI, which applies machine learning to accomplish a
certain goal. General AI is considered to be a match for humans, whereas artificial
superintelligence surpasses them.
Education is one of the many sectors that has come under the influence of artificial
intelligence. The significance of higher education belonging to that sector is widely
recognised since “Universities are the main sources of highly qualified workforce and
“knowledge”” and are “being instrumental in the development of societies and economies
based on knowledge” (Dinu, 2011, p. 343). To understand and improve the application of AI
technology for educational purposes, Artificial Intelligence in Education (AIEd) has been
drawing the attention of the scientific community for a couple of decades (Chen et al., 2020).
The research subjects were the different aspects of AIEd, including the perspectives of
Vol. 26 • No. 65 • February 2024 295

teachers (Sangapu, 2018; Pisica et al., 2023) and students (Sangapu, 2018; Kuleto et al., 2021;
Cheng et al., 2023; Idroes et al., 2023) as well.
In this paper, the focus was on students' perceptions of AI applications in the educational
process. This population was selected in mind that they are the main users of educational
services. Following the service-dominant logic approach and co-creation, students can be
identified even as customers, hereby, as active participants and co-creators of services, they
should be put in a central place of a value creation process (Bowden, 2011). According to
Celuch and Robinson (2016), for whom higher education represents a “unique experiential
service”, student involvement occurs in the academic domain and also in the total educational
experience as well; therefore, the process of creating value in the delivery of educational
service requires a better understanding of the relations between students and the institution.
Since AI technology has the potential to influence the student experience (Barrett et al.,
2019), with a better understanding of student aspects, the implementation of AI in education
could be facilitated and adjusted to their needs. Therefore, after presenting some main
features related to AIEd, attention was paid to research that included students. Therefore, the
eight-factor model was applied. It was based on an instrument that measures the use of AI in
education (UAIEd), developed by Cheng et al. (2023). The use of an already developed
instrument in the new context imposed several research questions. First, there was a need to
determine the relations of constructs within the model and conclude to which type of
hierarchical model it belonged. Second, the instrument needed to be tested for application in
the new cultural context. In addition, there were research questions that emerged from
additional theoretical and practical considerations. The third research question was related to
the comparison of the results obtained with previous research from abroad and Serbia.
Finally, there was a need to provide general recommendations (based on research findings)
to higher education institutions.
To the authors' knowledge, this is the first paper that applied the mentioned model,
particularly considering the way it was designed by using the reflective-formative approach.
This can be understood as a main theoretical contribution of the paper. Not only did we apply
the developed instrument for the first time in the new context, but we also paid attention to
its specification. It is of the great importance since “measurement model misspecification
severely biases structural parameter estimates and can lead to inappropriate conclusions
about hypothesised relationships between constructs” (Jarvis, MacKenzie and Podsakoff,
2003, p. 216). The results obtained were followed by a discussion and conclusion section
providing their consideration from the aspect of previous research, the cultural context in
which the research was conducted, and consequential recommendations.
1. The use of artificial intelligence in education

Like in many other areas, artificial intelligence has been applied in education, which has been
influenced by various factors and changes over the last few decades. Financial pressures on
universities associated with the increased number of students and larger staff and operating
costs, partly caused by the democratisation of higher education, make the use of AI
technology very attractive (Popenici and Kerr, 2017). The importance of applying artificial
intelligence in the educational sector was especially emphasised due to the COVID-19
pandemic, which resulted in a shift from traditional classroom teaching to a digital form of
learning (Maqbool, Ansari and Otero, 2021; Pantelimon et al., 2021; Mijwil et al., 2022).

AE
Using the methodology of writing narrative overviews, Chassignol et al. (2018) considered
the following main educational areas that could be influenced by artificial intelligence
content and teaching, assessment, and communication.
When it comes to content and teaching, the application of AI can contribute to its
customisation. Taking into account the heterogeneity among students, a better understanding
of their learning requirements is necessary and therefore the educational content to be
personalised to their needs (Bhutoria, 2022). As an alternative to a traditional concept,
personalised learning can be defined as an “approach in which teaching is customised
according to the needs and abilities of an individual student” (Magomadov, 2020, p. 1). There
are the so-called intelligent tutoring systems (ITS) that can offer step-by-step tutorials
adjusted to each individual student; some of them are Spark, developed by Domoscio, and
Gooru Navigator, known as Google Maps for learning (Holmes and Tuomi, 2022). When
talking about personalised education, a special place belongs to the Watson Education
Classroom, developed by IBM, which can help teachers gain deeper insight into the learning
process and the ability of each student (European Commission Joint Research Centre., 2018).
To increase the quality of education, personalised learning methods can be used in
conjunction with certain gamification techniques and elements, particularly those associated
with leaderboards and points (Chassignol et al., 2018). Digital games-based learning,
Augmented Reality (AR) and Virtual Reality (VR) simulations can be applied for educational
purposes as well (Holmes and Tuomi, 2022). Potentially, with the use of artificial
intelligence, every student could have her/his own personalised learning companion, which
besides the roles of a learning partner, guide, and instructor, would continuously record the
student's improvements and interests (Holmes, Bialik and Fadel, 2019).
The potential benefits of personalised learning are numerous. Not only can its models reduce
student dropout rates and support the learning process, but they can also help to achieve
sustainable development goals (SDGs) presented by the UN General Assembly, such as those
related to ensuring equal quality education for all (SDG4) and reducing inequalities between
online and on-site approaches with special emphasis on students with disabilities (SDG 10)
(Furini et al. 2022). However, there is a criticism of personalised learning, according to which
this technology is redundant for meaningful learning and highly motivated students, who
obtain all the necessary information by themselves (Chassignol et al., 2018).
Another area of education that was affected by artificial intelligence relates to student
assessment. The development of AI technology has enabled partial or full automation of
assessment practice, whereby it can be used for generating tasks, identifying adequate peers
for grading, and automatically scoring student work (Swiecki et al. 2022). For example, there
are systems called autograders, usually applied for the assessment of written tasks, and in
mathematics and computer sciences, that are capable not only of scoring students' answers
but also of diagnosing the type of error and suggesting its correction (Holmes and Tuomi,
2022). Chassignol et al. (2018) point to ITS which besides tutoring, can help in identifying
learning gaps in students’ works. Hereby, the models created by ITS and used for grading
students’ performance are based on the rules derived from previous teacher's evaluations of
a sample set of students. The comparative study of Hooda et al. (2022) showed that among
various artificial intelligence, machine learning, and learning analytics techniques for the
assessment and provision of feedback to students, the Improved Fully Connected Network
(I-FCN) had the highest performance in accuracy, precision, recall rate, and F1-score.
Vol. 26 • No. 65 • February 2024 297

The use of AI can facilitate the assessment process for teachers by saving their time and effort
(Holmes and Tuomi, 2022). Its techniques, such as electronic assessment platforms (EAPs)
and stealth assessment, can provide continuous tracking of students' performance, while
computerised adaptive testing systems (CATs) enable adjusting tasks according to students’
abilities, rather than offering the same questions to all students (Swiecki et al., 2022). Despite
the mentioned benefits, it should be noted that the application of AI in the assessment process
related to automatic scoring, especially in the case of high-stakes testing, represents “one of
the two high-risk use cases in the proposed EU AI Act, and so would be regulated by its
provisions“ (Holmes and Tuomi, 2022, p. 556).
AI technology influenced the area of communication in the educational process, with a focus
on providing adequate feedback to students (Chassignol et al., 2018). According to
Wongvorachan et al. (2022, p. 95), feedback represents „a crucial component of student
learning“ that enhances the level of their understanding. Communication between teachers
and students is especially important under online conditions, despite the fact that this type of
interaction significantly affects student satisfaction and performance in the learning process
(Seo et al., 2021). Among AI technologies that may affect this communication are, already
mentioned, ITS and CATs (Wongvorachan et al., 2022). Seo et al. (2021) gave an example
of an AI teaching assistant called “Jill Watson” that can improve teacher-student
communication by providing autonomous responses, announcements and answers to
common questions. There is also an example of the network orchestrator named “Open
Tutor”, which enables students to connect with other human tutors via an OT mobile app, to
better understand certain topics (Holmes and Tuomi, 2022). In addition, AI techniques can
be used for collaborative learning, i.e., an educational approach that includes communication,
and working in a team to realise set tasks (Tan, Lee and Lee, 2022; Ramadevi et al., 2023).
The benefits of applying AI techniques in the process of communication can be reflected in
the improvement of both its quantity and quality, through the efficient provision of feedback
and just-in-time support to students at scale with minimal space and time barriers and with
the possibility of processing a large amount of educational data (Seo et al., 2021;
Wongvorachan et al., 2022). However, the research of Seo et al. (2021) revealed the existence
of concerns among students and teachers that AI systems may result in responsibility, agency,
and surveillance problems.
As the use of AI can have many implications for almost all stakeholders in the educational
process, including both advantages and disadvantages, an important area of its
implementation refers to the ethical question (Holmes and Tuomi, 2022). Thus, the focus
should be on the protection of human rights (Holmes and Tuomi, 2022), the control of AI
developments (Popenici and Kerr, 2017), the responsibility of AI actions, their creators and
operators (Pisica et al., 2023), the digital divide and power relations between teachers,
students, and AI technology (Hwang et al., 2020).
2. Student perceptions of artificial intelligence

In a number of investigations, artificial intelligence was analysed in the context of students.
Among them are those in which the emphasis was on students' perceptions of AI, including
their attitudes, beliefs, and/or fears. Such was the research of Gherheş and Obrad (2018),
which involved undergraduate students. According to the results obtained, the majority of
them believed that they had a below-average level of information about AI. A significant

AE
number of respondents had a positive attitude towards the development of artificial

intelligence, and more than half of them believed that it will positively affect society. On the
other hand, the probability of humankind being destroyed by intelligent devices and the
disappearance of workplaces were the greatest fears associated with the emergence of AI.
A general positive perception of college students regarding artificial intelligence was also
found in the study by Jeffrey (2020). When it comes to the level of understanding of the
meaning of AI by the student and information about its current development, the largest
number of respondents fell in the moderately high and moderately low range, respectively.
Furthermore, the research findings indicated the existence of both affirmative beliefs related
to the positive impacts of AI on society and well-being and concerns related to the rapid
developments in AI, replacing human jobs and surpassing human intelligence. Thus, it was
proven that those perceptions were affected by the student's level of information about AI.
The focus of some studies was students' AI learning intention. Wang et al. (2022) examined
this variable in the context of AI anxieties (learning and job replacement), learning
motivations (intrinsic and extrinsic), and learning self-efficacy, taking university students as
participants. In addition to the positive impacts of learning self-efficacy and learning
motivations on AI learning intention, this research highlighted the importance of AI
anxieties, especially the learning one, which negatively influenced the two motivation
variables. Besides university students, the intention or readiness to learn AI was analysed
among secondary school students (Chai, Wang and Xu, 2020) and elementary school students
(Dai et al., 2020).
There are studies that examined artificial intelligence from the point of view of students with
special reference to the educational process. Students’ perception regarding the use of
artificial intelligence in education was examined in the research of Idroes et al. (2023). The
results of their survey conducted among undergraduate students pointed to the existence of a
generally positive perception of the usefulness of artificial intelligence in education.
Furthermore, the perceived advantages and disadvantages of AI in education were analysed.
In terms of the teaching process, the main advantage perceived by the students was the virtual
assistant. In the case of learning, the most important advantage was universal access, while
when considering the evaluation process, most students identified constant feedback as the
main benefit. The main disadvantage of AI in the educational process was found to be the
lack of a relationship between students and their teachers.
To develop an instrument to evaluate undergraduate students’ conceptions of artificial
intelligence in education, Cheng et al. (2023) conducted a study that resulted in eight AI
factors: intelligent tutoring system (AI1) – the use of AI to offer learning experiences to
students in the appropriate time and place; student grading and evaluation (AI2) – the use of
AI to grade and evaluate student performance; students’ retention and dropout of students
(AI3) – the application of AI in developing intervention and warning systems intended for
at-risk students so that adequate support could be provided in a timely manner; sentiment
analysis in education (AI4) – the application of AI techniques to detect student opinions
(negative and positive) concerning their learning experiences; recommendation systems
(AI5) – the use of AI in making helpful suggestions to students; classroom monitoring and
visual analysis (AI6) – the application of AI in assessing students’ face-to-face class
attendance with the emphasis on their emotions and engagement; personalised learning (AI7)
– the use of AI technologies, usually based on learner data, in creating learner profiles that
can be applied for providing timely leraning support, curricular pathways and feedbacks; and
Vol. 26 • No. 65 • February 2024 299

student performance prediction (AI8) – the application of AI models in anticipating students’

future performance. Based on these factors, the Student Conceptions of AI in Education Scale
(SCAIES) was developed.
Kuleto et al. (2021) paid attention to artificial intelligence and machine learning (ML) taking
into account their opportunities and challenges in higher education institutions (HEI). In addition
to content analysis, they implemented the survey among students; the results of the regression
model have shown that the enhancement of personalised learning driven by AI and ML can be
performed through the development of student skills, the provision of a collaborative learning
environment in the HEI, and the development of an accessible research environment.
The study by Sangapu (2018) explored the students' and teachers' perceptions of AI use in
the classroom, relying on online surveys. When it comes to students, half of the respondents
perceived artificial intelligence as useful, while the lowest percent of them felt that it was
harmful. Approximately two-thirds of the students had a positive perception of the teaching
pattern in which a teacher uses AI as a support, while nearly eighty percent of them had
negative responses about the complete replacement of teachers by robots. More than 50% of
the students received positive feedback that the use of AI by teachers would improve their
teaching performance. On the other hand, more than half of them thought that it was not
possible for robot classrooms without teachers to fill the gaps in teaching performance.
When analysing the use of AI in education (UAIEd) from the perspective of students, eight
factors proposed by Cheng et al. (2023) were considered. Bearing in mind that the mentioned
factors represent latent phenomena, partial least squares structural equation modelling (PLS-
SEM, also known as PLS path modelling (Hair et al., 2017, p. 31)) was applied. Following
Hair et al. (2017), SEM refers to second-generation advanced statistical techniques, which
use can help overcome the weaknesses of first-generation techniques (such as logistic
regression, multiple regression, and cluster analysis). Representing a combination of
regression and factor analysis, SEM is suitable for examining unobservable variables
(constructs) indirectly measured by one or more manifest variables (indicators) (Hair et al.,
2017). Moreover, PLS-SEM enables the analyses of hierarchical (higher-order) constructs,
which implementation leads to better theoretical parsimony and the reduction of model
complexity (Becker, Klein and Wetzels, 2012). Hereby, the UAIEd was presented as a
multidimensional latent variable that was modelled as a hierarchical construct. When it
comes to the construct mode (Crocetta et al., 2021), the AI factors were set as reflective
lower-order dimensions, while UAIEd was set as a formative higher-order construct, that is,
it was “formed” by AI factors. The relations of the items with corresponding dimensions can
be considered as reflective in accordance to Jarvis, MacKenzie and Podsakoff (2003) since
those items are actually manifestations of the dimension and changes in them should not
cause changes in the dimensions (but opposite); the items have similar content/share a
common theme and dropping one of them should not alter the conceptual domain of the
dimension; finally, a change in one of the items is associated with changes in another and
they have the same antecedents and consequences. On the other hand, the relations of the
dimensions of their underlying higher-order construct correspond to formative relations (as
suggested by Jarvis, MacKenzie and Podsakoff, 2003). Concretely, those dimensions are
defining characteristics of the construct and changes in them cause changes in it (not

AE
opposite); those dimensions do not have similar content, nor share a common theme, and
dropping one of the dimensions will alter the conceptual domain of the construct; dimensions
do not necessarily covary, nor have the same antecedents and consequences. This type of
hierarchical model is very popular among researchers in regard to the application of partial
least squares structural path modelling (Cheah et al., 2019).
Figure 1 presents the research model. The meaning of abbreviations and the source of the
corresponding items can be seen in this section of the paper when the instrument was described.
Figure no. 1. Research model

The instrument applied for the purpose of the research was an anonymous questionnaire. In
addition to some demographic questions, it included statements that were evaluated on a
5-point Likert scale. For assessing eight AI factors, 37 SCAIES statements were taken from
Cheng et al. (2023) and formulated as conditional sentences: 3 statements for intelligent
tutoring system (ITS) – ITS1, ITS2, and ITS3; 4 statements for students’ grading and
evaluation (SGE) – SGE1, SGE2, SGE3, and SGE4; 4 statements for students’ retention and
dropout (SRD) – SRD1, SRD2, SRD3, and SRD4; 6 statements for sentiment analysis in
education (SAE) – SAE1, SAE2, SAE3, SAE4, SAE5, and SAE6; 4 statements for
recommendation systems (RS) – RS1, RS2, RS3, and RS4; 4 statements for classroom
monitoring and visual analysis (CMVA) – CMVA1, CMVA2, CMVA3, and CMVA4; 6
statements for personalized learning (PL) – PL1, PL2, PL3, PL4, PL5, and PL6; and
6 statements for students’ performance prediction (SPP) – SPP1, SPP2, SPP3, SPP4, SPP5,
and SPP6. Hereby, it should be noted that the repeated indicator approach (Becker, Klein and
Vol. 26 • No. 65 • February 2024 301

Wetzels, 2012) was implemented to model the higher-order construct; hence, all
37 statements (indicators) employed for measuring previously mentioned lower-order
dimensions were also used for specifying the higher-order construct – UAIEd.
For evaluating both reflective lower-order dimensions (AI factors) and higher-order
formative construct (UAIEd) in PLS-SEM, recommendations provided by Hair et al. (2017)
were used. Therefore, when it comes to reflective dimensions, several steps were taken
including the analyses of: indicator reliability (each indicator’s outer loading must be greater
than 0.70), internal consistency reliability (the value of composite reliability value (CR) for
each dimension must be greater than 0.70), convergent validity (the value of the average
variance extracted (AVE) for each dimension must be above 0.50), and discriminant validity
(the evaluation of the Fornell-Larcker and HTMT criterion). On the other hand, the
assessment of the higher-order formative construct included convergence validity (the
application of redundancy analysis based on the use of a single global item), collinearity
statistics, and path coefficients.
The convenience sample consisted of 285 students from the University of Novi Sad. When
considering the sample size, our sample exceeded the “10 times rule”, according to which
“the minimum sample size should be 10 times the maximum number of arrowheads pointing
at a latent variable anywhere in the PLS path model” (Hair et al., 2017, p. 47), which was 80
in this case. Additionally, the sample size in this research exceeded samples sizes in most
similar studies. Idroes et al. (2023) researched 91 students, Kuleto et al. (2023) had a sample
of 103 participants, while in the case of Jeffrey's research (2020), there were 230 respondents.
The size of our sample was similar to the sample of 301 participants investigated by Wang
et al. (2022). Only in two cases (Cheng et al. (2023) relied on the sample of 445 students,
while Gherheş and Obrad (2018) researched 928 students) sizes of samples from similar
research were larger (to a higher extent) than in this research.
The questionnaire was available online and professors from different faculties belonging to
the University of Novi Sad had asked students to fill in the survey on the topic that may be
of special importance for the future of, among others, higher education. Most of them were
women (55.1%). The gender structure of the sample was very similar to the gender structure
of the student population in the autonomous province of Vojvodina in which, within the state
university, i.e. University of Novi Sad, there were 23,042 female students enroled in school
year 2022/23, out of the total of 40,867 students enroled in that school year, i.e. 56.43%
(Statistical Office of the Republic of Serbia, 2023, p. 118). Their average age was 22.4. In
addition to the fact that students represent the primary users of educational services, as stated
by Gherhes and Obrad (2018) this population is well informed and will be an active
population category under the greater influence of AI development in the future. Data were
collected and processed in 2023, using the SmartPLS 4 software.

When testing the UAIEd model, the lower-order reflective dimensions were analysed first, after
which the higher-order formative construct was examined. Table no. 1. presents the values
related to the assessment of the reflective dimensions – indicator’s outer loadings, composite
reliability, and the average variance extracted – and as can be seen all assumptions are satisfied.
The meaning of abbreviations and the source of the corresponding items can be seen within the
Research methodology section of this paper when the instrument was described.

AE
Table no. 1. Reflective dimensions: Outer loadings, CR and AVE

Indicators Outer loadings CR AVE
Intelligent tutoring system - ITS 0.945 0.851
- ITS1 0.903
- ITS2 0.937
- ITS3 0.926
Student grading and evaluation - SGE 0.942 0.804
- SGE1 0.896
- SGE2 0.849
- SGE3 0.915
- SGE4 0.924
Students’ retention and dropout - SRD 0.919 0.740
- SRD1 0.923
- SRD2 0.860
- SRD3 0.903
- SRD4 0.743
Sentiment analysis in education- SAE 0.948 0.752
- SAE1 0.827
- SAE2 0.849
- SAE3 0.896
- SAE4 0.845
- SAE5 0.895
- SAE6 0.888
Recommendation systems - RS 0.930 0.768
- RS1 0.875
- RS2 0.877
- RS3 0.867
- RS4 0.885
Classroom monitoring and visual analysis - CMVA 0.916 0.731
- CMVA1 0.847
- CMVA2 0.811
- CMVA3 0.905
- CMVA4 0.855
Personalised learning -PL 0.950 0.759
- PL1 0.891
- PL2 0.858
- PL3 0.911
- PL4 0.872
- PL5 0.839
- PL6 0.855
Student performance prediction - SPP 0.949 0.755
- SPP1 0.895
- SPP2 0.883
- SPP3 0.885
- SPP4 0.849
- SPP5 0.843
- SPP6 0.859
Source: Authors
Vol. 26 • No. 65 • February 2024 303

With regard to discriminant validity, no problem was identified, as shown by the Fornell-
Larcker criterion. As can be seen in Table no. 2 the square root of each construct’s AVE was
greater than its correlations with other constructs. The meaning of abbreviations can be seen
within the Research methodology section of this paper when the instrument was described.
Table no. 2. Reflective dimensions: Fornell-Larcker criterion
CMVA ITS PL RS SAE SGE SPP SRD
CMVA 0.855
ITS 0.306 0.923
PL 0.354 0.541 0.871
RS 0.166 0.552 0.623 0.876
SAE 0.488 0.482 0.556 0.378 0.867
SGE 0.283 0.349 0.276 0.190 0.327 0.897
SPP 0.514 0.427 0.447 0.294 0.649 0.232 0.869
SRD 0.429 0.519 0.577 0.555 0.543 0.260 0.451 0.860
Source: Authors
The establishment of discriminant validity was also confirmed by the application of the
HTMT approach, which according to Henseler, Ringle and Sarstedt (2015) has superior
performance over traditional methods of its assessment. The HTMT.90 criterion was met
because the HTMT ratio for each pair of constructs was less than 0.90 (Table no 3).
Table no. 3. Reflective dimensions: HTMT ratio
CMVA ITS PL RS SAE SGE SPP SRD
CMVA
ITS 0.331
PL 0.379 0.583
RS 0.189 0.606 0.671
SAE 0.527 0.522 0.592 0.409
SGE 0.313 0.381 0.298 0.208 0.352
SPP 0.554 0.460 0.474 0.314 0.691 0.252
SRD 0.469 0.569 0.620 0.620 0.579 0.287 0.479
Source: Authors
When it comes to the higher-order formative construct (UAIEd), the results of the redundancy
analysis (a path coefficient had a magnitude greater than 0.70 (0.711) and R2 was greater than
0.50 (0.506)) confirmed its convergent validity. For examining path coefficients, a
bootstrapping procedure with 5000 sub-samples was applied. As can be seen in table no. 4,
all path coefficients were positive and statistically significant with p<0.001, whereby VIF
values below 3 indicated that there were no collinearity issues. In addition, it should be
mentioned that due to the application of the repeated indicator approach, the value of R2 for
the UAIEd construct equalled 1.
When it comes to the results of the model, as already suggested, all lower-order constructs
form a higher-order construct positively and significantly. The largest path coefficients (all
above 0.2) can be attributed to sentiment analysis in education, personalised learning, and
prediction of student performance. The next group of first-order constructs have path
coefficients ranging from 0.123 to 0.152 and here belong (in the following order): students’
retention and dropout of students, recommendation systems, and intelligent tutoring system.

AE
Finally, two constructs with the smallest values of path coefficients are classroom monitoring
and visual analysis, and students’ grading and evaluation.
Table no. 4. Path coefficients and VIF values
Path
Relations Standard deviation T statistics P values VIF value
coefficient
CMVA → UAIEd 0.119 0.010 11.835 p<0.001 1.563
ITS → UAIEd 0.123 0.008 15.333 p<0.001 1.840
PL → UAIEd 0.250 0.010 25.818 p<0.001 2.192
RS → UAIEd 0.136 0.010 14.183 p<0.001 2.029
SAE → UAIEd 0.256 0.011 23.878 p<0.001 2.263
SGE → UAIEd 0.098 0.015 6.598 p<0.001 1.211
SPP → UAIEd 0.235 0.012 19.681 p<0.001 1.956
SRD → UAIEd 0.152 0.007 21.935 p<0.001 2.007
Source: Authors
Before discussing the obtained results, two remarks should be made. Firstly, the instrument
used in this research (Cheng et al., 2023) is rather new and has not been implemented in any
other studies so far. Therefore, the space for direct comparison of the obtained results with
results from other studies is rather limited. Second, the authors of the implemented instrument
focus on its development and present only mean values for each of the factors. In this
research, since the implementation of the instrument is in its focus, the authors consider the
SCAIES as a second-order reflective-formative construct. The justification for such an
approach can be found in literature dealing with that topic (Jarvis, MacKenzie and Podsakoff,
2003) and can be considered as a specific contribution of this paper due to the large presence
of construct misspecification within the published research. Therefore, it additionally limits
the comparison of the obtained results with those obtained when the instrument was
developed. However, both remarks do not prevent the authors from considering the results
from the point of view of their meaning and the specificities of the country in which the
research was conducted.
If we focus on the three most important factors in student perceptions of the use of AI in
higher education, we could attempt to find out if there is something common among them. It
can be seen that they are related to detecting the opinions of students (negative and positive)
considering their learning experiences; providing timely learning support, curricular
pathways, and feedback; and anticipating their future performance. All of them may be
related to the overcoming of different aspects of challenges arising from classical education,
including working with large groups of students and not providing them feedback
continuously. It is all in line with one previous research on student perception of higher
education in Serbia (Rodic Lukic, 2015), indicating that students do not receive as high a
level of quality as they expect when it comes to, among others, gaining personal attention,
understanding their own specific needs, and reacting promptly to their needs.
On the other hand, two factors that contribute the least to students’ perceptions of AI use in
higher education are related to assessing the attendance of students' face-to-face classes with
an emphasis on their emotions and participation, as well as grading and evaluating student
performance. Both of those factors belong to the control area and could be connected to
anxiety about AI. Namely, AI anxiety is the fear or agitation associated with AI and may
Vol. 26 • No. 65 • February 2024 305

affect learning motivation (Wang et al., 2022) and other variables, such as attitude toward
using AI and intention to learn AI (Chai, Wang and Xu, 2020).
When considering the obtained results, two other aspects should be emphasised. Namely, in
this research students perceptions are measured. The authors of the original questionnaire
also stress that these perceptions may be influenced by students’ pre-faculty educational
experiences or “other views of AI found in movies, television shows, news, social media, and
the like“ (Cheng et al., 2023, p. 303). In addition, AI literacy can be mentioned, i.e., the level
of knowledge and skills related to AI (Dai et al., 2020). It can be considered a significant
antecedent of students' attitudes and perceptions regarding AI (Chai, Wang and Xu, 2020).
Conclusions
The development of artificial intelligence is changing the world by affecting almost every
segment of human life and work. Based on the process of “making intelligent machines” it
was supposed to help people in different sectors and activities. Therefore, the implementation
of AI technology has drawn the attention of researchers in various scientific fields. One of
them refers to education, which represents a sector of special importance for the entire
society. Taking into account the sensitivity of the educational sector, the application of AI
systems should be approached with great caution.
To get deeper insights into this field of interest, the paper examined the use of artificial
intelligence in education. In its theoretical part, the focus was on certain educational areas
(content and teaching, assessment, and communication) that can be affected by AI
technology. In addition, the subject of the analysis was the perceptions of the students about
artificial intelligence. Therefore, in addition to their general attitudes and beliefs towards AI
technology, perceptions regarding the educational process were also analysed.
In the empirical part of the paper, the hierarchical model related to students' perceptions
regarding the use of AI in education was assessed. The model was based on the SCAIES
scale, where its eight factors were presented as reflective lower-order dimensions forming
the higher-order UAIEd construct. It can be understood as the main theoretical contribution
of the paper. Following the results obtained, all dimensions positively and significantly
contributed to the UAIEd. Therefore, for students, particularly significant domains were
sentiment analysis in education, personalised learning, and prediction of student
performance, while classroom monitoring and visual analysis and student rating and
evaluation were the ones with the lowest contribution.
The results obtained suggest that the largest contribution can be attributed to factors arising
from weaknesses of traditional education, while the lowest contribution is to factors that
could be connected to privacy concerns regarding AI use and the domain of control. When
considering the results in line with Cheng et al. (2023), the results obtained must be
interpreted from the aspect of implications for higher education institutions. Therefore, the
use of AI is supported in areas where students lack personal attention, understanding of their
own specific needs, and prompt response to their needs. Even before implementing AI,
management representatives in higher education institutions should improve the quality of
this aspect of service. On the other hand, if they want to implement AI in higher education,
institutions should address all fears students express about privacy, control, or lack of human
interaction. These suggestions present a practical contribution to the research.

AE
Similar to the study of Cheng et al. (2023), the limitation of this research refers to the
insufficient knowledge of students' experiences associated with artificial intelligence.
Therefore, future studies could be expanded with new variables, such as AI literacy and/or
AI anxiety.
References
Barrett, M., Branson, L., Carter, S., DeLeon, F., Ellis, J., Gundlach, C., and Lee, D., 2019.
Using Artificial Intelligence to Enhance Educational Opportunities and Student Services
in Higher Education. Inquiry: The Journal of the Virginia Community Colleges, 22(1),
pp. 1-10.
Becker, J.-M., Klein, K. and Wetzels, M., 2012. Hierarchical Latent Variable Models in PLS-
SEM: Guidelines for Using Reflective-Formative Type Models. Long Range Planning,
45(5-6), pp. 359-394. https://doi.org/10.1016/j.lrp.2012.10.001.
Bhutoria, A., 2022. Personalized education and Artificial Intelligence in the United States,
China, and India: A systematic review using a Human-In-The-Loop model. Computers
and Education: Artificial Intelligence, 3, pp. 1-18. https://doi.org/10.1016/
j.caeai.2022.100068
Boucher, P., 2020. Artificial intelligence: How does it work, why does it matter, and what
can we do about it? [pdf] European Parliamentary Research Service. Available at:
<https://www.europarl.europa.eu/RegData/etudes/STUD/2020/641547/EPRS_STU(202
0)641547_EN.pdf> [Accessed 7 July 2023].
Bowden, J.L-H., 2011. Engaging the Student as a Customer: A Relationship Marketing
Approach. Marketing Education Review, 21(3), pp. 211-228. https://doi.org/10.2753/
MER1052-8008210302.
Buabbas, A.J., Miskin, B., Alnaqi, A.A., Ayed, A.K., Shehab, A.A., Syed-Abdul, S. and
Uddin, M., 2023. Investigating Students’ Perceptions towards Artificial Intelligence in
Medical Education. Healthcare, 11, pp. 1-16. https://doi.org/10.3390/healthcare
11091298
Celuch, K. and Robinson, N.M., 2016. How the Customer Feedback Process Contributes to
Perceived Customer Orientation and Affective Commitment in the Higher Educational
Service Context. The Journal of Consumer Satisfaction, Dissatisfaction and Complaining
Behavior, 29, pp. 53-76.
Chai, C.S., Wang, X. and Xu, C., 2020. An Extended Theory of Planned Behavior for the
Modelling of Chinese Secondary School Students’ Intention to Learn Artificial
Intelligence. Mathematics, 8(11), pp. 1-18. https://doi.org/10.3390/math8112089.
Intelligence trends in education: a narrative overview. Procedia Computer Science, 136,
pp. 16-24. https://doi.org/10.1016/j.procs.2018.08.233.
Cheah, J-H., Ting, H., Ramayah, T., Memon, M.A., Cham, T-H. and Ciavolino, E., 2019. A
comparison of five reflective – formative estimation approaches: reconsideration and
recommendations for tourism research. Quality & Quantity, 53, pp. 1421-1458.
https://doi.org/10.1007/s11135-018-0821-7.
Vol. 26 • No. 65 • February 2024 307

Chen, X., Xie, H., Zou, D. and Hwang, G-J., 2020. Application and theory gaps during the
rise of Artificial Intelligence in Education. Computers and Education: Artificial
Intelligence, 1, pp. 1-20. https://doi.org/10.1016/j.caeai.2020.100002.
Cheng, L., Umapathy, K., Rehman, M., Ritzhaupt, A., Antonyan, K., Shidfar, P., Nichols, J.,
Lee, M. and Abramowitz, B., 2023. Designing, Developing, and Validating a Measure of
Undergraduate Students’ Conceptions of Artificial Intelligence in Education. Journal of
Interactive Learning Research, 34(2), pp. 275-311. https://www.learntechlib.org/
primary/p/222246/.
Crocetta, C., Antonucci, L., Cataldo, R., Galasso, R., Grassia, M.G., Lauro, C.N. and Marino,
M., 2021. Higher-Order PLS-PM Approach for Different Types of Constructs. Social
Indicators Research, 154(2), pp. 725-754. https://doi.org/10.1007/s11205-020-02563-w.
Dai, Y., Chai, C.-S., Lin, P.-Y., Jong, M.S.-Y., Guo, Y. and Qin, J., 2020. Promoting
Students’ Well-Being by Developing Their Readiness for the Artificial Intelligence
Age. Sustainability, 12(16), article no. 6597. https://doi.org/10.3390/su12166597.
Dinu, V., 2011. The Knowledge - Based Economy: Implications for Higher Education in
Economics and Business. Amfiteatru Economic, 13(30), pp. 343-344.
European Commission. Joint Research Centre., 2018. The impact of Artificial Intelligence on
learning, teaching, and education. [online] LU: Publications Office. Available at:
<https://data.europa.eu/doi/10.2760/12297> [Accessed 23 August 2023].
Furini, M., Gaggi, O., Mirri, S., Montangero, M., Pelle, E., Poggi, F. and Prandi, C., 2022.
Digital twins and artificial intelligence: as pillars of personalized learning models.
Communications of the ACM, 65(4), pp. 98-104. https://doi.org/10.1145/3478281.
Gasmi, G., and Prlja, D., 2021. Ugrožavanje ljudskih prava i veštačka inteligencija. Zbornik
radova Kopaoničke škole prirodnog prava – Slobodan Perović, pp. 323-335.
Gherheș, V. and Obrad, C., 2018. Technical and Humanities Students’ Perspectives on the
Development and Sustainability of Artificial Intelligence (AI). Sustainability, 10(9),
Hair, J.F., Hult, G.T.M., Ringle, C.M. and Sarstedt, M. 2017. A Primer on Partial Least
Squares Structural Equation Modeling. Thousand Oaks: Sage.
Henseler, J., Ringle, C.M., and Sarstedt, M., 2015. A new criterion for assessing discriminant
validity in variance-based structural equation modelling. Journal of the Academy of
Marketing Science, 43, pp. 115-135.
Holmes, W. and Tuomi, I., 2022. State of the art and practice in AI in education. European
Journal of Education, 57(4), pp. 542-570. https://doi.org/10.1111/ejed.12533.
Holmes, W., Bialik, M. and Fadel, C., 2019. Artificial intelligence in education: promises
and implications for teaching and learning. Boston, MA: The Center for Curriculum
Redesign.
Hooda, M., Rana, C., Dahiya, O., Rizwan, A. and Hossain, M.S., 2022. Artificial Intelligence
for Assessment and Feedback to Enhance Student Success in Higher Education.
Mathematical Problems in Engineering, 2022, pp. 1-19. https://doi.org/10.1155/
2022/5215722.
Hwang, G.-J., Xie, H., Wah, B.W. and Gašević, D., 2020. Vision, challenges, roles and
research issues of Artificial Intelligence in Education. Computers and Education:
Artificial Intelligence, 1, article no. 100001. https://doi.org/10.1016/j.caeai.2020.100001.

AE
Idroes, G.M., Noviandy, T.R., Maulana, A., Irvanizam, I., Jalil, Z., Lensoni, L., Lala, A.,
Abas, A.H., Tallei, T.E. and Idroes, R., 2023. Student Perspectives on the Role of
Artificial Intelligence in Education: A Survey-Based Analysis. Journal of Educational
Management and Learning, 1(1), pp. 8-15. https://doi.org/10.60084/jeml.v1i1.58.
Jarvis, C.B., MacKenzie, S.B. and Podsakoff, P.M., 2003. A Critical Review of Construct
Indicators and Measurement Model Misspecification in Marketing and Consumer
Research. Journal of Consumer Research, 30(2), pp. 199-218. https://doi.org/10.1086/
376806.
Jeffrey, T., 2020. Understanding College Student Perceptions of Artificial Intelligence.
Systemics, cybernetics and informatics, 18(2), pp. 8-13.
Kuleto, V., Ilić, M., Dumangiu, M., Ranković, M., Martins, O.M.D., Păun, D. and
Mihoreanu, L., 2021. Exploring Opportunities and Challenges of Artificial Intelligence
and Machine Learning in Higher Education Institutions. Sustainability, 13(18), article no.
10424. https://doi.org/10.3390/su131810424.
Magomadov, V.S., 2020. The application of artificial intelligence and Big Data analytics in
personalized learning. Journal of Physics: Conference Series, 1691(1), article no.
012169. https://doi.org/10.1088/1742-6596/1691/1/012169.
Manning, C., 2020. Artificial Intelligence Definitions. Stanford University: Human-Centered
Artificial Intelligence.
Maqbool, F., Ansari, S., and Otero, P., 2021. The Role of Artificial Intelligence and Smart
Classrooms during COVID-19 Pandemic and its impact on Education. Journal of
Independent Studies and Research Computing, 19(1), pp. 7-14. https://doi.org/
10.31645/JISRC.41.19.1.2.
Marrone, R., Taddeo, V. and Hill, G., 2022. Creativity and Artificial Intelligence – A Student
Perspective. Journal of Intelligence, 10(3), article no. 65. https://doi.org/10.3390/
jintelligence10030065.
Mijwil, M.M., Aggarwal, K., Mutar, D.S., Mansour, N. and Singh, R.S.S., 2022. The Position
of Artificial Intelligence in the Future of Education: An Overview. Asian Journal of
Applied Sciences, 10(2), pp. 102-108. https://doi.org/10.24203/ajas.v10i2.6956.
Pantelimon, F.-V., Bologa, R., Toma, A. and Posedaru, B.-S., 2021. The Evolution of AI-
Driven Educational Systems during the COVID-19 Pandemic. Sustainability, 13(23),
Identity and Human Skills. Amfiteatru Economic, 23(56). pp. 33-45.
https://doi.org/10.24818/EA/2021/56/33
Intelligence in Higher Education: Pros and Cons from the Perspectives of
Academics. Societies, 13(5), article no. 118. https://doi.org/10.3390/soc13050118.
Popenici, S.A.D. and Kerr, S., 2017. Exploring the impact of artificial intelligence on
teaching and learning in higher education. Research and Practice in Technology
Enhanced Learning, 12(1), article no. 22. https://doi.org/10.1186/s41039-017-0062-8.
Ramadevi, J., Sushama, C., Balaji, K., Talasila, V., Sindhwani, N., and Mukti, 2023. AI
enabled value-oriented collaborative learning: Centre for innovative education. The
Vol. 26 • No. 65 • February 2024 309

Journal of High Technology Management Research, 34(2), article no. 100478.

https://doi.org/10.1016/j.hitech.2023.100478.
Rodic Lukic, V., 2015. Service Quality in Function of Satisfaction and Behavioral Intentions
in Realization of Marketing Performance at Institutions of Higher Education in the
Western Balkans. Doctoral dissertation. University of Novi Sad. [online] Available at:
<https://www.cris.uns.ac.rs/DownloadFileServlet/Disertacija14564017456688.pdf?cont
rolNumber=(BISIS)95952&fileName=14564017456688.pdf&id=4978&licenseAccepte
d=true> [Accessed 7 July 2023].
Sangapu, I., 2018. Artificial Intelligence in Education - From a Teacher and a Student
Perspective. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.3372914.
Seo, K., Tang, J., Roll, I., Fels, S. and Yoon, D., 2021. The impact of artificial intelligence
on learner – instructor interaction in online learning. International Journal of Educational
Technology in Higher Education, 18(1), article no. 54. https://doi.org/10.1186/s41239-
021-00292-9.
Statistical Office of the Republic of Serbia, 2023. Statistical Yearbook оf the Republic of
Serbia. [pdf] Belgrade. Available at: <https://publikacije.stat.gov.rs/G2023/
Pdf/G20232056.pdf> [Accessed 14 July 2023].
Swiecki, Z., Khosravi, H., Chen, G., Martinez-Maldonado, R., Lodge, J.M., Milligan, S.,
Selwyn, N. and Gašević, D., 2022. Assessment in the age of artificial
intelligence. Computers and Education: Artificial Intelligence, 3, article no. 100075.
Tan, S.C., Lee, A.V.Y. and Lee, M., 2022. A systematic review of artificial intelligence
techniques for collaborative learning over the past two decades. Computers and
Education: Artificial Intelligence, 3, article no. 100097. https://doi.org/10.1016/
j.caeai.2022.100097.
van der Vorst, T., and Jelicic, N., 2019. Artificial Intelligence in Education: Can AI bring the
full potential of personalized learning to education? 30th European Conference of the
International Telecommunications Society (ITS): "Towards a Connected and Automated
Society". Helsinki, Finland, 16th-19th June, 2019, International Telecommunications
Society (ITS), Calgary.
Wang, Y.-M., Wei, C.-L., Lin, H.-H., Wang, S.-C. and Wang, Y.-S., 2022. What drives
students’ AI learning behavior: a perspective of AI anxiety. Interactive Learning
Environments, pp. 1-17. https://doi.org/10.1080/10494820.2022.2153147.
Wongvorachan, T., Lai, K.W., Bulut, O., Tsai. Y-S. and Chen, G., 2022. Artificial
Intelligence: Transforming the Future of Feedback in Education. Journal of Applied
Testing Technology, 23, pp. 95-116.

Contents
Economic Interferences
Measuring the Level of Performance of Country’s Cultural Economics:
The Case of EU11 ............................................................................................................. 312
Kristina Astikė and Viktorija Skvarciany
Research on Tax Compliance Incentive Effects of Platform Companies

from the Perspective of Incomplete Contract – An Empirical Study Based
on China ............................................................................................................................ 330
Xuefeng Shao and Shi Chen
Vol. 26 • No. 65• February 2024 311

AE Measuring the Level of Performance of Country’s Cultural Economics:
The Case of EU11
MEASURING THE LEVEL OF PERFORMANCE

OF COUNTRY’S CULTURAL ECONOMICS: THE CASE OF EU11
Kristina Astikė1* and Viktorija Skvarciany2
1)2)
Vilnius Gediminas Technical University, Vilnius, Lithuania

Astikė, K. and Skvarciany, V., 2024. Measuring the Received: 14 September 2023
Level of Performance of Country’s Cultural Revised: 15 November 2023
Economics: The Case of EU11. Amfiteatru Economic, Accepted: 16 December 2023
26(65), pp. 312-329.
Abstract
Culture itself forms society’s identity, social values, and formal and informal social relations.
This strengthens social capital, which is based on community, citizenship, volunteerism, and
social values. These aspects promote effective democratic governance, and thereby ensure
economic growth and, at the same time, the expansion of cultural economics. The main
purpose of this article is to identify which of the selected cultural performance factors are
more important in relation to each other and to create an index of cultural economics for the
post-communist Central and Eastern European countries of the European Union according to
these factors. The methods used in the article are as follows: analysis of scientific literature,
qualitative expert interviews process with FAHP, and multi-criteria decision-making method
TOPSIS. Two types of data were collected and analysed in the article: data received from the
expert survey and existing statistical data. The results revealed that after creating an index of
the post-communist Central and Eastern Europe countries of the European Union according
to selected cultural performance factors, the first five places are occupied by Romania,
Bulgaria, Slovakia, Latvia, and Lithuania. The practical implications of the research are that
cultural economics is related to both the private and public sectors and, as a result, to their
revenue; hence, the products it produces may be stated to contribute to the country’s economy
and development. The limitations of this study are statistics that have been used; the data is
for 2020 as this is the most recent available.
Keywords: cultural economics, heritage, cultural institutions, EU11
JEL Classification: C83, D1, H1, Z1
*
Corresponding author, Kristina Astikė – e-mail: [email protected]

Economic Interferences AE
Introduction
Culture is decisive in national progress, individual and community self-worth, and various
forms of creativity and innovation. Without culture, modern development of political
democracy, economic growth, society’s ability to think critically, social harmony, and the
overall well-being of every citizen are impossible. This view is supported by the cultural
theories. Considering these theories, it is essential to mention Geert Hofstede’s research, in
which he established a conceptual framework to comprehend variations in culture among
diverse nations and societies. Initially focused on identifying and measuring cultural
disparities in the workplace, the findings of this research have since been extrapolated to
encompass broader contexts (Hofstede et al., 2004). Hofstede’s cultural theories are closely
related to this research because they explore cultural institutions, cultural employment, and
professional training in the culture sector. All of this provides a framework for understanding
cultural differences and these dimensions are closely related to cultural institutions. Cultural
institutions are the structures and mechanisms within a society that shape the behaviour,
values, and norms of its members. The dimensions identified by Hofstede help to highlight
how certain aspects of cultural institutions vary across different societies. In various aspects,
culture occupies an important place in the modern world. Culture fosters the increase in the
value and competitiveness of products (Goffi et al., 2019; Azeem et al., 2021), and promotes
social inclusion and social development (Meir and Fletcher, 2019; Valverde-Moreno et al.,
2021). Culture is essential for spatial planning and urban expansion (Grenni et al., 2020;
Purkarthofer et al., 2021), and it also serves for the development of collective awareness and
thinking (Caron et al., 2020; Chabay, 2020). Culture is also influential from an economic
approach. The World Economic Forum announces that culture is seen as an essential sector
that develops human capital, promotes the creation of new innovative products, develops
infrastructure and tourism, and contributes to the digitisation of processes (World Economic
Forum, 2020). Hence, it could be stated that culture is a part of different systems, and
economics is not an exception. In order to include culture in the economics dimension, the
conception of cultural economics has been developed. The importance of cultural economics
is justified in official reports, such as United Nations documents. In 2019, the United Nations
Educational, Scientific and Cultural Organisation (UNESCO) organised a debate on Culture
and Sustainable Development. The debate highlights the connection between culture,
sustainable development, innovation, diversification, and economic expansion. These
documents illustrate how artistic education, cultural performance and culture itself, and
creative industries contribute to achieving development goals in providing up-to-date and
innovative solutions in urban and rural, national, and local contexts. In addition, they provide
opportunities for decent work, poverty reduction, social resilience, and gender equality.
United Nations argue that culture is the key to sustainable development and a powerful
contributor to economic and social stability interacting with technology, intellectual property,
and tourism objectives: it is a set of knowledge-based, and thus more localised, economic
activities with a development dimension and cross-cutting linkages at macro and micro levels
to the overall economy. All this shows the relevance of cultural economics. Based on that, it
could be argued that in order to reach the highest level of cultural economics, the present
situation should be assessed. Because of that, it is crucial to set the factors that could be used
for cultural economics measurement, which will help to understand where the countries are
now. It is especially important for post-communistic countries to facilitate smoother
integration into the European cultural space in the field of economics.
Vol. 26 • No. 65 • February 2024 313

The Case of EU11
After the early 1990s, the post-communist countries of Central and Eastern Europe achieved
similar development goals. The main ones were democratisation, integration into the EU,
development of relations with other European countries, the economic and political
transformation of financial systems, and also getting involved in the cultural space of the
European Union. Thus, this research aims to identify the factors for measuring the country’s
cultural economics and evaluating its state in EU11. EU11 is the abbreviation for all post-
communist countries that are members of the European Union. These are 11 countries in
Central and Eastern Europe (Croatia, Estonia, Slovenia, Bulgaria, Romania, Latvia,
Lithuania, Slovakia, Hungary, the Czech Republic, Poland) that had a communist regime
and, after the collapse of the Soviet Union, joined the European Union. These countries have
not only common political, cultural, geographical, and economic features but also faced
similar challenges: prohibitions of religion, and native language, which suppressed their
cultural identity. The main cultural similarity of all EU11 is the communist intangible and
material heritage. This heritage does not reflect the countries’ historical past or national
identity; that is, it shows the countries’ communist ideology, which was similar in all
countries. Therefore, it is not surprising that after the collapse of the Soviet Union, all these
countries tried to free themselves from their communist heritage, and this liberation is still
ongoing today. Thus, it is relevant to analyse the factors of the cultural economics of these
countries and compare them with each other in this aspect.
The article has the following structure: the theoretical part presents the analysis of the factors
for measuring the country’s cultural economics. The research part of the article describes the
research methodology and methods: AHP and TOPSIS methods. The article ends with a
discussion and conclusions.
1. Theoretical background
Different studies have shown that cultural activities bring both direct and indirect benefits.
For example, they actively participate in civic life and thereby increase social capital. It also
contributes to faster economic development, as it encourages the improvement of the
environment around us through the strengthening of public security, which creates an
attractive investment environment. And, of course, it protects the cultural heritage and
educates about the importance of passing on and preserving cultural values and the country’s
history, which forms group memory and identity (Attanasi et al., 2013; Orlando et al., 2022).
Culture itself forms society’s identity, social values, and formal and informal social relations.
All of this strengthens social capital, which is based on community, citizenship,
volunteerism, and social values. These aspects promote effective democratic governance and,
therefore, ensure economic growth. In other words, all of the mentioned aspects are part of
the country’s cultural economics.
At the global level, UNESCO has been assessing culture in different reports and studies. For
example, its 2013 Creative Economy Report proposes a system of factors that are divided
into three groups: resources, capacity, and outcomes. The group of resource factors consists
of the following: creative workforce, creative businesses, cultural institutions, cultural
heritage, cultural activities, and cultural infrastructures (UNESCO Culture Development
Indicators, 2014). Capacity factors include data about government and private sector
participation, social capital, civil society, education in arts and culture, communication, and
media. The group of outcome factors is divided into economic, social, cultural, and

environmental parts. The group of economic factors included data on the output of cultural
goods and services, employment, exports, business development, tourism, and equity in
economic outcomes. Social outcomes are measured using data about social cohesion, cultural
diversity, human rights and non-discrimination, and education. Cultural outcomes include
cultural consumption and engagement data, cultural participation and creative activity, art-
form development, and culture in external relations. Finally, environmental outcomes are
measured using data on educational strategies, arts as an example of green practice, and
traditional knowledge (UNESCO, 2013). This shows how broad the field of cultural
economics research is and that cultural economics factors can be divided into four groups to
better understand them. Another example showing the importance of cultural economics on
a global scale is the Creative Economy Outlook 2022, prepared by the United Nations
(Creative Economy Outlook, 2022). This report highlights the unique importance of culture
for the creative industries, the fastest-growing sector in the world. In this way, culture
acquires even greater economic importance, contributing to creating jobs and promoting
innovation and international trade. By strengthening the development of the creative sector,
culture contributes to the growth of the countries’ social, political, and economic sectors
(Creative Economy Outlook, 2022). These groups show the economic importance of culture
and define how culture contributes to the economic growth of countries: through the
promotion of the growth of social capital, the expansion of innovation, creative industries,
and international trade. In addition, the economic value of culture can be measured as cultural
capital, which can be tangible: monuments (architecture, works of art, structures or findings
of an archaeological nature), ensembles (isolated or connected groups of structures whose
architecture is related to the landscape), famous places (works of people and nature) or
intangible: long-established activities, images, forms of expression, knowledge, skills, i.e.
related tools, objects, products of human activity, and cultural spaces related to them,
recognised by certain communities as part of cultural heritage. All the mentioned factors are
a part of cultural economics; hence, it is important to identify elements the of cultural
economics carefully and describe them in order to create a system for measuring the current
level of cultural economics.
Another aspect why it is crucial to determine these factors is because they affect the country’s
competitiveness and social policies and increase the country’s investment attractiveness. The
control of cultural economics lies within the hands of countries, making it crucial to identify the
pertinent factors that serve as determinants. This identification is essential for crafting relevant
policy formulations. Policy-makers could use the obtained results for the targeted development
of cultural economics, which would further contribute to the growth of the country’s economy
and its competitiveness. Identifying these factors is the primary purpose of the literature review
of the current article. After discussing the importance of the cultural economy, the following ten
factors were identified that could be used to measure cultural economics.
Cultural heritage is considered the main element in the identity and uniqueness of cities and
regions. During the Soviet era, this national identity was suppressed by the construction of
new objects reflecting the communist ideology. Such objects reflecting the regime of
communism are being dismantled today. Currently, cultural heritage is understood as heritage
with artistic value. It contributes to people’s well-being, creates jobs, increases people’s
social employment and promotes tourism, which, in turn, promotes the economic
development of a country. Many authors emphasise that it is through the development of
tourism that heritage affects the country’s economy, so it is vital to properly protect and
Vol. 26 • No. 65 • February 2024 315

The Case of EU11
manage it (Bosone et al., 2021; Chhabra, 2021; Zhenrao et al., 2021). Therefore, cultural
heritage can be singled out as a measurement factor in cultural economics.
Another important factor specified in the scientific literature as allowing the measurement of
cultural economics is cultural institutions: museums, cultural centres and institutes, theatres
and concert institutions, and libraries (Grenersen, 2012; Scott, 2019). These cultural
institutions are one of the main tools for promoting regional development and reducing
disparities among them by creating new jobs and participating in social cohesion. Cultural
institutions create conditions for the country’s residents to use cultural services and
participate in cultural life. The development of the infrastructure of cultural institutions
contributes to the development of cultural products and facilities and the formation of cultural
policy in the region (Baculáková and Grešš, 2021). Cultural institutions need a suitable
marketing strategy that is focused on solid and long-lasting terms between institutions,
audiences, and sponsors. For this, it is crucial to have a good image, products of high quality,
or products that have a reasonable proportion between price and quality (Buljubašić et al.,
2016). These relationships can stimulate the interest of the audience and, in the end,
successful marketing in cultural economics.
Today, the effect of cultural activities on GDP, economic growth, social welfare and
development of international exchanges is very noticeable as it has changed cultural activities
to one of the basic foundations of the economic system of the world (Alta and Taghva, 2017).
It is a common belief that cultural products are not crucial to a country’s economy, but they
have received increasing attention in recent years. Authors agree that this contributed to the
growing need for culture and art, the development of technology and the growing
independence of culture (Barandiaran-Irastorza et al., 2020). Culture also influences
entrepreneurship and creative projects, which create new businesses or expand existing ones.
All this justifies the contribution of cultural activities to GDP and can be used as an indicator
to measure the cultural economy.
The factors of cultural employment and professional training in the culture sector are directed
at the role of culture as an “employer” in order to understand better its influence on the
national economic and social increase (UNESCO, 2014). Participation in cultural action is
an integral part of development, which monitors the employment of people, the income
generation from job creation, and the higher standard of living of those who work ( Baluku
et al., 2019; Boyd, 2021). The development of culture in the country also requires
competitiveness from the cultural segment staff. In order to develop the economic potential
of culture, professionals in the cultural sector must have the opportunity to acquire and
develop artistic, creative, technical, technological and managerial skills and competencies
(UNESCO, 2014). Employees will not only be able to successfully use the acquired
competencies in practice, working in cultural institutions but also in representing the country
at international events (Gautam and Basnet, 2020).
Household expenditures on culture are related to economic development and show how
society values culture through financial flows. Household costs on culture also directly
contribute to the expansion of the cultural segment and encourage the development of new
cultural products (Cellini and Cuccia, 2021). Spending on culture promotes the development
of the infrastructure of cultural objects, such as theatres, museums, and galleries. It also
contributes to increasing the number of cafes, restaurants, and bars. Zhang et al. (2022)
carried out research showing that during the COVID-19 pandemic, when households’

spending on culture decreased significantly, many cultural objectives were closed. This
shows how vital household spending is to cultural economics.
The distribution of cultural infrastructures plays a key role in promoting greater access to
culture. This is important in reducing disparities between the different regions of countries
and, herewith, improving the quality of life of the citizens. However, the authors Li et al.
(2020) and Elwell et al. (2020) emphasise that a good cultural infrastructure alone is not
enough, and it is necessary to ensure that all groups of society can participate in the
management of cultural processes and participate in cultural activities.
Another vital factor is civil society’s participation in cultural governance and participation in
going-out activities, which consists of attending live performances, going to the cinema,
visiting cultural sites, and practising artistic activities (UNESCO). Public participation in
various cultural events should be a part of everyday life; such cultural commitment would
contribute to the design of better-paid labour in the cultural segment and become an incentive
for increasing cultural content (Warburton and McLaughlin, 2007), which, in turn, will
promote the development of cultural economics.
Furthermore, the last one is participation in identity-building cultural activities, which
consists of participation in identity-building cultural creative activities and developing
cultural and social competencies. According to Capello (2018), participation in such
activities contributes to the design of cultural processes, the strengthening of the country’s
identity, as well as national security. According to the author, the importance of the European
Union as a mechanism for preserving European values is perceived through identity
formation (Capello, 2018).
As can be seen from the literature analysis, the factors are comprehensive, covering both the
private and public sectors. The presented factors reflect economic indicators such as GDP,
household expenditures, and social indicators: employment and identity-building. This
shows the importance of culture in the economic sectors of the country’s governance.
2. Methodology
In order to measure the level of cultural economics of the EU11, two types of data were
collected and analysed in the article: expert survey and statistical data in order to assess which
of the selected cultural economic factors are more important in relation to each other, and to
create an index of the cultural economics of the post-communist countries that are members
of the European Union according to these factors.
The post-communist countries of the European Union, after the collapse of the Soviet Union,
developed extremely rapidly in all areas in order to restore statehood and provide better living
conditions for the country’s inhabitants, who were invited to a new economic and political
community (Fihel and Okólski, 2019). Despite all being in the European Union community,
these post-communist member countries have a number of differences, which arose not only
due to different geographical conditions, economic levels, and other structural circumstances,
but also due to the different beginnings of the creation of these nation-states (Papava, 2018).
Some of these EU11 countries lost their statehood during the Soviet period and had limited
opportunities to maintain their national and cultural identity. For the latter reason, it is
relevant to analyse how, according to selected cultural economics factors, these EU11
countries differ and what is the level of each country’s cultural economics.
Vol. 26 • No. 65 • February 2024 317

The Case of EU11
The first step of the study was the expert quantitative survey. The experts for the study were
selected based on the publications prepared on the topic of cultural economics, which were
published in the Web of Science and/or Scopus databases. All experts were sent identical
questionnaires in which they had to complete the pairwise comparison of the selected factors
distinguished from the literature. The number of respondents was selected based on the
recommendations of Libby and Blashfield (1978). The authors found out that the optimal
number of experts is between five and nine. Therefore, in the current research, the number of
experts selected is five (Table no. 1). All the experts are from different EU11 countries (the
countries are not provided due to the anonymity of the survey).
Table no. 1. General information about the experts
The period when exploring
Education degree Present work
No. the topic of cultural
and field position
economics
E1 PhD in economics 3 years Associate professor
E2 PhD in economics 2 years Professor
E3 PhD in humanities 16 years Associate professor
As can be seen from Table no. 1, all the experts interviewed in the study have doctorate
degrees and pedagogical positions, and all the experts have experience in researching cultural
economics. Such a high competence of experts shows the reliability of the obtained results.
The experts were asked to complete the pairwise comparison matrices in order to compare
the distinguished cultural economics factors (Table no. 2). The Fuzzy Analytic Hierarchy
Process (FAHP) was selected to process the data and calculate the weight of the selected
factors. FAHP was chosen for the research, as sometimes the expert evaluation is connected
to uncertainty, which could be covered by using fuzzy numbers.
Table no. 2. Description of statistical data of factors of cultural performance
Explanation
of relationship
Criterion between the
through which primary factors
Primary factors Narrative
cultural Unit of of cultural
of cultural explanation Source
economics measure economics and the
economics of factors data
factor criterion through
is measured which cultural
economics factors
is measured
Cultural heritage Properties Number The UNESCO Factor data: UNESCO
inscribed for 100000 organisation is properties database
on the population directly related to inscribed on the (2020)
UNESCO cultural heritage, as UNESCO World
World Heritage one of the main Heritage List
List. functions of this include a
organisation is the number of each
preservation and country’s
restoration of material
cultural heritage. heritage.

Explanation
of relationship
economics factors
is measured
This shows the
importance of this
criterion and its
direct connection
with cultural
heritage.
Cultural Number of Number Number of Factor data: Eurostat
institutions enterprises in for 100000 enterprises in the number of database
the cultural population cultural sector enterprises in the (2020)
sector. directly reflects cultural sector
how many cultural include a
institutions there number of each
are in a particular country’s
country. cultural
institutions/
enterprises.
Contribution of Contribution of Percentage Criterion – Factor data: Eurostat
cultural activities cultural of GDP contribution of contribution database
to GDP activities to cultural activities to of cultural (2020)
GDP. GDP directly activities to
reflects primary GDP. Shows
factor. each country’s
percentage
expression of
how much the
cultural sector
contributes to
the country’s
GDP.
Cultural Cultural Number Criterion – cultural Factor data: Eurostat
employment employment for 100000 employment by cultural database
by total age. population total age directly employment by (2020)
reflects the primary total age. Shows
factor. how many
thousands
of people work
in the cultural
sector in each
country.
Household Household Annual This is a direct Factor data: Eurostat
expenditures on expenditures on average criterion that household database
culture cultural goods. index reflects household expenditures on (2020)
spending on cultural goods.
Vol. 26 • No. 65 • February 2024 319

The Case of EU11
Explanation
of relationship
economics factors
is measured
culture, expressed Is rendered by
through an average the annual
annual index. average index
of how much a
household spend
on cultural
services per year
Distribution General Percentage Government sector Factor data: Eurostat
of cultural government of GDP expenditures by general database
infrastructures expenditure functions: government (2020)
by function: recreation and expenditure by
recreation culture are directly function:
and culture. related recreation and
to the development culture. Shows
of cultural how much
infrastructure. percentage
of GDP
governments
spend on the
cultural sector.
Civil society Participation in Percentage Civil society Factor data: Eurostat
participation in cultural participation in participation in database
cultural activities by cultural activities cultural (2020)
governance types: cinema, by types: cinema, activities by
live live performances, types: cinema,
performances, or cultural sites is a live
or cultural sites. direct criterion that performances,
reflects society’s or cultural sites.
interest in cultural The percentage
activities and their expression
management. reflects how
much society
participates in
cultural
activities.
Participation in Final Price index Final consumption Factor data: final Eurostat
going-out consumption (implicit expenditure consumption database
cultural activities expenditure of deflator), of a household expenditure of (2020)
a household by 2015=100, by going-out households by
going-out euro recreational is a going-out
recreational. direct criterion recreational.
that reflects how Represented by
households price index how

Explanation
of relationship
economics factors
is measured
participate in many
going-out cultural households
activities. participate
in outbound
cultural
activities.
Participation in Students Number The chosen Factor data: Eurostat
identity-building participating in for 100000 criterion directly students database
cultural activities identity- population reflects the participating in (2020)
building participation in identity-building
cultural identity-building cultural
activities by activities activities by
education level, of a specific target education level,
programme group – students. programme
orientation, sex, orientation, s
and field of ex and field
education. of education.
It shows
the number of
students who
are participate in
identity-building
activities.
Professional Cultural Thousand Cultural Factor data: Eurostat
training in the employment by persons employment, cultural database
culture sector education. according to employment by (2020)
employees’ education.
education, reflects Shows how
the professionalism many thousands
of employees in the of people work
cultural sector. in the field of
culture in each
country by the
international
standard
classification of
education levels.
The second step was the usage of multi-criteria decision-making method TOPSIS, which was
selected in order to assess the present situation in the field of cultural economics and to create
an index of the EU11 countries. Statistics have been used for 2020 as this is the most recent
data available, which is the limitation of this study. The data was collected from Eurostat and
UNESCO World Heritage List databases (Table no. 2).
Vol. 26 • No. 65 • February 2024 321

The Case of EU11
3. Empirical Findings
In the first part of the research, the answers of the selected experts were collected, and the
weight of each factor was calculated based on the answers. The weights of cultural economics
factors are presented in table no. 3.
Table no. 3. Obtained weights of factors of cultural economics performance
Factors of cultural economics performance Weights of factors
Cultural heritage 0.1475

Cultural institutions 0.0931
Contribution of cultural activities to GDP 0.0555
Cultural employment 0.0803
Household expenditures on culture 0.0490
Distribution of cultural infrastructures 0.0647
Civil society participation in cultural governance 0.0915
Participation in going-out cultural activities 0.1172
Participation in identity-building cultural activities 0.1879
Professional training in the culture sector 0.1133
As can be seen from table no. 3, the maximum weights are assigned for participation in
identity-building cultural activities, cultural heritage, participation in going-out cultural
activities, and professional training in the culture sector. These four cultural economics
factors stand out significantly with the highest weighted result compared to other factors.
There may be several reasons for this, such as the influence of these factors in cultural policy,
in strengthening civic activity, and in the development of social capital.
After determining the weights, the multi-criteria assessment method TOPSIS was used in the
study in order to calculate the index of cultural economics of the EU11, and the results are
presented in table no. 4.
Table no. 4. Factors of cultural performance ranking of post-communist countries
Country Index Rank
Romania 0.8866 1
Bulgaria 0.69411 2
Slovakia 0.67583 3
Latvia 0.65928 4
Lithuania 0.57618 5
Estonia 0.56568 6
Czech Republic 0.55922 7
Hungary 0.55274 8
Poland 0.55237 9
Croatia 0.48718 10
Slovenia 0.46794 11
As can be seen from Table no. 4, the first five places were arranged as follows: Romania,
Bulgaria, Slovakia, Latvia, and Lithuania. At first glance, the outcomes pertaining to

Romania and Bulgaria might appear unexpected, yet scientific studies reveal a different
perspective. The studies developed by Parra and Ferraz (2021) show that the majority of
Romanian society perceives communism as a dark period in the country’s history and culture.
Dobre (2021) states in a research paper that Romania is currently applying new
communication strategies to promote its culture to avoid the communist heritage of the
country. A similar situation exists in Bulgaria (Ivanova, 2017; Dobre, 2021). According to
the authors, since joining the European Union, Bulgaria has been rapidly moving towards a
European one. The tourism image of the country is created through a peculiar culture, the
restoration, and preservation of its objects. The other three countries in the top five are
Slovakia, Latvia and Lithuania, which is a likely result because these countries are similar in
terms of both economic and social growth, and in the same year, 2004, they were one of the
first post-Soviet countries to join the European Union.
4. Discussion
This study singles out the factors of cultural economics and presents the weights of each
factor. The factor with the highest weight – participation in identity-building cultural
activities – is not an unexpected result. Participation in identity-building cultural activities is
included in the cultural policy-making strategies of the countries as one of the goals; this
already shows the importance of this factor. In order to form the critical thinking and
citizenship of a society that is creating its own identity and to create a sustainable socio-
economic value of culture for national progress, the criterion is a priority for the formation
of the countries’ cultural policy. The second factor with the highest weighted result is cultural
heritage. Heritage is the ethnically, historically, aesthetically, or scientifically critical cultural
values that have been inherited over several generations. It is essential in terms of community
or group identity and continuity and shows the diversity of cultures. More than one study
reveals the importance of cultural heritage in cultural economics. The importance of cultural
heritage is analysed through the development of tourism (Swensen and Nomeikaite, 2019;
Ciurea and Filip, 2019), the culture of historical memory formation (Bräuchler, 2019;
Esposito and Ricci, 2020), and the implementation of innovations in order to restore and
preserve heritage (Li, 2020). The importance of heritage is also revealed in various strategic
documents. The primary mission of UNESCO is “to encourage countries to sign the World
Heritage Convention and to ensure the protection of their natural and cultural heritage”. Also,
to encourage States Parties to the Convention to nominate sites within their national territory
for inclusion on the World Heritage List and help States Parties safeguard World Heritage
properties by providing technical assistance and professional training (UNESCO, 2014).
According to the Convention, cultural heritage and the primary cultural value of the countries
form the country’s identity and communality and contribute to the general well-being of the
population (UNESCO World Heritage Centre, 1972). The third factor is participation in
going-out cultural activities, which is also emphasised in the strategy of the countries’
cultural policy formation. Participation in such activities develops a person’s identity,
increases self-confidence, which also affects the improvement of the quality of life. Such
persons usually belong to various public organisations that contribute to the creation of civil
society (Caron et al., 2020).
In the strategies for the formation of cultural policy, this factor is singled out as increasing
the quality and efficiency of the activities of cultural and art institutions, promoting equal,
high-value, and diverse forms of cultural access to various groups of society, and developing
cultural competencies. This is important in the point of cultural economics view because the
Vol. 26 • No. 65 • February 2024 323

The Case of EU11
expansion and accessibility of cultural activities depend on various cultural policy-making

strategies, and all this promotes the development of cultural economics as such.
Professional training in the culture sector is the fourth cultural economics factor. It is
important that suitable conditions for creativity, professional development, and internships
abroad are created for those working in the cultural sector and those who create culture
(Gautam and Basnet, 2020). This would positively influence the promotion of the
internationality of culture and its quality, allow various areas of culture to constantly renew,
contribute to the creation of the country’s image, the growth of exports and strengthen
intercultural dialogue. In the strategies for the formation of cultural policy, professional
training in the cultural sector is singled out as one of the main factors that ensure the quality
and proper dissemination of culture. It is emphasised that the development of cultural
competencies promotes public citizenship, historical and national awareness, information
literacy, and cultural accessibility for various groups of society. After discussing the results
of the obtained weights, it is possible to move on to discussing the results of creating an index
of the EU11 countries.
The results show that the first five places belong to Romania, Bulgaria, Slovakia, Latvia, and
Lithuania. Such results are supported by scientific research conducted by other authors. For
example, Dmytrów and Bieszk-Stolorz (2021) analysed economic growth in post-communist
countries a few years after joining the European Union. According to the 2019 data of the
conducted study, Bulgaria, Slovakia, Latvia, and Lithuania had one of the highest indicators
of all post-communist countries. This shows how fast these countries developed. Another
important aspect that shows the independence and rapid development of these countries is
the Democracy Index. This index is prepared annually by “The Economist Group” in order
to provide information about fair elections held in countries, citizens’ trust in their political
leaders, and access to civil rights. According to 2022 data, four of these countries are among
the 61 countries in the world as having top ranking of democracy (Economist Intelligence,
2023). This is an important indicator because a strong democracy creates opportunities for
the expansion of all sectors; for the cultural sector, it is extremely significant due to its
specificity and uniqueness. Another indicator with which obtained results can be compared
is the Global Happiness Index. This index is prepared by the United Nations to determine
how happiness and well-being influence social and economic development. The 2022 World
Happiness Index data shows that four of these ranked countries are among the 41 happiest
countries in the world (Helliwell et al., 2021). It was chosen to compare the results of the
study with these indices because, in the analysed strategies for the formation of the cultural
policy of the countries, it is these indices that are specified as the criteria for evaluating the
strategic goal.
Thus, the analysed different scientific studies confirm the results of this research, that after
creating an index of the post-communist countries according to the selected cultural factors,
the first five places are occupied by Romania, Bulgaria, Slovakia, Latvia, and Lithuania.
Conclusions
In various aspects, culture occupies an important place in the modern world. Culture fosters
the increase of value and competitiveness of products; it promotes social inclusion and social
development; culture is very important for spatial planning and urban development, and
culture also serves for the development of collective awareness and thinking. It forms the

concept of the cultural economy and contributes to the perception. The analysis of the
scientific literature shows that factors in measuring the country’s cultural economics are
these: cultural heritage, cultural institutions, the contribution of cultural activities to GDP,
cultural employment, household expenditures on culture, professional training in the culture
sector, civil society participation in cultural governance, distribution of cultural
infrastructures, participation in identity-building cultural activities, participation in going-out
cultural activities.
The article collected and analysed two types of data: expert surveys and statistical data, in
order to assess which of the selected cultural economics factors are more important in relation
to each other. The obtained weights of factors of cultural economics show that the highest
weights are assigned to: participation in identity-building cultural activities, cultural heritage,
participation in going-out cultural activities, and professional training in the culture sector.
These obtained results are important for policy-makers who develop national cultural policy
strategies. Considering the obtained results, policy-makers can purposefully integrate these
factors in order to create sustainable social and economic value of culture for national
progress.
The results obtained from the post-communist countries that are members of the European
Union were organised according to these factors. The results showed that the first five places
are occupied by Romania, Bulgaria, Slovakia, Latvia, and Lithuania.
This is an expected result based on the analysis of different scientific studies, according to
which these countries developed faster both economically and socially than other EU11
countries after the collapse of the Soviet Union, and also compared to others, EU11 have the
highest democracy ratings.
References
Ayhan, M.B., 2013. A Fuzzy AHP Approach for Supplier Selection Problem: A Case Study
in a Gearmotor Company. International Journal of Managing Value and Supply Chains,
[e-journal] 4(3), pp.11-23. https://doi.org/10.5121/ijmvsc.2013.4302.
Azeem, M., Ahmed, M., Haider, S. and Sajjad, M., 2021. Expanding competitive advantage
through organizational culture, knowledge sharing and organizational innovation.
Technology in Society, [e-journal] 66, article no. 101635. https://doi.org/10.1016/
j.techsoc.2021.101635.
Baculáková, K. and Grešš, M., 2021. Spatial distribution model for targeting the support for
cultural institutions’ development: A case study of Slovakia. Muzeologia a Kulturne
Dedicstvo, [e-journal] 9(3), pp. 93-112. https://doi.org/10.46284/mkd.2021.9.3.5.
Baluku, M.M., Kikooma, J.F., Bantu, E., Onderi, P. and Otto, K., 2019. Impact of personal
cultural orientations and cultural intelligence on subjective success in self-employment
in multi-ethnic societies. Journal of Global Entrepreneurship Research, [e-journal] 9,
article no. 8(2019). https://doi.org/10.1186/s40497-018-0144-0.
Barandiaran-Irastorza, X., Peña-Fernández, S. and Unceta-Satrústegui, A., 2020. The
archipelago of cultural and creative industries: A case study of the basque country.
Economies, [e-journal] 8(1), article no. 21. https://doi.org/10.3390/
ECONOMIES8010021.
Vol. 26 • No. 65 • February 2024 325

The Case of EU11
Beşikçi, E.B., Kececi, T., Arslan, O. and Turan, O., 2016. An application of fuzzy-AHP to
ship operational energy efficiency measures. Ocean Engineering, [e-journal] 121,
pp. 392-402. https://doi.org/10.1016/j.oceaneng.2016.05.031.
Bilbao-Terol, A., Arenas-Parra, M., Cañal-Fernández, V. and Antomil-Ibias, J., 2014. Using
TOPSIS for assessing the sustainability of government bond funds. Omega, [e-journal]
49, pp. 1-17. https://doi.org/10.1016/j.omega.2014.04.005.
Bosone, M., De Toro, P., Girard, L. F., Gravagnuolo, A. and Iodice, S., 2021. Indicators for
ex‐ post evaluation of cultural heritage adaptivreuse impacts in the perspective of the
circular economy. Sustainability, [e-journal] 13(9), article no. 4759.
https://doi.org/10.3390/ su13094759.
Boyd, R.L., 2021. The Harlem Renaissance and Blacks’ Employment in Cultural Expression
Occupations. Journal of African American Studies, [e-journal] 25(1), pp. 82-101.
https://doi.org/10.1007/s12111-021-09514-w.
Bräuchler, B., 2019. Brokerage, Creativity and Space: Protest Culture in Indonesia. Journal
of Intercultural Studies, [e-journal] 40(4), pp. 451-468. https://doi.org/10.1080/
07256868.2019.1628721
Buljubašić, I., Ham, M. and Pap, A., 2016. Factors Influencing the Implementation of
Unconventional Marketing in Cultural Institutions – Evidence from Croatia. Scientific
Annals of Economics and Business, [e-journal] 63(2), pp. 247-272. https://doi.org/
10.1515/saeb-2016-0120.
Capello, R., 2018. Cohesion Policies and the Creation of a European Identity: The Role of
Territorial Identity. Journal of Common Market Studies, [e-journal] 56(3), pp. 489-503.
https://doi.org/10.1111/jcms.12611.
Caron, R., Lee, E.O.J. and Pullen Sansfaçon, A., 2020. Transformative Disruptions and
Collective Knowledge Building: Social Work Professors Building Anti-oppressive
Ethical Frameworks for Research, Teaching, Practice and Activism. Ethics and Social
Welfare, [e-journal] 14(3), pp. 298-314. https://doi.org/10.1080/17496535.2020. 1749690.
Cellini, R. and Cuccia, T., 2021. Female workforce participation and household expenditure
for culture and recreation: macroeconomic evidence from the Italian regions. Applied
Economics, [e-journal] 53(14), pp. 1659-1671. https://doi.org/10.1080/00036846.2020.
1841087.
Chabay, I., 2020. Vision, identity, and collective behavior change on pathways to sustainable
futures. Evolutionary and Institutional Economics Review, [e-journal] 17(1), pp. 151-165.
https://doi.org/10.1007/s40844-019-00151-3.
Chang, D.-Y., 1996. Applications of the extent analysis method on fuzzy AHP. European
Journal of Operational Research, [e-journal] 95(3), pp. 649-655. https://doi.org/10.1016/
0377-2217(95)00300-2.
Chhabra, D., 2021. Cultural and heritage tourism: an introduction. Journal of Heritage
Tourism, [e-journal] 16(5), pp. 612-614. https://doi.org/10.1080/1743873x.2021.
1942622.
Ciurea, C. and Filip, F.G., 2019. The globalization impact on creative industries and cultural
heritage: a case study. Creativity Studies, [e-journal] 12(2), pp. 211-223. https://doi.org/
10.3846/cs.2019.7753.

Cobo, A., Vanti, A.A. and Rocha, R., 2014. A Fuzzy Multi-criteria Approach for it
Governance Evaluation. Journal of Information Systems and Technology Management,
[e-journal] 112(2), pp.257-276. https://doi.org/10.4301/S1807-17752014000200003.
Dinulescu, R. and Dobrin, C., 2022. Applying the fuzzy analytical hierarchy process for
classifying and prioritizing healthcare quality attributes. Management & Marketing.
Challenges for the Knowledge Society, [e-journal] 17(1), pp.15-40.
https://doi.org/10.2478/mmcks-2022-0002.
Dobre, C.F., 2021. The patrimonialization of the communist past in Romania: Laws,
memorials, and monuments. Balcanistic Forum, [e-journal] 30(1), pp.179-195.
https://doi.org/10.37708/bf.swu.v30i1.11.
Dmytrów, K. and Bieszk-Stolorz, B., 2021. Comparison of changes in the labour markets of
post-communist countries with other eu member states. Equilibrium. Quarterly Journal
of Economics and Economic Policy, [e-journal] 16(4), pp.741-764.
https://doi.org/10.24136/eq.2021.027.
Economist Intelligence, 2023. Democracy Index 2022 Frontline democracy and the battle
for Ukraine. [online] Available at: <https://www.eiu.com/n/wp-content/uploads/
2023/02/Democracy-Index-2022_FV2.pdf?li_fat_id=f1fbad7e-a282-4b9e-9f8f-
6a6d5a9fe6b8> [Accessed 9 December 2023].
Elwell, T.L., López-Carr, D., Gelcich, S. and Gaines, S.D., 2020. The importance of cultural
ecosystem services in natural resource-dependent communities: Implications for
management. Ecosystem Services, [e-journal] 44, article no. 101123.
https://doi.org/10.1016/j.ecoser.2020.101123.
Esposito, P. and Ricci, P., 2020. Cultural organizations, digital Corporate Social
Responsibility and stakeholder engagement in virtual museums: a multiple case study.
How digitization is influencing the attitude toward CSR. Corporate Social Responsibility
and Environmental Management, [e-journal] 28(2), pp.953-964. https://doi.org/
10.1002/csr.2074.
Fihel, A. and Okólski, M., 2019. Population decline in the post-communist countries of the
European Union. Population & Societies, 567(6), pp.1-4.
Gautam, D.K. and Basnet, D., 2020. Organizational culture for training transfer: the
mediating role of motivation. International Journal of Organizational Analysis, [e-
journal] 29(3), pp.769-787. https://doi.org/10.1108/IJOA-04-2020-2147.
Goffi, G., Cucculelli, M. and Masiero, L., 2019. Fostering tourism destination
competitiveness in developing countries: The role of sustainability. Journal of Cleaner
Production, [e-journal] 209, pp.101-115. https://doi.org/10.1016/j.jclepro.2018.10.208.
Grenersen, G., 2012. What is a document institution? A case study from the South Sámi
community. Journal of Documentation, [e-journal] 68(1), pp.127-133. https://doi.org/
10.1108/00220411211200356.
Grenni, S., Horlings, L.G. and Soini, K., 2020. Linking spatial planning and place branding
strategies through cultural narratives in places. European Planning Studies, [e-journal]
28(7), pp.1355-1374. https://doi.org/10.1080/09654313.2019.1701292.
Helliwell, J. F., Layard, R., Sachs, J.D. and De Neve, J.-E., 2021. World Happiness Report
2021. [online] Available at: <https://www.wellbeingintlstudiesrepository.org/
hw_happiness/5/> [Accessed 8 December 2023].
Vol. 26 • No. 65 • February 2024 327

The Case of EU11
Hofstede, G., Noorderhaven, N., Thurik, R., Uhlaner, L. and Wennekers, A.R.M., 2004.
Culture's Role in Entrepreneurship: Self-Employment Out of Dissatisfaction. [online]
University of Illinois at Urbana-Champaign's Academy for Entrepreneurial Leadership
Historical Research Reference in Entrepreneurship. Available at:
<https://ssrn.com/abstract=1496783> [Accessed 8 December 2023].
Ivanova, M., 2017. The inclusion of the communist/socialist heritage in the emerging
representations of eastern Europe: The case of Bulgaria. Tourism, Culture and
Communication, [e-journal] 17(1), pp.31-46. https://doi.org/10.3727/109830417X
14837314056852.
Janjic, A., Savic, S., Velimirovic, L. and Nikolic, V., 2015. Renewable energy integration in
smart grids-multicriteria assessment using the fuzzy analytical hierarchy process. Turkish
Journal of Electrical Engineering & Computer Sciences, [e-journal] 23, pp.1896-1912.
https://doi.org/10.3906/elk-1404-287.
Li, J., Krishnamurthy, S., Pereira Roders, A. and van Wesemael, P., 2020. Community
participation in cultural heritage management: A systematic literature review comparing
Chinese and international practices. Cities, [e-journal] 96(September 2019).
https://doi.org/10.1016/j.cities.2019.102476.
Li, X., 2020. Cultural creative economy and urban competitiveness: How one matters to the
other. Journal of Urban Affairs, [e-journal] 42(8), pp.1164-1179. https://doi.org/10.1080/
07352166.2020.1727293.
Meir, D. and Fletcher, T., 2019. The transformative potential of using participatory
community sport initiatives to promote social cohesion in divided community contexts.
International Review for the Sociology of Sport, [e-journal] 54(2), pp.218-238.
https://doi.org/10.1177/1012690217715297.
Orlando, B., Ballestra, L.V., Scuotto, V., Pironti, M. and Giudice, M.D., 2022. The Impact
of R&D Investments on Eco-Innovation: A Cross-Cultural Perspective of Green
Technology Management. IEEE Transactions on Engineering Management, [e-journal]
69(5), pp.2275-2284. https://doi.org/10.1109/TEM.2020.3005525.
Parra, R. and Ferraz, J., 2021. From a Communist Heritage to an Unwanted Past: The Case
of Romania. Science Insights, [e-journal] 38(1), pp.298-304. https://doi.org/10.2139/
ssrn.3899261.
Papava, V., 2018. Catching Up and Catch-Up Effect: Economic Growth in Post-Communist
Europe (Lessons from the European Union and the Eastern Partnership States). European
Journal of Economic Studies, [e-journal] 7(2), pp.109-125. https://doi.org/10.13187/
es.2018.7.109.
Purkarthofer, E., Humer, A. and Mattila, H., 2021. Subnational and Dynamic
Conceptualisations of Planning Culture: The Culture of Regional Planning and Regional
Planning Cultures in Finland. Planning Theory and Practice, [e-journal] 22(2), pp.244-
265. https://doi.org/10.1080/14649357.2021.1896772.
Scott, D., 2019. Music Hall: Regulations and behaviour in a British cultural institution.
Muzikologija, [e-journal] 26, pp.61-74. https://doi.org/10.2298/muz1926061s.
Swensen, G. and Nomeikaite, L., 2019. Museums as narrators: heritage trails in a digital era.
Journal of Heritage Tourism, [e-journal] 14(5-6), pp.525-543. https://doi.org/10.1080/
1743873X.2019.1574803.

UNCTAD, 2018. Creative Economy Outlook. [online] Available at: <https://unctad.org/
en/PublicationsLibrary/ditcted2018d3_en.pdf> [Accessed 9 December 2023].
UNESCO, 2020. UNESCO Culture for Development Indicators. [online] Available at:
<https://www.unesco.org/creativity/en/activities/cdis> [Accessed 9 December 2023].
UNESCO, 2014. UNESCO culture for development indicators: methodology manual.
[online] Available at: <https://unesdoc.unesco.org/ark:/48223/pf0000229608>
[Accessed 9 December 2023].
Valverde-Moreno, M., Torres-Jimenez, M. and Lucia-Casademunt, A.M., 2021. Participative
decision-making amongst employees in a cross-cultural employment setting: evidence
from 31 European countries. European Journal of Training and Development, [e-journal]
45(1), pp.14-35. https://doi.org/10.1108/EJTD-10-2019-0184.
Warburton, J. and McLaughlin, D., 2007. Passing on our culture: How older Australians from
diverse cultural backgrounds contribute to civil society. Journal of Cross-Cultural
Gerontology, [e-journal] 22(1), pp.47-60. https://doi.org/10.1007/s10823-006-9012-4.
World Economic Forum, 2020. Argentina’s Travel and Tourism Competitiveness The Case
for Promoting and Preserving Cultural, Natural and Heritage Assets in Tourism. March.
[online] Available at: <www.weforum.org> [Accessed 2 August 2023].
Zhang, L., Chen, Y., Lyulyov, O. and Pimonenko, T., 2022. Forecasting the Effect of
Migrants’ Remittances on Household Expenditure: COVID-19 Impact. Sustainability,
[e-journal] 14(7), article no. 4361. https://doi.org/10.3390/su14074361.
Zhenrao, C., Chaoyang, F., Qian, Z. and Fulong, C., 2021. Joint development of cultural
heritage protection and tourism: the case of Mount Lushan cultural landscape heritage
site. Heritage Science, [e-journal] 9, article no. 86. https://doi.org/10.1186/s40494-021-
00558-5.
Vol. 26 • No. 65 • February 2024 329

AE Research on Tax Compliance Incentive Effects of Platform Companies from
the Perspective of Incomplete Contract – An Empirical Study Based on China
RESEARCH ON TAX COMPLIANCE INCENTIVE EFFECTS OF PLATFORM

COMPANIES FROM THE PERSPECTIVE OF INCOMPLETE CONTRACT –
AN EMPIRICAL STUDY BASED ON CHINA
Xuefeng Shao1 and Shi Chen2
1)2)
Economics School of Jilin University, Changchun, China

Shao, X. and Chen, S., 2024. Research on Tax Received: 22 September 2023
Compliance Incentive Effects of Platform Companies Revised: 20 November 2023
from the Perspective of Incomplete Contract – An Accepted: 18 December 2023
Empirical Study Based on China. Amfiteatru
Economic, 26(65), pp. 330-344.
Abstract
In this paper, based on the incomplete contract perspective, we select the implementation of
the Electronic Commerce Law of the People’s Republic of China as a quasi-natural
experiment to study the tax compliance incentive effects of platform firms. Our study finds
that the Chinese experience helps to improve the efficiency of tax compliance contract
enforcement and significantly increases the propensity of platform firms to comply with
taxes. Of course, these effects are also constrained by the contractual environment, social
responsibility, financing constraints, and market competition. Further mechanism tests show
that the incompleteness of the tax compliance contract is compensated by two mechanisms
of action, namely the reduction of information asymmetry and the reduction of transaction
costs of the tax department, which generate tax compliance incentive effects. The research
has important implications for optimising the tax compliance contract of platform firms and
reducing tax leakage in the platform economy.
Keywords: incomplete contract; platform enterprises; tax compliance; transaction cost
JEL Classification: H26, D86

Corresponding author, Shi Chen – e-mail: [email protected]

Introduction
Together with the wide application of digital information technology, the platform economy
as a new economic model of the market economy brings many advantages, such as reducing
the transaction costs of enterprises, increasing the social employment, breaking the
geographical and spatial restrictions, and reducing the waste of resources, and has become a
new engine of economic growth in various countries (Barbu et al., 2018; Niu et al., 2022;
Zha et al., 2022; Zhang et al., 2022a; Zhang et al., 2022b; Martin-Martin et al., 2022). From
a global perspective, China is in a relatively leading position in the development of the
platform economy. In terms of the number of platform companies, market capitalisation, and
transaction size, China ranks second in the world after the United States. From 2018 to 2021,
the transaction scale of China’s sharing economy increased from 2.9 trillion yuan to 3.7
trillion yuan, with a growth rate of 27.6%. Among them, the proportion of online take-out
revenue as well as online take-out per capita expenditure in China's food and beverage
industry reaches about 20% in 2021; the proportion of online cab passenger traffic among
online cabs reaches about 30%, and the proportion of online taxi per capita expenditure to
travel consumption expenditure reaches about 20%; the proportion of shared accommodation
to all accommodation room revenue reaches 6%, and the proportion of shared
accommodation per capita expenditure to accommodation consumption expenditure reaches
6% (National Information Center Sharing Economy Research Center, 2019, 2020). From
2017 to June 2021, the size of China’s online payment consumers grew from 530 million to
1.011 billion, an increase of up to 200% (China Internet Network Information Center,2021).
In 2021, China’s online retail sales reached 13.1 trillion yuan, up 14.1% year-on-year, a
growth rate 3.2 percentage points faster than the previous year (China Bureau of Statistics,
2022). The platform economy plays an increasingly important role in promoting economic
development by changing the traditional transaction model, reducing communication costs
between enterprises and consumers, accurately understanding consumer needs, and creating
personalised services. However, due to its characteristics of two-sided market, economies of
scale and network effects, the platform economy also raises a number of prominent problems,
such as data leakage, price abuse, tax evasion, disorderly capital expansion and winner-take-
all (Rodrigues et al., 2022).
In view of this, it is necessary to regulate the platform economy, guide the healthy
development of the platform economy and compensate for the imperfect nature of platform
economy contracts. In 2019, China issued and implemented the first comprehensive law in
the field of e-commerce, the E-Commerce Law of the People’s Republic of China (hereinafter
referred to as the E-Commerce Law). The E-Commerce Law clearly stipulates the tax
obligations of e-commerce operators and the need for e-commerce platforms to fulfil the
obligation to report the identity and tax information of operators on the platform to the tax
authorities. The implementation of this policy fills the black zone in the field of taxation of
the platform economy. Based on the above background, this paper conducts a study using
China as an example to explore the tax compliance incentive effect of platform enterprises.
Whether platform companies induce large-scale tax evasion as one of the problems caused
by the platform economy. At present, scholars have different views on the taxation of
platform companies. One view is that the tax collection of platform enterprises faces many
Vol. 26 • No. 65 • February 2024 331

difficulties. On the one hand, the incompatibility of the tax system with the platform economy
increases the transaction costs of taxpayers and tax authorities. Platform enterprises are
involved in several activities, and the industries are integrated. For example, the distinction
between business income, labour compensation, and royalties, and property transfer income.
The current tax system does not have appropriate rules on what type of income is appropriate
for the type of income of converging enterprises (Shao et al., 2022). There is also no
uniformity as to whether the various forms of electronic services provided by platform
companies are subject to pre-tax deduction and at what rate. All of these are additional
burdens for platform companies. Platform enterprises break the boundaries of geographical
space and facilitate transactions across regions and even countries (Agrawal and Fox, 2017;
Spinosa and Chand, 2018). At the same time, the transactions of platform enterprises involve
multiple subjects. It is difficult for the tax authority to verify the information of each business,
all revenue acquisition channels, revenue payment subjects, and costs and expenses
consumed by the enterprise, which increases the cost of enforcement and supervision by the
tax authority(Alm, 2021; Ma Argiles-Bosch et al., 2021). However, concealing the
transaction process of platform enterprises widens the information gap between taxpayers
and tax authorities. Platform enterprises have all the transaction information, but the lack of
effective incentives for platform enterprises to report information results in platform
enterprises not truthfully reporting all the tax-related information they have, such as the
identity of transaction subjects, bank account numbers, payment subjects, and income
amounts, to the tax authority(Agrawal and Fox, 2021). The data underlying the tax system is
decentralised and cannot be adapted to the platform economy. Tax governance has limited
application of technologies, such as cloud computing and blockchain, to intelligently and
accurately monitor anomalous changes in tax sources (Stabrowski, 2017; Agrawal and Fox,
2021).
Another view is that tax collection by platform companies has new advantages. On the one
hand, platform enterprises have a large amount of tax-related information data, which
expands the channels for tax authorities to obtain information, lays the foundation for tax
authorities to obtain massive data, and accelerates the process of building intelligent taxation.
With the first-hand data held by platform companies, tax authorities can easily confirm
whether the transactions are real and trustworthy, facilitate tax management, improve the
level of tax risk control, and increase the accuracy of tax audits (Agrawal, 2021). On the other
hand, platform enterprises collect a large number of tax sources, reduce the marginal cost of
tax administration, achieve the marginal cost of administration close to zero, and create the
scale effect of tax administration. Tax authorities and platform enterprises form a good
cooperation, and platform enterprises can play the function of supervision and management.
The taxation department can achieve the goal of supervising the relevant subjects of the
platform economy by supervising only the platform, and significantly reduce the time and
manpower costs of tax collection and administration. As the proportion of the platform
economy in the overall economy increases, the number of tax returns from platform
enterprises is also gradually increasing. Due to the tax agglomeration ability of platform
enterprises, the marginal cost of tax administration tends to zero, creating the scale effect of
tax administration (Alm, 2021).

2.1. Theoretical analysis and research hypothesis
2.1.1. Information mechanism
The nature of the platform economy leads to the opacity of data and higher levels of
manipulation (Ma Argiles-Bosch et al., 2021). The tax authority cannot take the control and
collect and verify the authenticity of tax information of enterprises with the help of
accounting books, etc. The platform economy includes many participating subjects, and the
subjects are distributed in different regions, which leads to the fragmented characteristics of
tax information (Thomas, 2018). Platform companies achieve the objective of tax non-
compliance through virtual non-existent transactions. Information on contract flow, capital
flow, and logistics is scattered, and the tax authority cannot control tax by invoice and verify
the authenticity of transactions. Therefore, the information asymmetry between platform
companies and tax authorities leads to the incompleteness of tax contracts. The E-commerce
Law stipulates that platform operators report the identity information and tax-related
information of taxpayers on the platform to the taxation department, which can play a
complementary and synergistic role among the data and obtain a panoramic view of things.
At the same time, from the taxpayer's point of view, the information reported by the platform
increases the channels for the taxation department to obtain information, and the data from
different sources can be corroborated with each other, which reduces the behaviour of
platform companies to provide “different” data to different departments for opportunistic
motives, and ensures the authenticity of the data (Agrawal and Fox, 2021; Alm, 2021). As a
result, tax authorities are able to detect corporate anomalies in a timely and accurate manner.
The company’s internal information is more transparent, the difficulty and cost of financial
manipulation is significantly increased, and the completeness of the tax compliance contract
is improved.
2.1.2. Transaction cost mechanism

The costs of enforcing and monitoring tax compliance in the platform economy are high (Ma
Argiles-Bosch et al., 2021). The paperless transactions of the platform economy increase the
possibility of transaction manipulation and concealment and make it more difficult for the
tax authority to conduct audits. There is an information asymmetry between the tax authority
and platform participants in several respects, such as whether they are involved in platform
activities, the type of business they are involved in, the channels of income generation, the
recipients of income, and the costs and expenses incurred (Stabrowski, 2017). At the same
time, there are many gaps in tax administration technologies. Different systems in the tax
registration and filing process have different inputs and exports. Different standards for data
transfer between departments make it difficult to integrate effectively and respond effectively
to the characteristics of the platform economy (Stabrowski, 2017; Agrawal and Fox, 2021).
All of the above increases the transaction costs of supervision by tax authorities. The E-
commerce Law mentions that e-commerce platforms must fulfil the tax department’s
obligation to report the identity information and tax-related information of platform
operators. By reporting the primary data they have to the taxation department, platform
operators will expand the channels through which the taxation department can obtain tax-
related information, increase the verifiability of tax-related information from different
channels, accurately distinguish whether the tax-related information is true or not, enhance
Vol. 26 • No. 65 • February 2024 333

the risk analysis ability of the taxation department, and improve the accuracy of audits (Alm,
2021; Agrawal, 2021). Good cooperation between platform companies and tax authorities.
With the ability of platform companies to collect tax sources, the tax authority realises the
marginal cost of tax administration, which is close to zero, by supervising all tax subjects on
the platform through platform companies. Transaction costs, such as time and manpower
required for tax collection and management, are greatly reduced, the incompleteness of tax
compliance contracts is reduced, and the efficiency of tax compliance contracts is increased.
Based on the above analysis, this paper proposes the following research hypotheses：
H1: The implementation of the E-Commerce Law effectively compensates the
incompleteness of enterprises' tax compliance contract and motivates enterprises' tax
compliance.
H2: The E-commerce Law improves the tax compliance contract by reducing information
asymmetry and stimulating companies' tax compliance motivation.
H3: The E-commerce Law can effectively reduce transaction costs and improve the
efficiency of tax compliance contracts, thus curbing the tendency of corporate tax non-
compliance.
2.2. Research design

2.2.1. Data sources
The data of platform enterprises in this paper are obtained from Wind online sales data. We
remove Hong Kong, overseas listed, delisted, IPO, ST enterprises, and non-manufacturing
enterprises, and finally obtain a sample of 187 platform enterprises. In this paper, we match
the platform enterprises with manufacturing listed companies in Shanghai and Shenzhen A-
shares, and process the data as follows: (1) delete ST and *ST enterprises; (2) delete
enterprises with serious missing data and outlier enterprises; (3) winsorise at the top and
bottom 1% levels. We finally get 6841 observations. The variable indicators in this paper are
mainly from Wind database, China Statistical Yearbook, China Marketisation Index Report
by Provinces and Hexun.com, etc. The data interval is 2015-2021.
2.2.2. Model
To test whether there is an incentive effect of policy implementation on tax compliance, i.e.,
hypothesis H1. Drawing on Bertrand and Mullainathan (2003) and Chen et al. (2012), this
paper constructs the following difference-in-difference model.
𝑇𝑎𝑥𝑖𝑡 = 𝛼0 + 𝛼1 𝑇𝑟𝑒𝑎𝑡𝑖 × 𝑇𝑖𝑚𝑒𝑡 + 𝛼2 𝑋𝑖𝑡 + 𝑌𝑒𝑎𝑟𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (1)
where i represents the firm and t represents time. 𝑇𝑟𝑒𝑎𝑡𝑖 indicates whether the firm is in the
treatment group. 𝑇𝑖𝑚𝑒𝑡 represents whether firm i is affected by the implementation of the
policy in year t. X is the control variables. 𝑌𝑒𝑎𝑟𝑡 and 𝜇𝑖 are year and individual fixed effects,
respectively, and𝜀𝑖𝑡 is a random disturbance term.

2.2.3. Variable setting
The explained variable is tax compliance (TAX). This paper draws on Desai and Dharmapala
(2006) to measure the tax compliance of firms by their Book-Tax Difference (BTD).
The explanatory variable is the treatment group dummy variable (Treat). The data for
platform enterprises are taken from Wind online sales data. Non-platform enterprises are
manufacturing enterprises listed on Shanghai and Shenzhen A-shares. In this paper, platform
enterprises are taken as the treatment group with a value of 1. Non-platform enterprises, as
the control group in this paper, are taken as 0. Policy implementation dummy variable (Time).
China started to implement the Electricity Business Law in January 2019. Therefore, 2019 is
selected as the time point for policy implementation in this paper. That is, 1 is taken in 2019
and after, and the value is 0 before 2019.
Control variables. Considering that other factors may affect corporate tax compliance,
referring to the studies of (Armstrong et al., 2015; Dyreng et al., 2010), the following control
variables were selected: firm size (Size), gearing ratio (Lev), return on assets (Roa),
shareholding ratio of the first largest shareholder (Top1), separation of powers ratio (Sep),
director size (Brdsize), combined leverage (DTL), total assets Growth rate (Zgrowth), Net
asset per share growth rate (NAG). (Table no. 1).
Table no. 1. Variable definitions
Variable Variable
Variable name Definitions and metrics
types code
Explained Tax compliance TAX BTD = (accounting profit before tax -
variables taxable income) / total assets
at the end of the period
Explanatory Processing group Treat Processing group. Platform enterprises
variables take the value of 1
Control group. Non-platform
enterprises take the value of 0
Policy implementation Time 1 for 2019 and beyond, 0 for previous
dummy variable values
Control Firm size Size the logarithm of total corporate assets
variables Gearing ratio Lev the ratio of total liabilities to total
assets
Return on assets Roa Net profit to average assets
Shareholding ratio of the Top1 the number of shares held by the first
first largest shareholder largest shareholder/total equity
Separation of powers Sep the difference between control and
ratio ownership of a listed company
by the effective controller
Director size Brdsize the number of directors
Combined leverage DTL financial leverage × operating
leverage
Total assets Growth rate Zgrowth (Total assets at the end of the current
period - Total assets at the beginning
of the current period)/ Total assets at
the beginning of the current period
Vol. 26 • No. 65 • February 2024 335

Variable Variable
Variable name Definitions and metrics
types code
Net asset per share NAG (Net asset per share at the end of the
growth rate current period - Net asset per share at
the end of the same period of the
previous year)/Net asset per share at
the end of the same period of the
previous year
Where, net assets per share = total
owner's equity / number of common
shares
Year Year Year dummy variable
Individual 𝜇 Individual dummy variable

3.1. Empirical results and analysis
3.1.1. Basic regression results
The regression results for the tax compliance incentive effect of platform firms are shown in
Table no. 2. Column (1) of table no. 2 shows the regression results without the inclusion of
control variables, and the cross product term 𝑇𝑟𝑒𝑎𝑡 × 𝑇𝑖𝑚𝑒 is significantly negative at the
1% level, which initially confirming hypothesis H1. Column (2) shows the regression results
controlling for firm-level control variables, year fixed effects, and firm fixed effects. The
results show that the coefficient 𝛼1 of 𝑇𝑟𝑒𝑎𝑡 × 𝑇𝑖𝑚𝑒 is still significantly negative at the 1%
level, which further supports hypothesis H1, indicating that the implementation of the policy
will significantly motivates corporate tax compliance and creates an incentive effect of tax
compliance.
Table no. 2. Baseline regression results
(1) (2)
Variables TAX TAX
-0.0066*** -0.0061***
(-3.4620) (-3.1910)
Control variables No Yes
Year Yes Yes
𝜇 Yes Yes
N 6841 6841
R2 0.637 0.682
Note: *, **, and *** denote 10%, 5%, and 1% significance levels, respectively; t-statistics
are given in parentheses.

3.1.2. Robustness tests
To reduce sample selectivity bias and ensure robustness of results, this paper has separately
parallel trend test, propensity score matching (PSM), DDBTD as the explained variables,
explanatory variables are lagged for one period, adding other control variables and placebo
test (Rosenbaum and Rubin, 1983; Heckman et al., 1998; Desai and Dharmapala, 2009; Boler
et al., 2015; Li et al., 2016).The results are shown in Figure no.1,Table no. 3 and Figure no.
2, all of which demonstrate the robustness of the regression results.
Figure no. 1. Parallel trend graph
Table no. 3. Robustness tests

(1) (2) (3) (4)
PSM Replacing the The Adding
explained explanatory other
variables variables are control
Variables lagged by one variables
period
TAX TAX TAX TAX
-0.0095** -0.0065*** -0.0080*** -0.0060***
(-2.0606) (-3.3963) (-2.7699) (-3.1685)
EM -0.0001
(-0.6646)
Roi -0.0000***
(-2.5853)
GR -0.0006
(-0.6763)
Control variables Yes Yes Yes Yes
Year Yes Yes Yes Yes
𝜇 Yes Yes Yes Yes
N 1235 6837 4213 6841
R2 0.829 0.689 0.736 0.682
Note: *, **, and *** denote 10%, 5%, and 1% significance levels, respectively; t-statistics are given in parentheses.
Vol. 26 • No. 65 • February 2024 337

Figure no. 2. Placebo test graph
3.2. Mechanism of action test

3.2.1. Information mechanism
Based on the previous theoretical analysis, in order to test hypothesis H2. This paper refers
to Baron and Kenny (1986) and constructs the following mediating effect model.
𝐼𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛𝑖𝑡 = 𝛽0 + 𝛽1 𝑇𝑟𝑒𝑎𝑡𝑖 × 𝑇𝑖𝑚𝑒𝑡 + 𝛽2 𝑋𝑖𝑡 + 𝑌𝑒𝑎𝑟𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (2)
𝑇𝑎𝑥𝑖𝑡 = 𝛿0 + 𝛿1 𝑇𝑟𝑒𝑎𝑡𝑖 × 𝑇𝑖𝑚𝑒𝑡 + 𝛿2 𝐼𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛𝑖𝑡 + 𝛿3 𝑋𝑖𝑡 + 𝑌𝑒𝑎𝑟𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (3)
Where Information is the mediating variable, which indicates the transparency of corporate
information. Other variables are defined as in equation (1).This paper draws on Jin and Myers
(2006) to measure corporate information transparency using share price synchronisation. The
higher the synchronisation of the stock price synchronisation, the less transparent the
corporate information is. We refer to the method of Durnev et al. (2003) to measure stock
price synchronisation.
Since the coefficients of 𝛽1 as well as 𝛿2 in the mediating effect model are significant in
columns (1) and (2) of table no. 4, the mediating effect exists, and the magnitude of this effect
is 0.00013. The coefficient of the variable 𝑇𝑟𝑒𝑎𝑡 × 𝑇𝑖𝑚𝑒 in column (1) of Table no. 4 is
significantly negative at the 5% level, indicating that the policy significantly improves the
information transparency of enterprises and exerts an information effect. Furthermore, the
coefficient of 𝑇𝑟𝑒𝑎𝑡 × 𝑇𝑖𝑚𝑒 in column (2) is significantly negative and that of Information
is significantly positive, indicating that the policy somehow promotes corporate information
transparency and further motivates the tax compliance of platform companies. The
contribution of this mediating effect to the total tax compliance effect is 2.2%. Hypothesis
H2 of this paper was tested.

Table no. 4. Information mechanism regression results
Variables (1) (2)
Information TAX
-0.1626** -0.0059***
(-2.3355) (-3.1213)
Information 0.0008**
(2.0232)
Control variables Yes Yes
Year Yes Yes
𝜇 Yes Yes
N 6841 6841
R2 0.553 0.682
Note: *, **, and *** denote 10%, 5%, and 1% significance levels, respectively; t-statistics are given in
parentheses.
3.2.2. Transaction cost mechanism

Based on the mechanistic analysis above, in order to test hypothesis H3. This paper draws on
Baron and Kenny(1986) to construct the following mediating effect model.
𝐶𝑜𝑠𝑡𝑖𝑡 = 𝛽0 + 𝛽1 𝑇𝑟𝑒𝑎𝑡𝑖 × 𝑇𝑖𝑚𝑒𝑡 + 𝛽2 𝑋𝑖𝑡 + 𝑌𝑒𝑎𝑟𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (4)
𝑇𝑎𝑥𝑖𝑡 = 𝛿0 + 𝛿1 𝑇𝑟𝑒𝑎𝑡𝑖 × 𝑇𝑖𝑚𝑒𝑡 + 𝛿2 𝐶𝑜𝑠𝑡𝑖𝑡 + 𝛿3 𝑋𝑖𝑡 + 𝑌𝑒𝑎𝑟𝑡 + 𝜇𝑖 + 𝜀𝑖𝑡 (5)
where Cost is the mediating variable and denotes the transaction costs of the tax authority.
The other variables are defined as in equation (1). In this paper, we use the tax administration
effort to represent the transaction costs of the tax authority. The higher the tax collection
effort, the higher the transaction cost that the tax authority needs to pay. Drawing on Xu et
al. (2011), Wen et al. (2020), Liang et al. (2021) and other measures of tax administration
intensity in China.
The coefficients corresponding to 𝛽1 as well as 𝛿2 in the mediating effect model are
significant in columns (1) and (2) of Table no. 5. Therefore, there is a mediating effect, and
the magnitude of this effect is 0.00015. The coefficient of the variable 𝑇𝑟𝑒𝑎𝑡 × 𝑇𝑖𝑚𝑒 in
column (1) of table no. 5 is significantly negative at the 10% level, indicating that the Chinese
experience effectively reduces the transaction costs of the tax department. Furthermore, the
coefficient of 𝑇𝑟𝑒𝑎𝑡 × 𝑇𝑖𝑚𝑒 in column (2) is significantly negative and Cost is significantly
positive, indicating that the policy somehow reduces transaction costs and further motivates
the tax compliance of platform enterprises, and the contribution of this part of the mediating
effect to the total effect of enterprise tax compliance is 2.8%. Hypothesis H3 of this paper
was tested.
Table no. 5. Regression results of transaction cost mechanism
Variables (1) (2)
Cost TAX
-0.0119* -0.0055***
(-1.7331) (-2.8758)
Cost 0.0129***
(2.6661)
Vol. 26 • No. 65 • February 2024 339

Variables (1) (2)

Cost TAX
Control variables Yes Yes
Year Yes Yes
𝜇 Yes Yes
N 6712 6712
R2 0.900 0.665
Note: *, **, and *** denote 10%, 5%, and 1% significance levels, respectively;t-statistics are given in parentheses.
3.3. Further analysis

3.3.1. Contractual environment
This paper measures the contractual environment using the indicators of the development of
market intermediary organisations and the legal institutional environment among the five
indicators of the Fan Gang Marketisation Index. We divide the sample into two groups, high
and low, according to the median contractual environment for the regressions, and obtain the
results in table no. 6. Based on columns (1) and (2) of Table no.6, we can see that the
coefficient of 𝑇𝑟𝑒𝑎𝑡 × 𝑇𝑖𝑚𝑒 is significantly negative at the 10% level in the high contract
environment, but has a limited impact in the low contract environment. The results suggest
that the policy significantly increases the tax compliance of platform companies, but the
effect is not significant in the low contract environment.
3.3.2. Corporate social responsibility

This paper measures the level of corporate social responsibility using Hexun’s Social
Responsibility Rating Score, where a higher score indicates that the company is more socially
responsible. In this paper, the sample is divided into two groups of high and low social
responsibility according to the median social responsibility score for regression, and the
results in Table no. 6 are obtained. Based on columns (3) and (4) of table no. 6 we can see
that the policy significantly increases tax compliance for actively socially responsible
platform firms, but has limited impact on socially irresponsible platform companies.
3.3.3. Financing constraints

This paper draws on Kaplan and Zingales (1997) to construct the KZ index to measure the
degree of financing constraints of firms. The higher the KZ index, the higher the degree of
financing constraints faced by the firms. This paper divides the sample into two groups of
high and low according to the median financing constraint and obtains the results in Table
no. 6. Based on columns (5) and (6) of Table no. 6, we can see that there is a significant
incentive for firms with low financing constraints to improve their tax compliance, while the
effect on firms with high financing constraints is limited.

3.3.4. Market competition
This paper constructs the Hirschman Index HHI using the sum of the squares of the total
assets of all firms in the industry/total industry assets, where a higher value of the index
indicates a lower level of market competition faced. In this paper, the sample is divided into
two groups of high and low market competition according to the median market competition
for regression, and the results are given in Table no. 6 are obtained. The coefficient in column
(7) is significantly negative at the 10% level and the coefficient in column (8) is significantly
negative at the 1% level, with a significant effect. The results suggest that tax compliance
can be significantly incentivised for platform firms facing low levels of market competition,
while the effect is not significant for firms facing high levels of market competition.
Table no. 6. Regression results of contractual environmental heterogeneity
(1) (2) （3）（4）（5）（6）（7）（8）
High Low High Low High Low High Low
contrac contra social social financin financin market market
Variables tual ctual responsi responsi g g competi competi
enviro enviro bility bility constrai constrai tion tion
nment nment nts nts
TAX TAX TAX TAX TAX TAX TAX TAX
- - -
-
-0.0050 0.0098** 0.0003 -0.0010 0.0072** -0.0065* 0.0059**
0.0052* * * *
(- (- (- (- (- (- (-
(0.0850)
1.8868) 1.1917) 3.8906) 0.2887) 2.9165) 1.7001) 2.5800)
Control
Yes Yes Yes Yes Yes Yes Yes Yes
variables
Year Yes Yes Yes Yes Yes Yes Yes Yes
𝜇 Yes Yes Yes Yes Yes Yes Yes Yes
N 3466 3375 3493 3348 3421 3420 3538 3303
R2 0.734 0.703 0.741 0.755 0.749 0.740 0.691 0.691
Note: *, **, and *** denote 10%, 5%, and 1% significance levels, respectively; t-statistics are given in parentheses.
Conclusions
Our study finds that the implementation of the policy significantly reduces the
incompleteness of the tax compliance contract of the platform firms. Further mechanism tests
show that the two main mechanisms, information effect and cost effect, improve the tax
compliance contracts of platform enterprises. Heterogeneity analysis shows that the effect of
the policy to improve the completeness of the tax compliance contracts of the platform
enterprises is more pronounced for platform enterprises with a good contractual environment,
active social responsibility, low financing constraints, and low market competition. We
expand the research perspective on tax compliance in the platform economy, fill a gap in the
existing literature, and have enriched the research on incomplete contract theory. Due to the
availability of data, this paper only considers the information and transaction cost
mechanisms of tax compliance for platform companies from the perspective of incomplete
contracts and could be expanded in the future to investigate other action mechanisms.
Vol. 26 • No. 65 • February 2024 341

Acknowledgement
This work is supported by the project of the National Social Science Fund of China
(No. 20BJY224). Disclosure statement: All authors reported no potential conflicts of interest.
References
Agrawal, D.R., 2021. The internet as a tax haven?. American Economic Journal: Economic
Policy, 13(4), pp. 1-35.
Agrawal, D.R. and Fox, W.F., 2017. Taxes in an e-commerce generation. International Tax
and Public Finance, [e-journal] 24(5), pp. 903-926. https://doi.org/10.1007/s10797-016-
9422-3.
Agrawal, D.R. and Fox, W.F., 2021. Taxing goods and services in a digital era. National Tax
Journal, [e-journal] 74(1), pp. 257-301. https://doi.org/10.1086/712913.
Alm, J., 2021. Tax evasion, technology, and inequality. Economics of Governance,
[e-journal] 22(4), pp. 321-343. https://doi.org/10.1007/s10101-021-00247-w.
Armstrong, C.S., Blouin, J.L., Jagolinzer, A.D. and Larcker, D.F., 2015. Corporate
governance, incentives, and tax avoidance. Journal of Accounting & Economics,
[e-journal] 60(1), pp. 1-17. https://doi.org/10.1016/j.jacceco.2015.02.003.
Barbu, C.M., Florea, D.L., Ogarca, R.F. and Barbu, M.C.R., 2018. From ownership to access:
how the sharing economy is changing the consumer behavior. Amfiteatru Economic,
[e-journal] 20(48), pp. 373-387. https://doi.org/10.24818/EA/2018/48/373.
Baron, R.M. and Kenny, D.A., 1986. The moderator mediator variable distinction in social
psychological-research - conceptual, strategic, and statistical considerations. Journal of
Personality and Social Psychology, [e-journal] 51(6), pp. 1173-1182.
https://doi.org/10.1037/0022-3514.51.6.1173.
Bertrand, M. and Mullainathan, S., 2003. Enjoying the quiet life? Corporate governance and
managerial preferences. Journal of Political Economy, [e-journal] 111(5), pp. 1043-1075.
https://doi.org/10.1086/376950.
Boler, E.A., Moxnes, A. and Ulltveit-Moe, K.H., 2015. R&D, International Sourcing, and
the Joint Impact on Firm Performance. American Economic Review, [e-journal] 105(12),
pp. 3704-3739. https://doi.org/10.1257/aer.20121530.
Chen, Q., Chen, X., Schipper, K., Xu, Y. and Xue, J., 2012. The sensitivity of corporate cash
holdings to corporate governance. Review of Financial Studies, [e-journal] 25(12),
pp. 3610-3644. https://doi.org/10.1093/rfs/hhs099.
China Bureau of Statistics, 2022. Times Online. [online] Available at:
<http://www.stats.gov.cn/> [Accessed 18 July 2022].
China Internet Network Information Center, 2021. The 48th Statistical Report on the
Development Status of the Internet in China. [online] Available at:
<https://www.cnnic.net.cn/n4/2022/0401/c88-1132.html> [Accessed 18 July 2022].
Desai, M.A. and Dharmapala, D., 2006. Corporate tax avoidance and high-powered
incentives. Journal of Financial Economics, [e-journal] 79(1), pp. 145-179.
https://doi.org/10.1016/j.jfineco.2005.02.002.
Desai, M.A. and Dharmapala, D., 2009. Corporate tax avoidance and firm value. The review
of Economics and Statistics, [e-journal] 91(3), pp. 537-546. https://doi.org/
10.2139/ssrn.689562.

Durnev, A., Morck, R., Yeung, B. and Zarowin, P., 2003. Does greater firm-specific return
variation mean more or less informed stock pricing? Journal of Accounting Research,
[e-journal] 41(5), pp. 797-836. https://doi.org/10.1046/j.1475-679X.2003.00124.x.
Dyreng, S.D., Hanlon, M. and Maydew, E.L., 2010. The effects of executives on corporate
tax avoidance. Accounting Review, [e-journal] 85(4), pp. 1163-1189.
https://doi.org/10.2308/accr.2010.85.4.1163.
Heckman, J.J., Ichimura, H. and Todd, P., 1998. Matching as an econometric evaluation
estimator. Review of Economic Studies, [e-journal] 65(2), pp. 261-294.
https://doi.org/10.1111/1467-937x.00044.
Jin, L. and Myers, S.C., 2006. R2 around the world: New theory and new tests. Journal of
financial Economics, [e-journal] 79(2), pp. 257-292. https://doi.org/10.1016/
j.jfineco.2004.11.003.
Kaplan, S.N. and Zingales, L., 1997. Do investment-cash flow sensitivities provide useful
measures of financing constraints? Quarterly Journal of Economics, [e-journal] 112(1),
pp. 169-215. https://doi.org/10.1162/003355397555163.
Li, P., Lu, Y. and Wang, J., 2016. Does flattening government improve economic
performance? Evidence from China. Journal of Development Economics, [e-journal] 123,
pp. 18-37. https://doi.org/10.1016/j.jdeveco.2016.07.002.
Liang, Q., Li, Q., Lu, M. and Shan, Y., 2021. Industry and geographic peer effects on
corporate tax avoidance: Evidence from China. Pacific-Basin Finance Journal,
[e-journal] 67. https://doi.org/10.1016/j.pacfin.2021.101545.
Ma Argiles-Bosch, J., Ravenda, D. and Garcia-Blandon, J., 2021. E-commerce and labour
tax avoidance. Critical Perspectives on Accounting, [e-journal] 81.
https://doi.org/10.1016/j.cpa.2020.102202.
Martin-Martin, D., Maya Garcia, J. and Romero, I., 2022. Determinants of digital
transformation in the restaurant industry. Amfiteatru Economic, [e-journal] 24(60), pp.
430-446. https://doi.org/10.24818/EA/2022/60/430.
National Information Center Sharing Economy Research Center, 2019. Report on the
Development of China' s Sharing Economy (2019). [pdf] National Information Center
Sharing Economy Research Center. Available at: <http://www.sic.gov.cn/
sic/82/568/0301/8232.pdf> [Accessed 18 June 2022].
National Information Center Sharing Economy Research Center, 2020.Report on the
Development of China's Sharing Economy (2020). [pdf] National Information Center
Sharing Economy Research Center. Available at: <http://www.sic.gov.cn/sic/82/
568/0309/10425_pc.html> [Accessed 18 June 2022].
Niu, B., Dong, J., Dai, Z. and Liu, Y., 2022. Sales data sharing to improve product
development efficiency in cross-border e-commerce. Electronic Commerce Research and
Applications, [e-journal] 51, article no. 101112. https://doi.org/10.1016/
j.elerap.2021.101112.
Rodrigues, V.F., Policarpo, L.M., da Silveira, D.E., da Rosa Righi, R., da Costa, C.A.,
Barbosa, J.L.V., Antunes, R.S., Scorsatto, R. and Arcot, T., 2022. Fraud detection and
prevention in e-commerce: A systematic literature review. Electronic Commerce
Research and Applications, [e-journal] article no. 101207. https://doi.org/
10.1016/j.elerap.2022.101207.
Vol. 26 • No. 65 • February 2024 343

Rosenbaum, P.R. and Rubin, D.B., 1983. The central role of the propensity score in
observational studies for causal effects. Biometrika, [e-journal] 70(1), pp. 41-55.
https://doi.org/10.1093/biomet/70.1.41.
Shao, X., Chen, S., Song, Y. and Yan, S., 2022. Research on the Tax Compliance Behavior
of Online Anchors in the Sharing Economy. Transformations in Business & Economic,
21(3C), pp. 431-450.
Spinosa, L. and Chand, V., 2018. A long-term solution for taxing digitalized business models:
Should the permanent establishment definition be modified to resolve the issue or should
the focus be on a shared taxing rights mechanism? Intertax, 46(6-7), pp. 476-494.
Stabrowski, F., 2017. ‘People as businesses’: Airbnb and urban micro-entrepreneurialism in
New York City. Cambridge Journal of Regions, Economy and Society, [e-journal] 10(2),
pp. 327-347. https://doi.org/10.1093/cjres/rsx004.
Thomas, K.D., 2018. Taxing the gig economy. University of Pennsylvania Law Review,
166(6), pp. 1415-1473.
Wen, W., Cui, H. and Ke, Y., 2020. Directors with foreign experience and corporate tax
avoidance. Journal of Corporate Finance, [e-journal] 62. https://doi.org/10.1016/
j.jcorpfin.2020.101624.
Xu, W., Zeng, Y. and Zhang, J., 2011. Tax enforcement as a corporate governance
mechanism: Empirical evidence from China. Corporate Governance-an International
Review, [e-journal] 19(1), pp. 25-40. https://doi.org/10.1111/j.1467-8683.2010.00831.x.
Zha, X., Zhang, X., Liu, Y. and Dan, B., 2022. Bonded-warehouse or direct-mail? Logistics
mode choice in a cross-border e-commerce supply chain with platform information
sharing. Electronic Commerce Research and Applications, [e-journal] 54, article no.
101181. https://doi.org/10.1016/j.elerap.2022.101181.
Zhang, L., Yang, W., Yang, L. and Gao, L., 2022a. How does the value co-creation of
e-commerce platform and manufacturers affect channel performance? The moderating
role of innovation capability and market turbulence. Electronic Commerce Research and
j.elerap.2022.101120.
Zhang, T., Tang, Z. and Han, Z., 2022b. Optimal online channel structure for multinational
firms considering live streaming shopping. Electronic Commerce Research and
j.elerap.2022.101198.

Contents
Amfiteatru Economic recommends

Cultured Meat – From Scientific Challenge to Consumer Acceptance
as Sustainable Food Source ............................................................................................. 346
Lelia Voinea, Dorin Vicențiu Popescu, Teodor Mihai Negrea and Răzvan Dina
Architecture to Transform Classic Academic Courses into Adaptive Learning

Flows with Artificial Intelligence .................................................................................... 363
Andrei Bobocea, Răzvan Bologa, Lorena Bătăgan and Bogdan-Ștefan Posedaru
Victor Tufescu or the Perfect Elegance of Man and Spirit ........................................... 381
Silviu Neguţ
Vol. 26 • No. 65• February 2024 345

AE Cultured Meat – From Scientific Challenge to Consumer Acceptance
as Sustainable Food Source
CULTURED MEAT – FROM SCIENTIFIC CHALLENGE TO CONSUMER

ACCEPTANCE AS SUSTAINABLE FOOD SOURCE
Lelia Voinea1 , Dorin Vicențiu Popescu2 , Teodor Mihai Negrea3
and Răzvan Dina4*
1)2)3)4)

Voinea, L., Popescu, D.V., Negrea, T.M. and Dina, Received: 25 September 2023
R., 2024. Cultured Meat – From Scientific Challenge Revised: 24 November 2023
to Consumer Acceptance as Sustainable Food Source. Accepted: 20 December 2023
https://doi.org/10.24818/EA/2024/65/346
Abstract
Cultured meat obtained in vitro from animal cells represents one of the recent concerns of
specialists, who thus try to solve part of the issues related to animal welfare and sustainable
meat production. Beyond the technical difficulties that must be overcome, the acceptance of
cultured meat by consumers is not an easy goal to achieve, as it depends on factors acting
both at the global and local levels. Since this paper focuses on Romanian consumers, the
authors emphasize the importance of developing a general information framework that would
allow first their understanding and further their acceptance of cultured meat. For this purpose,
a systematic literature review, based on 34 articles, was carried out in order to observe good
practices at international level, which led, in the past, to the acceptance of other significant
innovations in the food field. Thus, five directions of action were identified: communication,
knowledge, trust, perception, and attitudes. Based on these, a model of cultured meat
acceptance by Romanian was developed. This initiative is unique in Romania so far and its
results can be of interest for Romanian consumers and also for many categories of specialists,
from scholars and researchers to practitioners interested in launching new products on the
market or policymakers in the field of nutrition, public health, and agriculture. Further
research is needed on the relationships between food security and safety, traditional meat
production, and cultured meat as an alternative to real meat, before the marketplace launch
of this new product.
Keywords: cultured meat, Romanian consumers, model of consumer acceptance of cultured
meat, sustainable food consumption
JEL Classification: D12, D18, M31, O33, I12
*
Corresponding author, Răzvan Dina – e-mail: [email protected]

Amfiteatru Economic recommends AE
Introduction
Meat overconsumption, especially the red one, specific to nowadays’ Western diet has had a
negative impact both on environment and consumers’ health (Allievi, Vinnari and
Luukkanen, 2015; De Graaf, 2019). Indeed, it is widely accepted that food of animal origin
has a greater impact on the environment than food of plant origin, since it emits higher
amounts of greenhouse gases (GHGs), requires more soil and nitrogen, and affects terrestrial
and aquatic biodiversity. From an environmental perspective, meat is considered the most
harmful form of food consumption of all foods of animal origin (Austgulen et al., 2018). In
addition, meat, alongside other energy-rich foods, is an important factor in the wide spread
of non-communicable diseases (NCDs), such as type 2 diabetes, coronary diseases, obesity,
cancer (Pieniak et al., 2016; Bobe, Procopie and Bucur, 2019). An overconsumption of meat
particularly affects the hormonal, the immunity, and the digestive systems, mainly because
the meat produced on a large scale, in industrial conditions, has, in most of the cases,
hormones and antibiotics. In this context, the transition toward a low-meat diet appears to be
the right choice from a human health, animal welfare and environmental protection point of
view. On the basis of this increased call to reduce meat consumption, a new trend of industrial
meat substitutes has emerged, where cultured meat or lab-grown meat, which represents the
focus of our research, is a part of.
Cultured meat (also called lab-grown meat, cultivated meat, artificial meat or in vitro meat)
is an innovative product of cellular agriculture, obtained from stem cells, which, after being
taken from animals through a small biopsy of muscle tissue, are placed in a special nutrient-
rich medium in a bio-reactor independent from the animal to promote growth (Post, 2012;
Ben-Arye and Levenberg, 2019).
Although it is not yet available to consumers, cultured meat is described in recent literature
as having positive effects in several directions, such as: environmental sustainability (because
it would reduce greenhouse gas emissions from agriculture while using significantly less
water and less land), animal welfare (because it is produced without the need for animals to
be raised and slaughtered), food safety (because it would be less susceptible to antibiotic
resistance and to zoonotic diseases spread), and world hunger issue (because it would prevent
meat becoming a luxury food for most consumers) (Post, 2012; Sinke et al., 2023). Thus,
cultured meat appears to have the potential to address the main challenges that meat industry
is currently facing, providing potentially healthier, ethical, sustainable, and more efficient
alternative to conventionally produced meat. But cultured meat also comes with drawbacks;
the most frequently reported in the scientific literature refer to the safety concerns generated
by the use of a potentially contaminated animal serum in the culture medium (Alexander et
al., 2017), the higher energy use in cultured meat production because some biological
functions are replaced by technological processes (Munteanu et al., 2021) and the potential
perception of in vitro technology as unnatural and interfering too much with nature
(Laestadius, 2015; Stephens et al., 2018). Moreover, Chriki and Hocquette (2020) noticed
the vague status of cultured meat from a religious point of view, its compliance with certain
dietary laws (e.g., Kosher for Jewish consumers or Halal for Islamic consumers) being still
debatable.
Because the success of new foods and food technologies depend to a considerable extent on
whether consumers accept those innovations (Fischer and Reinders, 2022; Socaciu et al.,
2022) and given the tendency of people to be reluctant to adopt food innovations, perceiving
them as too industrial, unnatural or “fake” (EIT Food Trust Report, 2021), consumer
Vol. 26 • No. 65 • February 2024 347

acceptance of cultured meat is an issue that requires further attention. Due to its novelty,
consumer acceptance is extremely important for cultured meat, because, as Laestadius (2015)
pointed out, without it there might be a product but not a market.
Although there is abundant information in scientific literature regarding the advanced
technological aspects of cultured meat, significantly little research has been done with regard
to its acceptance by consumers. However, here are to be mentioned the studies of Laestadius
(2015), Verbeke et al. (2015), Stephens et al. (2018), Bryant and Barnett (2020) and Siegrist
and Hartmann (2020) that investigated consumers’ reactions towards cultured meat in the
early stage of product development. Although consumers seemed open to try cultured meat,
the common insights provided by these studies showed feelings of disgust and considerations
of unnaturalness exhibited by consumers and highlighted consumer acceptance as important
issues related to the commercial success of cultured meat.
In Romania, alongside consumers’ natural tendency of reluctance towards food novelty,
another important obstacle in the acceptance of cultured meat could be the privileged status
of natural meat in local food culture, emphasised by the high share of meat-based products
and dishes in the traditional pattern of food consumption (Voinea et al., 2020).
Because, to the authors’ knowledge, there is no research addressing the issue of consumer
acceptance of cultured meat in Romania until now, the aim of present paper is to fill this gap
by designing a general framework for consumer acceptance which will contribute to a correct
understanding of cultured meat before its commercial release. Therefore, to identify
examples of good practices successfully used in the past in similar situations, a review of the
specialised literature was carried out. This allowed us to select five directions of action as a
basis for designing the model of cultured meat acceptance. This approach, unique in Romania
so far, can be useful to consumers, specialists, and other decision-makers in the field of
nutrition, public health, or agriculture.
Although paper’s rhetoric is focused on cultured meat, our study should not be regarded as a
pleading for agriculture transition to cultured meat production as the primary solution for the
sustainable diet of the future and for cultured meat as the best alternative for conventional
meat. The paper intends to bring to the foreground the need to inform and educate consumers
so that when the product will be available on the market, they can make an informed purchase
decision.
The study begins with an introductory section providing some background elements and
presenting the general framework of contemporary food that has led to the development of
cultured meat. Items of previous research in the area are also reviewed in this section.
Furthermore, the methodological approach is presented, emphasising the steps followed in
conducting the study. Within the next section, the results are presented and their significance
is discussed. Finally, there is a section comprising the conclusions of the study, and its
implications.
1. Methodology
As the main purpose of our study is to develop a model of consumer acceptance of cultured
meat, we conducted a systematic review of the scientific literature to uncover different best

practices used to achieve consumer acceptance for several previous food innovations. Based
on the identified predictors, we will further build a model for consumer acceptance of
cultured meat, the newest food innovation that consumers are to deal with in the near future.
The review followed the steps outlined in literature (Khan et al., 2003; Dziopa and Ahern,
2011; Xiao and Watson, 2019): formulating the question, identifying relevant publications,
assessing study quality, summarising the evidence, and interpreting the findings.
In the attempt to answer to the question “What were the predictors of consumer acceptance
of previous significant food innovations?”, we firstly established the following food
innovations as areas of interest for our approach: genetic modification of food, food
irradiation, entomophagy and food nanotechnology.
Numerous articles, empirical studies or review papers, were identified in international
databases, using key-words such as: “acceptance of new foods”, “acceptance of new food
technologies”, “acceptance of biotechnology”, “acceptance of genetically modified foods”,
“acceptance of food irradiation”, “acceptance of irradiated foods”, “acceptance of edible
insects”, “acceptance of nanotechnology foods”. A number of 34 articles from all those
identified were selected as the most suitable for the purpose of our research, based on the
following criteria: focus on consumer acceptance, presentation of empirical data or findings
of the literature review, publication in a peer-reviewed journal, publication in English
language. Then, the selected articles were grouped into five categories, as follows: genetic
modification, food irradiation, entomophagy, the use of nanotechnology, and new foods and
food technologies in general (Table no. 1). A summary of the evidence was done for each
category from which a key-word predictor of consumer acceptance was extracted.
Table no. 1. Results of the literature review
Key-factors
Food
No. Study Summary of evidence of consumer
innovation
acceptance
1 New foods Frewer, Howard and The way consumers will evaluate a Communi-
and food Sheperd (1995); new food product depends on cation
technologies Siegrist (2007); effective communication of both
in general Costa-Font, Gil tangible benefits and hazards.
and Traill (2008);
Shepherd (2008)
Siegrist, Gutscher and Consumers tend to rely on the Trust
Earle (2005); information provided by experts
Macready et al. (2020); (scientists, authorities, and industry);
Wu et al. (2021); high levels of trust reduce the
The EIT Food Trust perceived risks.
Report (2021)
Alhakami and Slovic Food technologies that are judged The
(1994); Rembischevski high in risk tend to be judged low in perception of
and Caldas (2020); benefit, and vice versa. benefits and
Rabbanee, Afroz risks
and Naser (2021)
Siegrist (2007); Frewer Positive attitudes towards a new food Attitudes
et al. (2014); Román, or technology are associated with
Sánchez-Siles favouring behaviour, while negative
and Siegrist (2017); attitudes are associated with
Bearth and Siegrist avoidance behaviour.
(2019)
Vol. 26 • No. 65 • February 2024 349

Key-factors
Food
No. Study Summary of evidence of consumer
innovation
acceptance
2 Genetically Hoban (1997); A positive correlation between Knowledge
modified Byrne et al. (2002); knowledge of biotechnology and
foods House et al. (2002) consumer acceptance of genetically
Lusk and Sullivan modified foods.
(2002); Costa-Font,
Gil and Traill (2008);
Wunderlich and Gatto
(2015)
Frewer, Howard The acceptance of genetically The
and Sheperd (1995) modified foods is largely determined perception of
by both perceived risks and benefits. benefits and
risks
Tenbült et al. (2005) The more a product was seen as Attitudes
Frewer et al. (2014) natural by consumers, the less
acceptable was its genetic engineered
version.
Lusk and Sullivan The acceptance of the genetically Trust
(2002); Costa-Font modified foods starts from the trust in
and Gil (2009) scientists and in public authorities, as
they are considered the most reliable
sources.
3 Edible Lensvelt and Knowledge about entomophagy is Knowledge
insects Steenbekkers (2014) positively correlated with acceptance
Woolf et al. (2020) and willingness to consume edible
Padulo et al. (2022) insects.
Ruby, Rozin and Chan Beliefs about the benefits and risks of The
(2015) eating insects represent one of the perception of
best predictors of insect acceptance. benefits and
risks
4 Food Bruhn (1998); Oliveira Knowledge is positively correlated Knowledge
irradiation and Sabato (2004) with the acceptance of irradiated
Behrens et al. (2009) food.
Sapp and Downing- Acceptance of irradiated foods is a Trust
Matibag (2009) function of trust in societal
institutions that they can manage
irradiation technology.
Bord and O'Connor When food irradiation is perceived as The
(1990); Bearth and a risky technology, fewer benefits perception of
Siegrist (2019) and more risks are associated with benefits and
irradiated foods risks
5 Nano- Besley (2010); Gomez- The more knowledge, the less fear for Knowledge
technology Llorente et al. (2022) nanotechnology foods; awareness is
associated with more positive views
about nanotechnology
Gupta, Frewer Consumers tend to rely to a greater Trust
and Fischer (2017) extent on the information provided by
experts.
Siegrist (2007); Willingness to buy nanotechnology The
Stampfli, Siegrist foods is strongly influenced by perception of
and Kastenholz (2010); perceived benefits and weakly benefits and
Besley (2010) influenced by perceived risks. risks

The review of the scientific literature in the field of food innovations led to the identification
of the following key-factors of consumer acceptance: communication, knowledge, trust,
perception of benefits and risks and attitudes.

The model of consumer acceptance of cultured meat is designed to serve as an instrument for
consumers in making informed decision and for the actors in food chain in approaching the
cultured meat-related communication strategy.
Based on the results of the undertaken systematic literature review we built up the cultured
meat acceptance model based on the following drivers: communication, knowledge, trust,
perception of benefits and risks, and attitudes. (Figure no. 1)
Communication
Knowledge
Trust
Perception
Positive perception Negative perception
High High
benefits risks
Attitude
Positive attitude Negative attitude
Acceptance Rejection
Figure no.1. The cultured meat acceptance model
Communication
New technologies enable innovations in the food sector. Since, in general, the benefits
associated with an innovation in the food offer are not tangible for consumers, they must be
explicitly communicated to the large public. This has never been an easy task, but it is all the
more difficult when it comes to a specific technology that tends to evoke negative
associations in consumer’s mind (Siegrist, 2007, p. 240).
According to Costa-Font, Gil and Traill (2008), the way consumer will evaluate a new food
product depends on effective provision of information regarding firstly its tangible benefits
from several sources – public and private, formal and informal, etc. – being conditioned at
the same time by the credibility and trustworthiness of each relevant information source. But,
alongside the benefits, the drawbacks and hazards have to be communicated to the public as
well. Frewer, Howard and Shepherd (1995) pointed out that it is important the dialogue be
Vol. 26 • No. 65 • February 2024 351

established between communicators and the large public, in such way that the addressed
issues reflect the real concerns of the public. Shepherd (2008) advocated the involvement of
the public in the management of food risk issues, citing the Fischhoff ‘s (1995) schema of
risk communication where consumers are seen as partners in this process.
Based on previous communication experience regarding food innovations, we consider that
an effective communication about cultured meat should focus primarily on the benefits,
emphasising its potential of mimicking the sensory and nutritional characteristics of natural
meat, as well as its positive impact on human health, animal welfare, economic efficiency,
and environmental sustainability. Similarly, risk messages should address specific concerns
related to food safety, lack of naturalness, and energy use in order to be more relevant to
Romanian consumers. In this process, the main responsibility lies with the entrepreneurs in
the food industry, who should mainly focus on providing correct and honest data on the
benefits and risks of artificial meat consumption and not just on its exclusively commercial
promotion.
Knowledge
Research has shown that consumer knowledge of new products or technologies correlates
positively with acceptance and willingness to consume those products.
For example, prior research on consumer acceptance of edible insects (Lensvelt and
Steenbekkers, 2014; Ruby, Rozin and Chan, 2015; Woolf et al., 2021; Padulo et al., 2022)
demonstrated that participants who attended an educational session that provided information
about entomophagy, followed by a cooking and tasting session of edible insects, felt more
knowledgeable about entomophagy, which positively correlated with acceptance and
willingness to consume edible insects.
Several surveys that have been carried out worldwide (Bruhn, 1998; Oliveira and Sabato,
2004; Behrens et al., 2009) to investigate consumers’ attitude to food irradiation found out
that knowledge is positively correlated with acceptance of irradiated food. Generally, results
suggest that when provided with science-based information, people hold more positive
attitudes toward food irradiation and tend to favour irradiated food.
The study of Gómez-Llorente et al. (2022) on consumer perception on the use of
nanotechnology in the agri-food sector, showed that the more knowledge, the less fear for
nanotechnology foods. This resonates with the findings of Besley (2010), who reported that
awareness is associated with more positive views about nanotechnology.
The findings of a large number of studies on consumer acceptance of biotechnology and
genetically modified foods (Hoban, 1997; Byrne et al., 2002; Lusk and Sullivan, 2002; House
et al., 2004; Costa-Font and Gil, 2009; Wunderlich and Gatto, 2015) highlighted a positive
correlation between knowledge of biotechnology and consumer acceptance of genetically
modified foods, since people with higher levels of scientific knowledge were usually more
optimistic about biotechnology.
The same positive correlation was revealed by Rolland, Markus and Post (2020) in their study
on the effect of information content on the acceptance of cultured meat, as they observed that
provision of information to participants, especially on personal benefits (with regard to
product safety) and societal benefits (related to reducing environmental impact, avoiding

harm to animals or reducing world hunger), importantly improves acceptance and willingness
to taste. They concluded awareness of cultured meat was the best predictor of its acceptance.
Similar findings can be found at Bryant and Barnett (2020) who argued that acceptance of
cultured meat could be increased with positive information.
Given the predominant eating behaviour in many areas of the world, including Romania,
characterised by a high consumption of meat and meat-based products, we consider that a
direction of increasing consumer acceptance of cultured meat could be a strong and well-
argued emphasis on the potential of cultured meat to mimic the nutritional value of real meat
from the actors in the food system. In addition, as it has been proven that consumers’ high
level of scientific knowledge plays a decisive role in building a favourable attitude towards
innovative food products, we believe that academia, nationally and internationally recognised
research institutes, as well as experts in the field of food safety are the driving factors in the
transition from a favourable attitude towards cultured meat to its acceptance and further
consumption.
Trust
Consumer trust is an important factor influencing the perception of food innovation: the
perceived risks as well as on perceived benefits. As long as acceptance of new foods and
willingness to buy them is directly determined by the perceived risks and the perceived
benefits, trust can be said to have an indirect impact on the acceptance of new foods or new
food technologies (Siegrist, 2007, p. 242).
Both risks and benefits, the latter being embodied in the characteristics of the new products,
such as health, sustainability, or safety, since they cannot be verified by the average
consumer, must therefore be communicated by a credible actor in the food chain (Macready
et al., 2020; EIT Food Trust Report, 2021; Wu et al., 2021). Therefore, as Gupta, Frewer and
Fischer (2017) pointed out, when consumer knowledge and experience on a new food product
and technology are limited, consumers tend to rely to a greater extent on the information
provided by experts.
As for the trustworthiness of different information sources, Lusk and Sullivan (2002), in their
study on consumer acceptance of genetically modified food, claim that consumers tend to
trust more the information from government institutions, such as USDA (United States
Department for Agriculture) or FDA (Food and Drug Administration), considering that these
actors are much more likely to be credible in disseminating important information about
biotechnology and genetically modified foods than mass media and agribusiness industry.
We found similar findings in Costa-Font and Gil (2009), who showed the reasoning process
that underpins the acceptance of the genetically modified foods departs from trust in scientists
and in public authorities.
The influence of consumer trust in scientists, authorities, and industry on consumer
perception of risks and benefits was studied by Saba and Messina (2003) in relation to the
use of pesticides in agriculture, confirming that trust had a positive influence on perceived
benefits and a negative influence on perceived risks. The study of Siegrist, Gutscher and Earle
(2005) emphasised that trust has an impact on the perception of new technologies, in the
sense that it exerts a negative influence on the perception of risk. High levels of trust reduced
perceived risks, compared to low levels of trust.
Vol. 26 • No. 65 • February 2024 353

As for cultured meat, Bryant and Barnett (2020) argued that trust in the regulatory bodies
responsible for ensuring food safety is all the more important for its acceptance.
Consequently, in order to raise the level of trust, we believe that the responsibility of
informing Romanian consumers cannot be the exclusive task of entrepreneurs, but it must be
shared with high credibility experts, government bodies such as the Ministry of Health, the
Ministry of Agriculture, the National Authority for Consumer Protection or the National
Sanitary-Veterinary and Food Safety. These bodies may be involved in adapting the existing
legislative framework or drafting new regulations related to the production, marketing and
fair labelling of cultured meat, respectively in developing new methods that allow
quantitative and qualitative measurements of the chemical components of this new product.
Perception of benefits and risks

Judgments of risk and benefit related to a specific issue have been found to be inversely
related in many pieces of research. Alhakami and Slovic (1994) in their study on the
relationship between perceived risk and perceived benefit showed technologies that are
considered high risk tend to be considered low benefit and vice versa.
A great deal of study examined the influence of people’s perceptions of risk and benefit on
their acceptance of new food products and technologies. While Frewer, Howard and
Shepherd (1995), concluded in their study that the most important determinant of consumer
acceptance of genetic engineering is likely to be perceptions of benefit resulting from the
application of this technology, the study of Siegrist, Gutscher and Earle (2005) and Siegrist
(2007) revealed that acceptance of genetically modified foods is largely determined by both
perceived risks and benefits associated with biotechnology.
A common conclusion of several studies investigating the acceptance of nanotechnology
foods (Siegrist, 2007; Besley, 2010; Stampfli, Siegrist and Kastenholz, 2010) is that
perceived benefits appear to be the most important predictor for willingness to buy. This is
strongly influenced by perceived benefits and weakly influenced by perceived risks.
In the case of edible insects, as it was demonstrated in the study of Ruby, Rozin and Chan
(2015), the beliefs about benefits (nutrition and environmental sustainability) and risks
(disease and illness) of entomophagy were indicated among the best predictors of consumer
acceptance.
However, as Rembischevski and Caldas (2020) highlighted in their review analysis,
consumer perception of food-related benefits and risks is a very complex issue, depending
not only on the objective and measurable aspects, but also on subjective aspects that define
individual’s values, such as social, cultural, psychological, ethical or moral ones. More than
scientific knowledge, the emotional side of individuals strongly contributes to the perception
of the ratio between risks and benefits. This issue was addressed by Bearth and Siegrist
(2019) in their research on consumer acceptance of food irradiation. They found that
consumers who exhibited negative emotions towards nuclear power perceived less benefits
and more risks associated with irradiated foods and expressed a lower overall acceptance of
this technology. Similar findings were previously revealed by the study of Bord and
O’Connor (1990), who showed that food irradiation had been perceived as a risky
technology, as consumers had associated it with nuclear power.

In the case of cultured meat, as Bryant and Barnett (2020) reported in their study, consumer
acceptance could be increased with frames which invoke more positive associations.
Therefore, we believe that building a favourable perception leading to the acceptance of
cultured meat is possible if consumers act on the principle of cognitive coherence. Thus, only
by achieving such consistency in beliefs, they can end up adopting beneficial decisions, based
both on objective, measurable elements (such as official communication, correct declaration
of chemical composition, educational sessions supported by experts where the risks and
benefits of consumption are extensively presented, tasting sessions, etc.) and on subjective
ones (such as values, culture, ethics, psychological motivations, etc.), also specific to human
nature. From an entrepreneurial point of view, the development of a successful cultured meat
business directly depends on consumers’ perceptions regarding the benefits and risks
associated with the consumption of this new innovative product. For this reason, it is
recommended that these perceptions be investigated through complex sociological and
anthropological research, which, being able to go beyond the simple quantitative approach
of market mechanisms, can provide valuable premises for the development of future
communication strategies.
Attitudes
The concept of attitude has been used in research to explain public reactions to new foods
and technologies (Frewer and Miles, 2003). Positive attitudes towards a new food or
technology are associated with favouring behaviour, while negative attitudes are associated
with avoidance behaviour (Siegrist, 2007, p. 237).
According to Ajzen (2001), attitude is a summary evaluation of an object, which can vary
from positive to negative and is experienced as an affect. The same idea was emphasised by
Southerton (2011) who defined attitude as an individual’s tendency to evaluate an object as
positive or negative. Generally, attitudes help people to make sense and give meaning to their
experiences.
In the attitude formation process, Costa-Font, Gil and Traill (2008) emphasised the leading
role of knowledge, claiming that the attitude is based on knowledge about the product itself
as well as its attributes. Based on Fishbein Multiple Attribute Model (Fishbein, 1963),
probably one of the most widely accepted theory of consumer attitude formation, Moinpour
and MacLachlan (1971) highlighted the role of perception, suggested that attitude towards a
product is a function of the sum of perceived attributes weighted according to the importance
they hold.
As it has been demonstrated in previous research, a highly important attribute that influences
attitude toward new foods is the perceived naturalness, which refers to product origin, the
technology and ingredients used and the properties of the final product (Román, Sánchez-
Siles and Siegrist, 2017). According to Butu et al. (2019) and Bearth and Siegrist (2019),
consumers frequently exhibit strong negative feelings of fear or even aversion toward
unfamiliar and artificially sounding food products or food technologies. For example, in the
case of genetically modified foods, as the study of Tenbült et al. (2005) and Frewer et al.
(2014) suggested, the more a product was seen as natural by consumers, the less acceptable
was its genetically engineered version.
Vol. 26 • No. 65 • February 2024 355

In the case of cultured meat, consumer perception of its naturalness has aroused scholars’
interest since the early stage of product development. Research to date (Kadim et al., 2015;
Laestadius, 2015; Laestadius and Caldwell, 2015; Verbeke et al., 2015; Hocquette, 2016;
Siegrist, Sütterlin and Hartmann, 2018; Stephens et al., 2018; Bryant et al., 2019; Bryant and
Barnett, 2020; Siegrist and Hartmann, 2020) has shown this perception can range from
natural to neither natural nor unnatural and to unnatural and emphasised perceived
naturalness as a crucial factor for the acceptance of cultured meat. While perceived
naturalness is positively associated with consumer acceptance and willingness to consume,
the perceived unnaturalness may act as a psychological barrier, contributing to low
acceptance of cultured meat.
Of all food products, naturalness is even more important for meat, because it represents one
of the four attributes that make up the moral defense mechanism of meat-eaters, known as
“4Ns justification” scheme, proposed by Piazza et al. (2015), according to which meat
consumption is perceived as “natural”, “normal”, “necessary” and “nice” (the latter referring
to hedonic pleasure derived from eating meat).
In Romania, to the inherent psychological barrier against the consumption of cultured meat,
the cultural one would be added, as natural meat is deeply embedded in local food culture,
being the main ingredient in most of traditional food products and dishes (Purcărea et al.,
2013). The study by Voinea et al. (2020) revealed that most Romanian consumers, being
typical meat-eaters, are not willing to compromise on rich taste and textural experience
provided by natural meat-based products. Because, generally, the cultural pattern of food
consumption and the traditional elements of food preferences are still important in Romania,
it is quite likely that most consumers would exhibit a negative attitude towards cultured meat,
perceiving its artificial nature in opposition to their demands for naturalness and authenticity
that traditional meat can fulfil. Therefore, the acceptance of cultured meat in Romania will
depend on how this conflict of cultural values will be resolved, both at the individual and
collective level.
However, there are also Romanian consumers aware of the meat paradox defined by Bastian
and Loughnan (2017) as a moral conflict of people concerned for animal wellbeing and yet
consume products which have caused animal suffering) and also worried about their own
health who are willing to reduce into some extent their meat consumption (meat-reducers) or
even to give it up (meat-avoiders). We estimate these latter categories of consumers are the
most likely to more easily develop a positive attitude towards cultured meat.
Overall, to increase the acceptance of cultured meat among Romanian consumers, as Siegrist
and Hartmann (2020) outlined in their study, communication strategies should be focused on
emphasising the similarity of cultured meat to traditional meat, rather than the technical
aspects of the production process, which are more susceptible to generate negative
associations and disgust feelings. In other words, this emerging technology of in vitro meat
should be framed in ways that resonate with people’s existing views (Besley, 2010). We
believe that in the process of transforming consumer perceptions into strong attitudes of
acceptance or rejection towards culture meat the associations of consumer protection play a
decisive role, as they can contribute to the correct information of consumers, through the
studies and campaigns carried out by their own experts.

Conclusions
The food development model specific to Western societies in recent decades, widely
encountered today even in Romania, is based on a diet in which highly processed and high-
energy foods are predominant. This type of eating behaviour, characterised especially by
overconsumption of red meat, possibly explainable in the first decades after World War II by
the need to increase people's living standards, has recently led to an emphasis on issues
related to the preservation of both environment and consumer health. (Allievi, Vinnari and
Luukkanen, 2015; De Graaf, 2019).
Meat is the food of animal origin that raises the most serious environmental sustainability
issues, as its production process requires a lot of soil and nitrogen and also emits the most
greenhouse gases (Austgulen et al., 2018). Moreover, meat is responsible for the dramatic
increase in the incidence of the so-called diseases of civilisation, such as obesity, type 2
diabetes, and cancer (Pieniak et al., 2016). Therefore, the need to redefine the contemporary
food style through a consistent decrease in the consumption of traditional meat has become
imperative. In this context, the production of industrial meat substitutes with the help of
avant-garde technologies has emerged as a new trend. Among these newly developed meat
analogues is cultured meat too, also called artificial meat or in vitro meat, which presents
both advantages (environmental preservation, animal welfare, food safety, and consumer
food security) and risks (potentially toxic animal serum in the culture environment, high
energy consumption in the production process, or too much interference of its technology
with nature or its uncertain status from a religious perspective, specifically in relation to
dietary laws present in some religions).
The objective analysis of the ratio between the advantages and disadvantages of cultured
meat is very important, considering the novelty of the product and the need to build a
sustainable marketplace for it. The originality of the present work lies in the study of the
acceptance potential of cultured meat among Romanian consumers, in today's food
consumption background characterised by a strong traditional imprint, since foods of animal
origin, including meat, are common items in Romanians’ diet.
The main aim of our research is to design and, thus, put into public debate the principles of
a general framework for cultured meat acceptance by Romanian consumers, before this
product becomes a reality in the food market. However, our approach should not be
understood as a plea for the large-scale transition of agriculture to the production of artificial
meat, as a unique alternative to the traditional meat consumed today. We believe that our
approach is useful, considering the biunivocal relationship between the main actors of any
market - producers and consumers - and the need of the latter to be properly educated and
informed when making purchasing decisions, which remains a necessary condition in a
market characterised by loyal commercial relationships.
To achieve the proposed goal, the methodological approach consisted of a systematic review
of the scientific literature to identify the best practices that led to consumer acceptance of
several food innovations in the past, such as: food genetic modification, food irradiation,
entomophagy, and food nanotechnology.
The scientific endeavour of the authors aims to conceive a general model of acceptance of
cultured meat by Romanian consumers, whose purchasing decisions we consider should be
made rationally and with full awareness, being correctly informed by producers who develop
these products, and whose communication and promotion strategies should adhere to the
Vol. 26 • No. 65 • February 2024 357

principles of fair trade. The proposed model is based on five key factors: communication,
knowledge, trust, perceptions of benefits and risks, and consumer attitudes.
The authors consider that it would be useful to carry out quantitative research, which would
analyse the typology of Romanian consumers and their possible predisposition to the
consumption of artificial meat, to the detriment of the natural product. At the same time, the
factors that could lead these consumers to partially change their consumption habits should
be investigated, along with the study of product range derived from artificial meat.
References
Alexander, P., Brown, C., Arneth, A., Dias, C., Finnigan, J., Moran, D. and Rounsevell,
M.D.A., 2017. Could consumption of insects, cultured meat or imitation meat reduce
global agricultural land use? Global Food Security, [e-journal] 15, pp. 22-32.
https://doi.org/10.1016/j.gfs.2017.04.001.
Alhakami, A.S. and Slovic, P., 1994. A Psychological Study of the Inverse Relationship
Between Perceived Risk and Perceived Benefit. Risk Analysis, [e-journal] 14(6),
pp. 1085-1096. https://doi.org/10.1111/j.1539-6924.1994.tb00080.x.
Allievi, F., Vinnari, M. and Luukkanen, J., 2015. Meat consumption and production –
analysis of efficiency, sufficiency and consistency of global trends. Journal of Cleaner
Production, [e-journal] 92, pp. 142-151. https://doi.org/10.1016/j.jclepro.2014.12.075.
Austgulen, M., Skuland, S., Schjøll, A. and Alfnes, F., 2018. Consumer Readiness to Reduce
Meat Consumption for the Purpose of Environmental Sustainability: Insights from
Norway. Sustainability, [e-journal] 10(9), article no. 3058. https://doi.org/10.3390/
su10093058.
Bastian, B. and Loughnan, S., 2017. Resolving the Meat-Paradox: A Motivational Account
of Morally Troublesome Behavior and Its Maintenance. Personality and Social
Psychology Review, [e-journal] 21(3), pp. 278-299. https://doi.org/10.1177/
1088868316647562.
Bearth, A. and Siegrist, M., 2019. “As long as it is not irradiated” – Influencing factors of
US consumers’ acceptance of food irradiation. Food Quality and Preference, [e-journal]
71, pp. 141-148. https://doi.org/10.1016/j.foodqual.2018.06.015.
Behrens, J.H., Barcellos, M.N., Frewer, L.J., Nunes, T.P. and Landgraf, M., 2009. Brazilian
consumer views on food irradiation. Innovative Food Science & Emerging Technologies,
[e-journal] 10(3), pp. 383-389. https://doi.org/10.1016/j.ifset.2009.01.001.
Ben-Arye, T. and Levenberg, S., 2019. Tissue Engineering for Clean Meat Production.
Frontiers in Sustainable Food Systems, [e-journal] 3(46). https://doi.org/10.3389/
fsufs.2019.00046.
Besley, J., 2010. Current research on public perceptions of nanotechnology. Emerging Health
Threats Journal, [e-journal] 3, article no. e8. https://doi.org/10.3134/ehtj.10.164.
Bobe, M., Procopie, R. and Bucur, M., 2019. Exploring the Role of Individual Food Security
in the Assessment of Population’s Food Safety. Amfiteatru Economic, [e-journal] 21(51),
pp. 347-360. https://doi.org/10.24818/EA/2019/51/347.

Bord, R.J. and O’Connor, R.E., 1990. Risk Communication, Knowledge, and Attitudes:
Explaining Reactions to a Technology Perceived as Risky. Risk Analysis, [e-journal]
10(4), pp. 499-506. https://doi.org/10.1111/j.1539-6924.1990.tb00535.x.
Bruhn, C.M., 1998. Consumer acceptance of irradiated food: theory and reality. Radiation
Physics and Chemistry, [e-journal] 52(1-6), pp. 129-133. https://doi.org/10.1016/S0969-
806X(98)00088-7.
Bryant, C. and Barnett, J., 2020. Consumer Acceptance of Cultured Meat: An Updated
Review (2018-2020). Applied Sciences, [e-journal] 10(15), article no. 5201.
https://doi.org/10.3390/app10155201.
Butu, A., Vasiliu, C.D., Rodino, S., Brumă, I.-S., Tanasă, L. and Butu, M., 2019. The
Anthropological Analysis of the Key Determinants on the Purchase Decision Taken by
the Romanian Consumers Regarding the Ecological Agroalimentary Products.
Sustainability, [e-journal] 11(18), article no. 4897. https://doi.org/10.3390/su11184897.
Byrne, P.F., Namuth, D.M., Harrington, J., Ward, S.M., Lee, D.J. and Hain, P., 2002.
Increasing public understanding of transgenic crops through the World Wide Web. Public
Understanding of Science, [e-journal] 11(3), pp. 293-304. https://doi.org/10.1088/0963-
6625/11/3/306.
Chriki, S. and Hocquette, J.-F., 2020. The Myth of Cultured Meat: A Review. Frontiers in
Nutrition, [e-journal] 7, article no. 7. https://doi.org/10.3389/fnut.2020.00007.
Costa-Font, M. and Gil, J.M., 2009. Structural equation modelling of consumer acceptance
of genetically modified (GM) food in the Mediterranean Europe: A cross country study.
Food Quality and Preference, [e-journal] 20(6), pp. 399-409. https://doi.org/10.1016/
j.foodqual.2009.02.011.
Costa-Font, M., Gil, J.M. and Traill, W.B., 2008. Consumer acceptance, valuation of and
attitudes towards genetically modified food: Review and implications for food policy.
Food Policy, [e-journal] 33(2), pp. 99-111. https://doi.org/10.1016/j.foodpol.
2007.07.002.
De Graaf, M., 2019. Poor diet kills more than smoking and hypertension: More than 500,000
Americans and 90,000 Brits a year die from too much meat and too few veggies, Bill
Gates study reveals. Dailymail. [online] 3 April. Available at:
<https://www.dailymail.co.uk/health/article-6881937/Poor-diet-kills-smoking-
hypertension-Bill-Gates-study-reveals.html> [Accessed 25 August 2019].
Dziopa, F. and Ahern, K., 2011. A Systematic Literature Review of the Applications of Q-
Technique and Its Methodology. Methodology, [e-journal] 7(2), pp. 39-55.
https://doi.org/10.1027/1614-2241/a000021.
EIT Food Trust Report, 2021. Sustainable food choices and the role of trust in the food
chain. The EIT Food Trust Report. [online] Available at: <https://www.eitfood.eu/
media/news-pdf/EIT_Food_Trust_Report_2021.pdf> [Accessed 5 December 2022].
Fischer, A.R.H. and Reinders, M.J., 2022. Consumer acceptance of novel foods. In: C.M.
Galanakis ed., 2022. Innovation Strategies in the Food Industry. 2nd ed. s.l.: Academic
Press, pp. 307-333. https://doi.org/10.1016/B978-0-323-85203-6.00013-X.
Frewer, L.J., Gupta, N., George, S., Fischer, A.R.H., Giles, E.L. and Coles, D., 2014.
Consumer attitudes towards nanotechnologies applied to food production. Trends in Food
Science & Technology, [e-journal] 40(2), pp. 211-225. https://doi.org/10.1016/
j.tifs.2014.06.005.
Vol. 26 • No. 65 • February 2024 359

Frewer, L.J., Howard, C. and Shepherd, R., 1995. Genetic engineering and food: what
determines consumer acceptance? British Food Journal, [e-journal] 97(8), pp. 31-36.
https://doi.org/10.1108/00070709510100118.
Frewer, L.J. and Miles, S., 2003. Temporal stability of the psychological determinants of
trust: implications for communications about food risks. Health, Risk and Society,
Gómez-Llorente, H., Hervás, P., Pérez-Esteve, É., Barat, J.M. and Fernández-Segovia, I.,
2022. Nanotechnology in the agri-food sector: Consumer perceptions. NanoImpact,
[e-journal] 26, article no. 100399. https://doi.org/10.1016/j.impact.2022.100399.
Gupta, N., Frewer, L.J. and Fischer, A.R.H., 2017. Acceptance of Agri-Food
Nanotechnology: Insights from the Evolution of Food Technology, Novel Foods and the
Psychology of Novel Food Acceptance and Evidence from Present Research. In: Q.
Chaudhry, L. Castle and R. Watkins eds., 2017. Nanotechnologies in food. Nanoscience
& Nanotechnology series, 2nd ed. London, UK: Royal Society of Chemistry, pp. 39-59.
Hoban, T.J., 1997. Consumer acceptance of biotechnology: An international perspective.
Nature Biotechnology, [e-journal] 15(3), pp. 232-234. https://doi.org/10.1038/nbt0397-
232.
House, L., Lusk, J.L., Jaeger, S., Traill, W.B., Moore, M., Valli, C., Morrow, B. and Yee,
W.M.S., 2004. Objective And Subjective Knowledge: Impacts On Consumer Demand For
Genetically Modified Foods In The United States And The European Union. [online]
Available at: <https://doi.org/10.22004/AG.ECON.20125> [Accessed 20 August 2023].
Khan, K.S., Kunz, R., Kleijnen, J. and Antes, G., 2003. Five Steps to Conducting a
Systematic Review. Journal of the Royal Society of Medicine, [e-journal] 96(3),
pp. 118-121. https://doi.org/10.1177/014107680309600304.
Laestadius, L.I., 2015. Public Perceptions of the Ethics of In-vitro Meat: Determining an
Appropriate Course of Action. Journal of Agricultural and Environmental Ethics,
[e-journal] 28(5), pp. 991-1009. https://doi.org/10.1007/s10806-015-9573-8.
Lensvelt, E.J.S. and Steenbekkers, L.P.A., 2014. Exploring Consumer Acceptance of
Entomophagy: A Survey and Experiment in Australia and the Netherlands. Ecology of
Food and Nutrition, [e-journal] 53(5), pp. 543-561. https://doi.org/10.1080/
03670244.2013.879865.
Lusk, J.L. and Sullivan, P., 2002. Consumer Acceptance of Genetically Modified Foods.
Food Technology Magazine, [online] Available at: <https://www.ift.org/news-and-
publications/food-technology-magazine/issues/2002/october/features/consumer-
acceptance-of-genetically-modified-foods> [Accessed 5 November 2018].
Macready, A.L., Hieke, S., Klimczuk-Kochańska, M., Szumiał, S., Vranken, L. and Grunert,
K.G., 2020. Consumer trust in the food value chain and its impact on consumer
confidence: A model for assessing consumer trust and evidence from a 5-country study
in Europe. Food Policy, [e-journal] 92, article no. 101880.
https://doi.org/10.1016/j.foodpol.2020.101880.
Munteanu, C., Mireşan, V., Răducu, C., Ihuţ, A., Uiuiu, P., Pop, D., Neacşu, A., Cenariu, M.
and Groza, I., 2021. Can Cultured Meat Be an Alternative to Farm Animal Production for
a Sustainable and Healthier Lifestyle? Frontiers in Nutrition, [e-journal] 8, article no.
749298. https://doi.org/10.3389/fnut.2021.749298.

Oliveira, I.B. and Sabato, S.F., 2004. Dissemination of the food irradiation process on
different opportunities in Brazil. Radiation Physics and Chemistry, [e-journal] 71(1-2),
pp. 495-499. https://doi.org/10.1016/j.radphyschem.2004.03.081.
Padulo, C., Carlucci, L., Balsamo, M. and Fairfield, B., 2022. A dynamic hop to cricket
consumption: factors influencing willingness to try insect-based food. Journal of Insects
as Food and Feed, [e-journal] 8(10), pp. 1157-1168. https://doi.org/10.3920/
JIFF2021.0112.
Piazza, J., Ruby, M.B., Loughnan, S., Luong, M., Kulik, J., Watkins, H.M. and Seigerman,
M., 2015. Rationalizing meat consumption. The 4Ns. Appetite, [e-journal] 91,
pp. 114-128. https://doi.org/10.1016/j.appet.2015.04.011.
Pieniak, Z., Żakowska-Biemans, S., Kostyra, E. and Raats, M., 2016. Sustainable healthy
eating behaviour of young adults: towards a novel methodological approach. BMC Public
Health, [e-journal] 16(1), article no. 577. https://doi.org/10.1186/s12889-016-3260-1.
Post, M.J., 2012. Cultured meat from stem cells: Challenges and prospects. Meat Science,
[e-journal] 92(3), pp. 297-301. https://doi.org/10.1016/j.meatsci.2012.04.008.
Purcărea, T.V., Orzan, G., Orzan, M. and Stoica, I., 2013. Romanian Consumer Behavior
Regarding Traditional Foods: Contributing to the Rebuilding of a Healthier Food Culture.
Journal of Food Products Marketing, [e-journal] 19(2), pp. 119-137. https://doi.org/
10.1080/10454446.2013.766074.
Rembischevski, P. and Caldas, E.D., 2020. Risk perception related to food. Food Science
and Technology, [e-journal] 40(4), pp. 779-785. https://doi.org/10.1590/fst.28219.
Rolland, N.C.M., Markus, C.R. and Post, M.J., 2020. The effect of information content on
acceptance of cultured meat in a tasting context. PLOS ONE, [e-journal] 15(4), article no.
e0231176. https://doi.org/10.1371/journal.pone.0231176.
Román, S., Sánchez-Siles, L.M. and Siegrist, M., 2017. The importance of food naturalness
for consumers: Results of a systematic review. Trends in Food Science & Technology,
[e-journal] 67, pp. 44-57. https://doi.org/10.1016/j.tifs.2017.06.010.
Ruby, M.B., Rozin, P. and Chan, C., 2015. Determinants of willingness to eat insects in the
USA and India. Journal of Insects as Food and Feed, [e-journal] 1(3), pp. 215-225.
https://doi.org/10.3920/JIFF2015.0029.
Saba, A. and Messina, F., 2003. Attitudes towards organic foods and risk/benefit perception
associated with pesticides. Food Quality and Preference, [e-journal] 14(8), pp. 637-645.
https://doi.org/10.1016/S0950-3293(02)00188-X.
Shepherd, R., 2008. Involving the public and stakeholders in the evaluation of food risks.
Trends in Food Science & Technology, [e-journal] 19(5), pp. 234-239.
https://doi.org/10.1016/j.tifs.2007.12.005.
Siegrist, M., 2007. Consumer attitudes to food innovation and technology. In: L. Frewer and
H. van Trijp eds., 2007. Understanding Consumers of Food Products. s.l.: Woodhead
Publishing, pp. 236-253. https://doi.org/10.1533/9781845692506.2.236.
Siegrist, M., Gutscher, H. and Earle, T.C., 2005. Perception of risk: the influence of general
trust, and general confidence. Journal of Risk Research, [e-journal] 8(2), pp. 145-156.
https://doi.org/10.1080/1366987032000105315.
Siegrist, M. and Hartmann, C., 2020. Consumer acceptance of novel food technologies.
Nature Food, [e-journal] 1(6), pp. 343-350. https://doi.org/10.1038/s43016-020-0094-x.
Vol. 26 • No. 65 • February 2024 361

Siegrist, M., Sütterlin, B. and Hartmann, C., 2018. Perceived naturalness and evoked disgust
influence acceptance of cultured meat. Meat Science, [e-journal] 139, pp. 213-219.
https://doi.org/10.1016/j.meatsci.2018.02.007.
Sinke, P., Odegard, I., van der Giesen, C., Swartz, E. and Sanctorum, H., 2023. Ex‑ ante life
cycle assessment of commercial – scale cultivated meat production in 2030. Food Chains.
[online] Delft: CE Delft. Available at: <https://cedelft.eu/publications/rapport-lca-of-
cultivated-meat-future-projections-for-different-scenarios/> [Accessed 14 March 2023].
Socaciu, M.-I., Câmpian, V., Dabija, D.-C., Fogarasi, M., Semeniuc, C.A., Podar, A.S. and
Vodnar, D.C., 2022. Assessing Consumers’ Preference and Loyalty towards Biopolymer
Films for Food Active Packaging. Coatings, [e-journal] 12(11), article no. 1770.
https://doi.org/10.3390/coatings12111770.
Stampfli, N., Siegrist, M. and Kastenholz, H., 2010. Acceptance of nanotechnology in food
and food packaging: a path model analysis. Journal of Risk Research, [e-journal] 13(3),
pp. 353-365. https://doi.org/10.1080/13669870903233303.
Stephens, N., Di Silvio, L., Dunsford, I., Ellis, M., Glencross, A. and Sexton, A., 2018.
Bringing cultured meat to market: Technical, socio-political, and regulatory challenges
in cellular agriculture. Trends in Food Science & Technology, [e-journal] 78,
pp. 155-166. https://doi.org/10.1016/j.tifs.2018.04.010.
Tenbült, P., de Vries, N.K., Dreezens, E. and Martijn, C., 2005. Perceived naturalness and
acceptance of genetically modified food. Appetite, [e-journal] 45(1), pp. 47-50.
https://doi.org/10.1016/j.appet.2005.03.004.
Verbeke, W., Marcu, A., Rutsaert, P., Gaspar, R., Seibt, B., Fletcher, D. and Barnett, J., 2015.
‘Would you eat cultured meat?’: Consumers’ reactions and attitude formation in Belgium,
Portugal and the United Kingdom. Meat Science, [e-journal] 102, pp. 49-58.
https://doi.org/10.1016/j.meatsci.2014.11.013.
Voinea, L., Popescu, D.V., Bucur, M., Negrea, T.M., Dina, R. and Enache, C., 2020a.
Reshaping the Traditional Pattern of Food Consumption in Romania through the
Integration of Sustainable Diet Principles. A Qualitative Study. Sustainability, [e-journal]
Woolf, E., Maya, C., Yoon, J., Shertukde, S., Toia, T., Zhao, J., Zhu, Y., Peter, P.C. and Liu,
C., 2021. Information and taste interventions for improving consumer acceptance of
edible insects: a pilot study. Journal of Insects as Food and Feed, [e-journal] 7(2),
pp. 129-139. https://doi.org/10.3920/JIFF2020.0057.
Wu, W., Zhang, A., van Klinken, R.D., Schrobback, P. and Muller, J.M., 2021. Consumer
Trust in Food and the Food System: A Critical Review. Foods, [e-journal] 10(10), article
no. 2490. https://doi.org/10.3390/foods10102490.
Wunderlich, S. and Gatto, K.A., 2015. Consumer Perception of Genetically Modified
Organisms and Sources of Information. Advances in Nutrition, [e-journal] 6(6),
pp. 842-851. https://doi.org/10.3945/an.115.008870.
Xiao, Y. and Watson, M., 2019. Guidance on Conducting a Systematic Literature Review.
Journal of Planning Education and Research, [e-journal] 39(1), pp. 93-112.
https://doi.org/10.1177/0739456X17723971.

ARCHITECTURE TO TRANSFORM CLASSIC ACADEMIC COURSES

INTO ADAPTIVE LEARNING FLOWS WITH ARTIFICIAL INTELLIGENCE
Andrei Bobocea1 , Răzvan Bologa2 , Lorena Bătăgan3
and Bogdan-Ștefan Posedaru4
1)2)3)4)

Bobocea, A., Bologa, R., Bătăgan, L. and Posedaru, Received: 30 September 2023
B.Ș., 2024. Architecture to Transform Classic Academic Revised: 27 November 2023
Courses into Adaptive Learning Flows with Artificial Accepted:26 December 2023
Intelligence. Amfiteatru Economic, 26(65), pp. 363-380.
Abstract
The literature on adaptive learning suggests that it can provide significant improvements to
the educational process and numerous studies have found a necessity for personalised
learning, which is one of the strong suits of adaptive learning. Adaptive learning platforms
require that content be effective, and lack thereof has hindered large-scale adoption by adding
the cost of content creation to the upfront implementation cost and creating a 'critical mass'
type problem where a platform without content is ineffective and unattractive, leading to lack
of interest from users and lack of funding for developing new content. Artificial intelligence
(AI) technology has the potential to aid in content creation by taking on a significant part of
the workload. This paper aims to explore this possibility and propose an architecture based
on current artificial intelligence technologies that will help teachers and experts transform
classic course materials into adaptive learning flows. The system is not autonomous and will
not replace a human expert but rather will take on some of the more straightforward, but time-
consuming, work. The proposed approach results in a distinct system, independent of the
adaptive learning platform itself, that can help rephrase, restructure and enrich the content,
resulting in an automated digital narrative, or fragment thereof, that can be exported in a
format based on open standards and used within an adaptive learning platform of choice.
Keywords: AI content generation, artificial intelligence, adaptive learning, learning flows,

personalised learning, educational content.
JEL Classification: A20, I20, I21
 Corresponding author, Bobocea Andrei – e-mail: [email protected]

Vol. 26 • No. 65 • February 2024 363

AE Architecture to Transform Classic Academic Courses
into Adaptive Learning Flows with Artificial Intelligence
Introduction
The report Markets and Markets (2020) shows that the adaptive learning market was valued
at USD 1.9 billion in 2020 and is expected to reach USD 5.3 billion by 2025. This growth is
driven by the demand for eLearning solutions, especially personalised learning, initiatives
government, and the perception of online education (Tartavulea et al., 2020). Adaptive
learning is thus of significant interest from an economic, technological and educational point
of view.
Adaptive learning itself aims to improve learning for every student; however, adaptive
learning platforms require a considerable amount of content to be effective and impactful. As
such, they have not been widely adopted due to the cost associated with creating appropriate
educational content.
Recent developments in educational technologies, adaptive learning and artificial intelligence
have given way to new possibilities in education and have the potential to generate significant
improvements across the board. The developments observed in the field of artificial
intelligence, especially with the recent GPT-3 Large Language Model, show promise to help
solve the classic "critical mass" problem described by Evans and Schmalensee (2010),
namely the lack of content for adaptive learning platforms.
The purpose of this article is to create an architectural proposal to transform existing
educational material into adaptive learning flows using artificial intelligence. This endeavour
involves testing current generative artificial intelligence technologies to determine whether
they can assist in the process of converting educational materials or the creation of new
educational materials based on a prompt (a textual input detailing desired outcomes,
interaction method specific to artificial intelligence). The end result of such a transformation
process is an automated digital narrative, or part of it, that can be used within an adaptive
learning platform.
The paper proposes an architecture based on current artificial intelligence technologies that
will help teachers and experts transform classic course materials into adaptive learning flows.
The proposed architecture does not depend on the use of a certain adaptive learning platform,
being independent of it.
Adaptive learning is a method of teaching that tries to address the individual needs of learners
using artificial intelligence and other computer algorithms to customise the learning
experience of each unique user (Kaplan, 2021). The need for adaptive learning stems from
the difficulty of teaching in a uniform way in a class of students with varying levels of
understanding of the subject and of the prerequisite concepts or knowledge necessary. The
need for personalised learning is well recognised (Duncan, 2013) and three kinds of adaptive
learning approaches (or levels of personalization) have been proposed: Individualisation,
Differentiation and Personalization. (U.S. Department of Education, 2010).
Individualisation refers to pace, it is about the speed with which different students reach the
same learning goals. In differentiation, the approach is customised based on the learning
preferences of the student. Finally, in Personalization, the method, as well as the learning
objectives and content, are tailored to the individual learner, in addition to being paced

according to the needs of the student. Personalization encompasses both Individualisation
and Differentiation (U.S. Department of Education, Office of Educational Technology, 2010;
Kerr, 2016).
In a classic instructional learning paradigm, the teacher is tasked to transfer knowledge to all
of his students, bringing them to a targeted minimum level of understanding (equivalent with
a passing grade in most scenarios). As such, the teacher needs to discover their current level
of understanding and account for variations between students and the mismatches between
them in other factors such as learning style and motivational structure that affect their
performance in class and the effectiveness of any one teaching strategy (Simelane and Mji,
2014).
Starting from Wright’s basis for learning curve theory (Wright, 1936) we can attribute a
similar graph to student learning and consider that they are in varying positions on the said
graph, with the graph being slightly different for each student due to their individual
differences in learning style, performance, motivation, mental state, cognitive ability and any
multitude of other factors that can make the learning curve appear steeper, requiring more
effort upfront or smoother, suggesting an accelerated learning process. Therefore, it is
reasonable to consider that a graph showing the distribution of the student’s learning
performance (as expressed in time needed to progress, or the rate of learning) would follow
a Gaussian distribution. This graph can also be interpreted as a distribution of the needs of
the students in terms of the attention needed and the appropriate challenge needed to
encourage learning, but not overwhelm the student.
The outer edges of the Gaussian distribution described earlier represent students with special
needs, one end representing students with learning deficiencies (challenged students) while
the opposite end represents gifted students, which classify as special needs due to the
increased attention and level of challenge necessary (Osborn, 1996; Kelemen, 2014) as
opposed to the usual meaning of special needs students (students with learning deficiencies)
which require a smaller challenge and a slower approach to learning.
An adaptive learning system aims to provide an appropriate challenge to each student and
adjust the pace to their needs (Essalmi et al., 2010), dynamically customising their personal
learning trajectory, building interactive digital narratives and identifying concepts that have
not been fully understood yet or competencies that can be further developed. This is an
inherently difficult task due to differences between learners and their learning styles,
difficulty which an adaptive learning system aims to alleviate (Anane, Alshammari and
Hendley, 2014). While a normal teacher, in an instructional paradigm, can only teach at one
specific difficulty level and speed at one time, aiming to synchronise to the median student
in the class from our normal distribution. This situation can become unmanageable in online
learning environments or situations where the number of students per teacher can exceed their
capacity. This situation is also known as “The teacher bandwidth problem” (Kotzee and
Palermos, 2021). An adaptive learning system can provide unique learning experiences for
each student, reducing the overall workload of a teacher and allowing more attention and time
to be dedicated where necessary.
This is achieved through connecting the student with the necessary resources and tools to
better understand and retain the material being studied and at the same time adapting the
material to their needs and abilities to keep them engaged.
Vol. 26 • No. 65 • February 2024 365

Additionally, in the above context of an instructional paradigm, there is an overhead of time

spent on tasks that are not part of the educational process themselves but must be performed
and fall upon the teacher. This is the case with tech support in classes that utilise hardware
such as any kind of lab equipment or even basic devices such as screen projectors, computer
science classes where students may encounter unknown errors, and particularly in the case of
distance learning where setup-specific problems may arise that are outside of a teacher's
control and often beyond a student’s expertise. These tasks are mandatory, but at the same
time, they reduce the overall efficiency of the teacher (Windelspecht, 2001). Such tasks block
the teaching itself until the issues are resolved; however, in the case of a class that utilises an
adaptive learning platform, the teacher is free to deal with them while the platform delivers
the educational content to all the other students. A sufficiently advanced adaptive learning
platform can diagnose and provide solutions to student issues automatically based on a series
of knowledge bases that can be queried and that can be expanded so that every new error or
problem encountered is automatically added and addressed in the future.
Adaptive learning systems also have the benefit of covering multiple learning styles. While
some students may prefer or perform best with a direct knowledge transfer approach, research
suggests that students obtain better results when they are actively engaged in discovery, such
as the self-explaining method, which has students generate an explanation using their reason,
rather than taking a fact that is presented to them for granted (Hausmann and VanLehn, 2007).
This process involves accessing prior knowledge and reasoning based on it, so that the new
knowledge is connected or linked to prior knowledge that improves learning. Knowing the
level of prior knowledge of a student is a difficult task that teachers can perform with limited
precision through various means, such as testing and direct observation; however, an adaptive
learning system can evaluate this with more precision. Knowing a student’s current level of
understanding, prior knowledge, and underlying presuppositions can greatly affect
knowledge acquisition (Shing and Brod, 2016), both explicit and implicit (Ziori and Dienes,
2008), this is also deeply tied to the student’s interests due to the linear correlation between
interests and prior knowledge (Tobias, 1994), and as such an adaptive learning system can
greatly improve the learning process by helping the teacher better understand a student’s
needs and providing the appropriate content.
An adaptive learning system requires content to be efficient and provide the most value it
can. Intuitively, if there is no content, then the adoption rate will be low, and if the adoption
rate is low, then the interest and resources needed to make new content will also be minimal,
creating a classic 'critical mass' type of problem and a race to the bottom, rather than the top.
As such, ways to produce significant amounts of decent quality content with limited resources
are an interesting proposition and potential solution to this problem.
Using AI to generate educational content, part of educational content, or just to enhance and
transform existing content is not a new idea. (Diwan et al., 2023) used artificial intelligence
to generate relevant definitions, section overviews and reflection quizzes based on existing
educational resources. The process of generating summaries involves the use of the Artificial
Intelligence component to process multiple sections of text and obtain a coherent description
of the topic addressed within them.
The generation of verification questions has the role of producing content that helps the
student retain the information presented, and the teacher (or an automatic system) to identify
gaps in knowledge early, thus improving both the educational content itself and the
effectiveness of the educational act.

The authors suggest automating the generation of educational content for training in divergent
question asking using a large language model (LLM). Through comparisons with hand-
generated information and human expert annotations, they assess the effectiveness of this
strategy. The outcomes illustrate the potential of artificial intelligence methods in assisting
children's learning and point to the relevancy and use of the LLM-generated content.
Although large language models, such as GPT-3, have shown significant promise in content
generation, their effectiveness depends on training on an enormous amount of high-quality
training data. LLMs attempt to develop answers based on patterns they have identified in
their training data (Brown et al., 2020). In the context of education, this suggests that they
may provide in-depth explanations on a wide range of subjects or even replicate problem-
solving techniques used in scientific or mathematical procedures. However, there is a risk of
overgeneralization or an excess of very broad answers that might not be accurate or nuanced
enough to address specific academic issues.
With the surging popularity and availability of transformer-based large language models,
such as GPT-3, generation of full content can be attempted. The models may still be trained
further or contextualised using existing educational resources and research to improve
accuracy with the purpose of generating valid, relevant and useful educational content that
can be verified by a human being and subsequently be improved based on feedback.
2. Methodology
In an adaptive learning context we consider a learning flow to be the equivalent of a lesson,
lecture, or laboratory session from a classic classroom, taking a constructivist approach with
emphasis on learning by doing and practical experiments that develop knowhow and a
practical understanding of the subject, rather than a purely theoretical one. A learning flow
consists of a series of steps; these steps are a simple form of instructions or information and
may contain other educational resources (videos, images, 3D animations, documents, code
sections, equations, etc.). These steps are often related to concepts which are presented within
the step, and are described as atomic because they represent the simplest and indivisible form
of the information or instructions presented within them. The purpose of a learning flow is to
provide these simple atomic steps that help the student move towards an end goal and
understand the process, whether that end goal is building and programming a robot, training
a neural network, making a simple electric circuit, producing a chemical reaction, or any other
lesson subject imaginable.
The steps of a learning flow reveal new concepts and ideas or reinforce previously learnt ones
to achieve progressive disclosure, with the aim of reducing friction in the learning process
and minimising the chances that the students become overwhelmed by the new information
(Spillers, n.d.). Each step provides not just instructions but also open educational resources
such as short videos, documents, diagrams, graphs, etc. to create an additional learning
opportunity for students who learn best by discovery and want or need more resources. The
steps may also contain an evaluation component, asking relevant questions, both to reinforce
the learnt concepts and to evaluate the degree of understanding, providing an early detection
mechanism for students who are falling behind or are disengaged.
Based on a series of engagement metrics (correct response rate, completion rate, time spent
per step, etc) the system adjusts each student’s experience by providing the appropriate
Vol. 26 • No. 65 • February 2024 367

content in terms of difficulty, additional information, or additional challenges. Flows can

have a branching tree-like structure to account for prerequisite concepts that were not well
understood previously or concepts that require further explanation.
The basic premise is that such a system will allow and encourage the student to ask questions,
request assistance and have access to a vast library of open educational resources. The
branches would eventually merge back into the main flow, leading to the same goal and
allowing each student to reach it at his own pace. Due to the interactive nature of an adaptive
learning system, the pace is performed automatically through the student’s interaction and is
based on the analytical data collected. With the appropriate pacing and challenges that are
adjusted to be rewarding without seeming insurmountable, the system will have a higher
chance to keep the students engaged for longer periods of time, thereby transferring
knowledge and creating skills.
In order to evaluate the degree of understanding of the students and identify gaps in their
knowledge, a series of question types and practical assessment exercises can be used within
the flows, the most common answer types being multiple choice, survey/scale, true/false,
numeric answer, audio/video, file or short text. A significant part of these can actually be
represented by multiple-choice or open ended questions (short text answer). Audio/video or
file responses are less common in automated systems due to the technical difficulty of
automatically grading them.
While multiple-choice questions are a common occurrence and go-to method of evaluation,
we consider that natural language-based, open-ended questions are more flexible, more
appropriate, and in line with the philosophy behind such a system. In an online only system
(which is a requirement to access the generous amount of open educational resources from
the public domain and access the functionality of the system), it is reasonable to assume that
students have access to the Internet as a whole, as such it cannot be treated as an isolated
system, and any testing should be assumed to be an open book, regardless of the type of
question or the answer format. In this scenario, in the case of multiple choice questions
students that do not know the correct answer may feel inclined to pick the closest answer, or
that which they think is most likely to be correct, while a natural language question may
encourage them to search for the right answer before answering. Additionally, ideas
expressed in natural language can be evaluated automatically with Natural Language
Processing, providing a more in-depth conclusion than a simple right/wrong answer. Current
NLP technologies can extract a range of information from a textual response, their most
common uses being for extracting entities from text, recognising terms, recognising
relationships between extracted entities, sentiment analysis and automatic summarization.
Using a small subset of the common functions of NLP technologies is sufficient to classify a
student's answers, based on a predefined training set, specific to each question, and this
classification to be validated by a human operator (a teacher for example) to compensate for
possible mistakes of the automatic process, thus retraining the classifier and improving its
performance with each intervention.
The purpose of a flow within an adaptive learning system is to present concepts and ideas in
such a way that it creates know-how and practical skills in the long term, with an emphasis
on learning by doing and on experimentation as a method of discovery.

3. Converting classic learning materials into adaptive learning flows
In its most basic form, a flow is a linear sequence of steps, as described earlier. A flow must
have a series of additional elements such as suggestions, help, or additional Open Educational
Resources that are relevant to the step they are associated with. With these elements in mind,
the flow will have a branching structure, allowing the student more freedom in the way they
interact with the content. The flow could potentially contain jump points into other flows, in
the case that it is identified that a necessary concept is not yet well understood, or to certain
materials that aim to further emphasise the current knowledge and help students understand
before moving forward. Such a structure is visible in Figure no. 1.
Figure no. 1 - Structure of a learning flow

From all of the above we understand that existing learning materials in various formats
(PDFs, E-books, presentations, etc.) from which text can be extracted automatically can be
put through a conversion process, as in Figure no. 2, which will use a Generative AI model
and Open Educational Resources to transform the content into an interactive learning flow
that can be imported directly into an adaptive learning solution.
Figure no. 2 - Conversion process diagram

The conversion process must be partially automated to reduce the overall workload of a
content creator and, as such, has a net benefit over creating content from scratch. Human
intervention is still required as the content must be validated and the AI can be directed
towards the intended level of difficulty and complexity for the content being generated.
Vol. 26 • No. 65 • February 2024 369

The content can be greatly enhanced by adding Open Educational Resources such as tutorials,
3d animations, code snippets or repositories, diagrams and other resources. Some of these
resources can be provided directly by the system by querying a database of resources such as
Wikimedia Commons or automatically by the AI model, provided that they were part of the
training data of the model.
We have identified several approaches, based on generative Artificial Intelligence models,
which we propose and which can be used independently or combined to achieve the best
results:
 Direct material transformation
 Topic extraction
-Topic based generation
-Augmented topic based generation
The performance of the AI models is largely dependent on the training data and on the prompt
that was used for generation. From our experiments we have concluded that current
generation generative AI models such as GPT-3 and GPT-4 can achieve satisfactory
performance. Techniques that fall under the category of prompt engineering can be useful in
improving results by adjusting the input rather than retraining the AI model, where possible.
The most straightforward approach is direct material transformation, and it involves using
paragraphs of existing educational materials as part of the prompt and generating a series of
steps that summarise the content and the concepts within the educational material. This will
have drawbacks if the text itself is already a plain series of factual statements, as there may
not be much to summarise or paraphrase. In the case of such material, the second approach
may be useful: Topic extraction of the text or paragraph and content generation from scratch
(described and exemplified as 'Topic-based generation' in figure no. 3) or using both the topic
and the original text as part of a prompt engineered to generate a lesson or the structure of
one, which can subsequently be further improved (described and exemplified as 'Augmented
topic-based generation' in Figure no. 3).
Due to the nature of AI, the limitations of the model, the training dataset and the prompt itself,
all results will have to go through a validation process to ensure the accuracy of the factual
information within. Once the content has been generated, validated and corrected if
necessary, it can be enhanced by inserting Open Educational Resources, either manually or
automatically, and by generating relevant verification questions. Following a final validation
step, the assets generated according to the steps above can be easily exported to any available
format, such as JSON, and subsequently imported into an adaptive learning platform of
choice.

Figure no. 3 - Detailed conversion process diagram
To validate the viability of these approaches and the technology behind them, several texts
with educational value were put through this process. One of these experiments is described
below, the original input text has been replaced by the [text] token, and the results have been
obtained using generative artificial intelligence tools based on an LLM (OpenAI, n.d.). In this
example, a text discussing sorting algorithms and their implementation in JavaScript engines
was processed. The method used was Topic based Augmented Generation (table no. 1).
The prompt used to extract the topic was of the form: “Extract the topic from the following
Vol. 26 • No. 65 • February 2024 371

text: [text]”, and the response was: “The topic of the text is 'Sorting algorithms and their
implementation in JavaScript engines”.
Table no. 1. Sorting Algorithms and their Implementation
in JavaScript Engines – Augmented Topic Based Generation
Task Process a text discussing sorting algorithms and their implementation in
JavaScript engines.
Method Augmented topic based generation
Prompt Generate the structure of a 5-step lesson on “Sorting Algorithms and Their
Implementation in JavaScript Engines” based on the following text: [text]
Result Step 1: Introduction to sorting concepts
Objective: Introducing the fundamental concept of sorting.
Key points:
Sorting involves arranging data in a specific order, such as numerically,
alphabetically, or based on specific criteria (for example, sorting people by
height or age).
In programming, sorting often deals with arrays of numbers or strings.
Didactic approach:
Discuss real-world examples of sorting (e.g., sorting numbers, names, or
people by attributes) to contextualise the concept.
Use code snippets or visual aids to illustrate the sorting process.
Step 2: Importance of Sorting in JavaScript
Objective: To understand the importance of sorting algorithms in JavaScript
programming.
Key points:
Sorting is a fundamental operation in programming that optimises the
arrangement of data for efficient manipulation.
In interviews or coding tasks, sorting algorithms are typically tested to sort
arrays of numbers or strings.
Didactic approach:
Engage students in scenarios where sorting algorithms are essential in real-
world programming tasks.
Provide sample interview questions or coding challenges related to sorting
algorithms.
Step 3: Overview of JavaScript Engines and Array.sort()
Objective: Introduce JavaScript engines and the built-in Array.sort()
method.
Key points:
JavaScript engines execute JavaScript code in browsers (e.g., Chrome V8,
Firefox SpiderMonkey).
Array.sort() is a built-in method in JavaScript to sort arrays.
Didactic approach:
Explain how different browsers use different JavaScript engines.
Demonstrate the basic use of the Array.sort() method with simple code
examples.

Step 4: Comparison of Sorting Algorithms in JavaScript Engines
Objective: Explore the differences in the sorting algorithms used by different
JavaScript engines.
Key points:
Different JavaScript engines use different sorting algorithms for
Array.sort().
For example, the Chrome V8 engine uses Timsort, while Mozilla Firefox
uses sort sort.
Didactic approach:
Give an overview of common sorting algorithms and their efficiency.
Illustrate how these algorithms are implemented in various JavaScript
engines.
Step 5: Practical applications and optimization strategies
Objective: Discuss practical applications and optimisation techniques for
sorting in JavaScript.
Key points:
Optimise the classification algorithms based on factors such as data size,
engine differences, and time complexity.
Discuss strategies to select appropriate sorting algorithms for different
scenarios.
Didactic approach:
Engage students in exercises or coding challenges to implement and
optimise sorting algorithms in JavaScript.
Highlight the importance of selecting the right algorithm for improved
performance in web development.
This structured lesson plan covers fundamental concepts, the specifics of the
JavaScript engine, and practical considerations related to JavaScript sorting
algorithms.
Source: OpenAI, n.d.
The other methods described in (Figure no. 3) obtained similar results, with small structural
differences. Other notable mentions include using the Topic Based Generation method
without starting from an existing text, or using the processing result as input to another
processing step. The lesson plan generated by Augmented topic-based generation within
the example in (table no. 1) can be used as input data in an Augmented topic-based
generation step. The description of the individual steps in the lesson can be used to generate
lesson content based on the generated plan,
The content of the prompt is the main determining factor of the format of the response
generated by the generative model. Increasing the degree of specificity, including a context,
and describing in as much detail as possible the form of the desired end result (structure of a
lesson, segment of text, etc.) increases the chances that the output generated will have a
predictable format. Due to the nature of AI, the limitations of the model, the training dataset,
and the prompt itself, all results will need to go through a validation process to ensure the
accuracy of the information within. Once the content has been generated, validated and
corrected if necessary, it can be enhanced by inserting Open Educational Resources, either
manually or automatically, and by generating relevant verification questions (Table no. 2):
Vol. 26 • No. 65 • February 2024 373

Table no. 2. Generation of verification questions

Prompt Response
Generate an open-ended Question

question and a short answer How do JavaScript engines improve the efficiency of
based on the following text: sorting algorithms and what factors do they consider?
JavaScript engines
implement efficiency- Answer
optimised sorting algorithms, JavaScript engines improve the efficiency of sorting
taking into account factors algorithms by customising them to be more efficient in
such as time complexity and handling data. They consider factors such as time
data characteristics. These complexity and data characteristics. Time complexity
engines optimise sorting refers to the efficiency of the algorithm in terms of the time
methods to efficiently handle required for execution relative to the size of the input. Data
different data sizes and characteristics include the type, structure, and
minimise resource arrangement of the sorted data. By optimising sorting
consumption. methods based on these factors, JavaScript engines aim to
efficiently handle different data sizes while reducing
resource usage.
Source: OpenAI, n.d.

After a final validation step, the elements generated according to the steps above can be easily
exported to any available format, such as JSON, and subsequently imported into an adaptive
learning platform of choice. The validation step and potential manual correction can take up
to two hours per hour of content produced. This step involves reviewing the content, the
duration of the review being influenced by the validator's expert quality, but also by the
increased attention required to observe possible errors. From our experiments, exemplified
above, this two-hour time covers most cases.
This time is significantly less than the time required to manually develop this educational
content or other types of content. Manual development time for educational content for
adaptive learning platforms is estimated at more than 50 hours per hour of content and 25
hours per hour of content using specialised software tools (Khosravi, 2019).
For classic instructional content of the “Instructor-led training” type and eLearning content
of various levels of complexity, the time spent on average is significantly higher (Figure no.
4), using the definitions for the level of complexity of the e-Learning content presented by
Chapman (2010).

Figure no. 4. Time spent per hour of content

Source: Chapman, 2010
The types of content presented do not involve the same types of development activities, and
thus, in order to be able to have a comparison term, only those activities were extracted that
are comparable in purpose to the transformation of existing materials using Artificial
Intelligence (Table no. 5).
Table no. 5 - Time spent per hour of content (comparable activities)
Time spent per hour of content (hours)
Activity ILT eLearning-L1 eLearning-L2 eLearning-L3
Instructional design 6.84 10.88 24.69 61.97
Development of the lesson 5.06 0 0 0

plan
Creation of handouts 3.38 0 0 0
Student Guide/Workbook 4.83 0 0 0

development
Test and exam creation 3.42 0 0 0
Storyboarding 0 9.03 20.88 53.22
Total 23.53 19.91 45.57 115.19

Source: Chapman, 2010
Comparing these data (Figure no. 5) we notice that by introducing an automated process
based on Artificial Intelligence we can significantly reduce the time needed to develop
educational material and thus the related cost.
Vol. 26 • No. 65 • February 2024 375

Figure no. 5 - Time spent per hour of content (comparable activities)

Other technologies based on Artificial Intelligence can be introduced to bring additional
efficiency improvements to the other activities specific to the development of educational
content, such as the generation of video content (Text-to-video) or images (Text-to-image).
Additionally, open educational resources, in the public domain, produced by other entities
(such as OER Commons or UNESCO) can be introduced for reuse in educational content.
4. Comparative analysis of the proposed approach and other works

Previous works have proposed using AI mainly as a conversation partner, in the form of a
chatbot, that helps the student learn, or to provide lightweight content creation tasks such as
section summarization or generation of multiple-choice verification questions.
Using AI as a chatbot has been proposed, tested, and even implemented with promising
results (Bii, 2023). The implementation of a voice chatbot called 'Ellie' in Korean EFL classes
has had a high success rate and improved classroom participation (Yang et al., 2022). The
feedback based on talking with a chatbot during class is presented in Figure no. 6.
This approach presents a couple of challenges and limitations: If the underlying AI model is
not a generative AI model, then it must be trained explicitly for all the tasks it is expected to
perform, for all the questions it is supposed to answer and all the explanations it is supposed
to provide. This involves a significant amount of effort and time, possibly that of a data
scientist. If the model is generative, such as the current generation GPT-3 and GPT-4, then it
is susceptible to 'hallucinations', a phenomenon in which the AI model generates an answer
that seems plausible, but on closer inspection it is invalid, as it is factually incorrect or not
related to the given prompt (Daniel, 2023; Marr, 2023). This phenomenon may be avoided
by using a highly specialised model or with significant efforts on the part of prompt
engineering by providing more contextual information, but it remains a concern when used
for education.

Figure no. 6 - Improved classroom engagement

Source: Yang et al., 2022
The second approach we have mentioned, proposed by (Diwan et al., 2023), is closer to our
proposal. It involves using generative AI models (in particular, GPT-2) to enhance existing
content, with the goal of increasing learner engagement. This is achieved by providing
relevant section summaries, generated upon the learning materials that the learner is about to
encounter, and by generating relevant verification questions based upon those same materials.
Our proposal is the use of generative AI models such as current-generation GPT-3 and GPT-
4 to transform existing learning materials into adaptive learning flows, thereby lowering the
implicit adoption cost of creating educational materials for a new learning platform. This
involves a conversion process in which much of the structure of the existing learning
materials is significantly altered, but the underlying information and concepts are retained.
The AI component is used in this process to do the conversion itself, fill the gaps where
necessary, generate new content where existing materials are insufficient, and enrich the
learning flow with elements such as Open Educational Resources and relevant verification
questions. This architecture does not aim to replace, but rather to aid creators of educational
content, to ease their workload and improve productivity.
Conclusions
Current generative AI models are able to facilitate the transformation of classic academic
course materials into learning flows, lowering adoption costs for adaptive learning systems,
thus confirming our research hypothesis. The comparative analysis shows that this ability
exists largely due to the significant advances in the field of generative Artificial Intelligence
models and further developments are possible as the models themselves are improved,
specialised, or fine-tuned for specific tasks including educational content generation.
Future research may involve collecting and analysing data on the adoption, user experience,
efficiency, and cost benefits of integrating generative Artificial Intelligence models into a
content generation workflow and of using the resulting content in a broader academic context.
Vol. 26 • No. 65 • February 2024 377

References
Anane, R., Alshammari, M.T. and Hendley, R.J., 2014. Adaptivity in E-Learning Systems.
[online] Available at: <https://www.researchgate.net/publication/263569590_
Adaptivity_in_E-Learning_Systems> [Accessed 12 September 2023].
Bii, P., 2023. Chatbot technology: A possible means of unlocking student potential to learn
how to learn. Educational Research, 4, pp. 218-221. [online] Available at:
<https://www.researchgate.net/profile/Bii-Kiptonui/publication/291813780_Chatbot_
technology_A_possible_means_of_unlocking_student_potential_to_learn_how_to_learn
/links/56a61bd208ae6c437c1ad941/Chatbot-technology-A-possible-means-of-
unlocking-student-potenti> [Acessed 2 December 2023].
Brown, T.B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A.,
Shyam, P., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan,
T., Child, R., Ramesh, A., Ziegler, D.M., Wu, J., Winter, C., Hesse, C., Chen, M., Sigler,.
E., Litwin, M., Gray, S., Chess, B., Clark. J., Berner, C., McCandlish, S., Radford, A.,
Sutskever, I. and Amodei, D., 2020. Language Models are Few-Shot Learners. [online]
Available at: <https://arxiv.org/abs/2005.14165> [Accessed 2 December 2023].
Chapman, B., 2010. How Long Does it Take to Create Learning? [pdf] Available at:
<https://www.cedma-europe.org/newsletter%20articles/misc/How%20long%20does%
20it%20take%20to%20develop%20training%20by%20Brian%20Chapman%20(Sep%
2010).pdf> [Accessed 26 November 2023].
Daniel, W., 2023. 'Hallucination problems' plague A.I. chatbots, Google CEO says. [online]
Fortune. Available at: <https://fortune.com/2023/04/17/google-ceo-sundar-pichai-
artificial-intelligence-bard-hallucinations-unsolved/> [Accessed 29 April 2023].
Diwan, C., Srinivasa, S., Suri, G., Agarwal, S. and Ram, P., 2023. AI-based learning content
generation and learning pathway augmentation to increase learner engagement.
Computers and Education: Artificial Intelligence, 4. <https://doi.org/10.1016/
j.caeai.2022.100110>.
Duncan, A., 2013. Enabling the Future of Learning. Obama White House. [online] Available
at: <https://obamawhitehouse.archives.gov/blog/2013/12/17/enabling-future-learning>
[Accessed 23 March 2023].
Essalmi, F., Ayed, L.J.B., Jemni, M., Kinshuk and Graf, S., 2010. A fully personalization
strategy of E-learning scenarios. Computers in Human Behavior, 26(4), pp. 581-591.
https://doi.org/10.1016/j.chb.2009.12.010.
Evans, D.S. and Schmalensee, R., 2010. Failure to Launch: Critical Mass in Platform Businesses.
Review of Network Economics, 9(4). https://doi.org/10.2202/1446-9022.1256.
Hausmann, R.G.M., and VanLehn, K., 2007. Self-explaining in the Classroom: Learning
Curve Evidence. Proceedings of the Annual Meeting of the Cognitive Science Society, 29.
[online] Available at: <https://escholarship.org/uc/item/7767m2c8> [Accessed 2
December 2023].
Kaplan, A., 2021. Higher education at the crossroads of disruption: the university of the 21st
century. First edition ed. Great debates in higher education. Bingley, UK: Emerald
Publishing.
Kelemen, G., 2014. Gifted children, special needs and special education. [online] Available
at: <https://www.researchgate.net/publication/273699596_GIFTED_CHILDREN_
SPECIAL_NEEDS_AND_SPECIAL_EDUCATION> [Accessed 15 March 2023].

Kerr, P., 2016. Adaptive learning: ELT Journal, 70(1), pp. 88-93. https://doi.org/
10.1093/elt/ccv055.
Khosravi, H., 2019 Development of Cost-Effective Adaptive Educational Systems via
Crowdsourcing. EDUCAUSE Review. [online] Available at: <https://er.educause.edu/
blogs/2019/12/development-of-cost-effective-adaptive-educational-systems-via-
crowdsourcing> [Accessed 26 November 2023].
Kotzee, B. and Palermos, S.O., 2021. The Teacher Bandwidth Problem: MOOCs,
Connectivism, and Collaborative Knowledge. Educational Theory, 71(4), pp. 497-518.
https://doi.org/10.1111/edth.12495.
Markets and Markets, 2020. Adaptive Learning Market Growth Drivers & Opportunities.
Markets and Markets. [online] Available at: <https://www.marketsandmarkets.
com/Market-Reports/adaptive-learning-market-257528889.html> [Accessed 30
November 2023].
Marr, B., 2023. ChatGPT: What Are Hallucinations And Why Are They A Problem For AI
Systems. [online] Available at: <https://bernardmarr.com/chatgpt-what-are-
hallucinations-and-why-are-they-a-problem-for-ai-systems/> [Accessed 29 April 2023]
OpenAI, n.d. OpenAI. [online] Available at: <https://openai.com/> [Accessed 30 November
2023].
Osborn, J., 1996. Special Educational Needs of Gifted and Talented Children. [online]
Available at: <https://www.researchgate.net/publication/234577187_Special_
Educational_Needs_of_Gifted_and_Talented_Children> [Accessed 15 March 2023].
Shing, Y.L. and Brod, G., 2016. Effects of Prior Knowledge on Memory: Implications for
Education. Mind, Brain, and Education, 10(3), pp. 153-161. https://doi.org/10.1111/
mbe.12110.
Simelane, S. and Mji, A., 2014. Impact of Technology-engagement Teaching Strategy with
the Aid of Clickers on Student’s Learning Style. Procedia - Social and Behavioral
Sciences, [online] 136, pp. 511-521. https://doi.org/10.1016/j.sbspro.2014.05.367.
Spillers, F., n.d.. Progressive Disclosure | The Glossary of Human Computer Interaction.
[online] Interaction Design Foundation. Available at: <https://www.interaction-
design.org/literature/book/the-glossary-of-human-computer-interaction/progressive-
disclosure> [Accessed 19 March 2023].
Tartavulea, C.V., Albu, C.N., Albu, N., Dieaconescu, R.I. and Petre, S., 2020. Online
Teaching Practices and the Effectiveness of the Educational Process in the Wake of the
COVID-19 Pandemic. Amfiteatru Economic, 22(55), pp. 920-936. https://doi.org/
10.24818/EA/2020/55/920.
Tobias, S., 1994. Interest, Prior Knowledge, and Learning. Review of Educational Research,
64(1), pp. 37-54. https://doi.org/10.3102/00346543064001037.
U.S. Department of Education, Office of Educational Technology, 2010. Transforming
American Education: Learning Powered by Technology [online]. Department of
Education. Available at: <https://www.ed.gov/sites/default/files/netp2010.pdf>
[Accessed 2 December 2023].
Windelspecht, M., 2001. Technology in the Freshman Biology Classroom: Breaking the Dual
Learning Curve. The American Biology Teacher, 63(2), pp. 96-101.
https://doi.org/10.2307/4451050.
Vol. 26 • No. 65 • February 2024 379

Wright, T.P., 1936. Factors Affecting the Cost of Airplanes. Journal of the aeronautical
sciences, 3, pp.122-128. [pdf] Available at: <https://pdodds.w3.uvm.edu/research/papers/
others/1936/wright1936a.pdf> [Accessed 2 December 2023].
Yang, H., Kim, H., Lee, J.H. and Shin, D., 2022. Implementation of an AI chatbot as an
English conversation partner in EFL speaking classes. ReCALL, 34(3), pp. 327-343.
https://doi.org/10.1017/S0958344022000039.
Ziori, E. and Dienes, Z., 2008. How does Prior Knowledge Affect Implicit and Explicit
Concept Learning? Quarterly Journal of Experimental Psychology, 61(4), pp. 601-624.
https://doi.org/10.1080/17470210701255374.

Victor Tufescu or the Perfect Elegance

of Man and Spirit
Silviu Neguţ
University of Economic Studies, Bucharest, Romania
Please quote this article as follows:

Neguţ, S., 2023. Victor Tufescu or the Perfect Elegance of Man and Spirit. Amfiteatru
Economic, 26(65), pp. 381-391.
In the 1950s and 1960s, the Department of Economic

Geography at the Bucharest University of Economic
Studies (BUES, in Romanian known as “ASE”) was part of
the Faculty of Commerce (now the Faculty of Business and
Tourism). Also, after 2011, the geography collective
became part of the newly established Department of
Tourism and Geography. Given the esteem in which
geography was regarded in ASE, for a long time, until
2013, it was an admission discipline, either compulsory or
optional. For example, in 1963, there were even two exams
– written and oral – for the Geography of the People's
Republic of Romania. Between 1941 and 1968, the most
prominent personality of geography taught at ASE was
Professor Victor Tufescu. Between 1951 and 1956 he was
abusively removed from university teaching and at one
point arrested. Before his university career, he taught
geography at the secondary school level. Since 1958, he
has been the author of Geography of the RPR / RSR /
Romania textbooks for the final grades of general school
and high school, respectively, most often in collaboration with Claudiu Giurcăneanu and
Gheorghe Ghica. For at least 40 years, students have learnt the geography of the country
from successive editions of these books. The observation is that the attribution of the
textbooks would have been done directly, possibly as moral reparation for the three
authors, colleagues at the former Commercial Academy, who had also previously suffered
obvious professional restrictions. Professor Victor Tufescu has been awarded the
Romanian Academy Prizes twice, under different political regimes: for A region of living
circulation: Poarta Târgului-Frumos (published in 1940), the “George Vâlsan” Prize, and
for Natural relief modelling and accelerated erosion (published in 1966), the “Gheorghe
Munteanu-Murgoci” Prize. In 1992, after the new change of political regime, he will be
elected as academician. The Institute of Geography and the Faculty of Geography of the
University of Bucharest, through V. Urucu and, respectively, by M. Ielenicz, M. Grigore,
Vol. 26 • No. 65• February 2024 381

AE Victor Tufescu or the Perfect Elegance of Man and Spirit
D. Bălteanu, F. Grecu and I. Nicolae, will dedicate to Professor Victor Tufescu two tribute
volumes, in 2000 and 2008. (Nicolae Lupu)
Victor Tufescu was an outstanding personality of Romanian geography. He established

himself with brilliance as a university professor in two prestigious institutions of higher
education: the Bucharest University of Economic Studies (formerly the Academy of Higher
Commercial and Industrial Studies, then the Institute of Economic Sciences “V.I. Lenin”)
and the University of Bucharest, as well as an excellent researcher, mainly at the Institute of
Geography of the Romanian Academy. He was honoured with the title of Academician.
Professor Victor Tufescu was born on 19 November 1908, in a family of intellectuals, in
the city of Botoșani, where he attended secondary and high school. As for his university
studies, he attended the Faculty of Sciences of the “Alexandru Ioan Cuza” University of
Iași, and after a first attempt to become a physician, as his mother wanted, he became a
graduate in Geography and Law, but quickly turned to the first field, to which he was
naturally inclined. He ascended to heaven on 11 March 2000.
I did not have the opportunity to be a student of Professor Victor Tufescu, as I graduated in
the year (1968) in which he was transferred from the Bucharest University of Economic
Studies to the Faculty of Geography of the University of Bucharest. But once I became a
professor at ASE, where the great geographer officiated for almost three decades, I had the
opportunity to learn many beautiful things about him, thanks to meetings and discussions
with many teachers or personalities from the financial-banking, commercial, business, etc.,
I was therefore able to realise how appreciated and loved Professor Victor Tufescu was –
eloquent reports about his special way of giving lectures, for the culture he showed, for his
physical and moral integrity, for his extremely deferential way of addressing, for everything
that made him a normal but truly special man. Furthermore, of extraordinary politeness, not
even the youngest, even students, were addressed as “tu” (“you”), but as “dumneavoastră”
(“your”) or, more frequently, as “mata”. If not as a student, I later got to know the professor
very well, first as an editor, then thanks to other circumstances, having the chance and, at
the same time, the honour of being one of the people very close to His Excellency, in the
last part of his life, and thanks, perhaps, to “geographical proximity”, living very close to
him, on Nicolae Titulescu Street.
Working with the professor. Once I became a redactor at the Romanian Encyclopaedic
Publishing House (Editura Enciclopedică Română), I met him in the geography editorial
office, where he came to bring the articles that had been commissioned for the Small
Encyclopaedic Dictionary, in fact a large work, despite the title, and for what would
eventually be called the Encyclopaedic Dictionary (7 volumes).
I was simply stunned when I saw his incredibly beautiful, artistic writing. I liked the
professor very much: he was always elegant (suit and tie obligatory, even in the warm
season), but not ostentatious, he spoke extraordinarily beautifully, with an even and clear
pronunciation, he was – and it didn't just seem to me – extremely deferential to my fellow
geographers, about the same age as me, addressing themselves, as I said before, as
“dumneavoastră” (“your”) or “mata”.
Then I proposed him as a contributor to the monograph Romania, requested by the
prestigious West German publishing house Schauble – we were in the “age of openness”,

but unfortunately it did not last long. Professor Victor Tufescu was very pleased with the
proposal; he could hardly believe that it was possible to publish a book written in
communist Romania for West Germany. He had a… German punctuality: the first to bring
the article, impeccable in form and content. During our collaboration, he invited me to his
home several times and I still remember the address: Jean Texier no. 19, in Floreasca
neighbourhood, opposite the skating rink. It was an old villa, beautiful and elegant, he lived
on the ground floor. Later it struck me that he moved “to the block”, buying an apartment
on Nicolae Titulescu Street, probably living in that villa by renting it.
Prof. Victor Tufescu with a group of young geographers (1972), from left to right:
Silviu Neguț, Ion Nicolae (current university professors), Gabriel Matei and
Constantin Romanciuc (editors at Encyclopaedic Publishing House) and Nicolae Pavel
(emigrated to the USA)
The book about Romania, for Germany, although it was ready, even translated into
German, was never published, since after the Ceaușescu visit to China and North Korea
everything took a different turn, especially since Elena Ceaușescu, inspired by the model of
Mao's wife, who, among other things, led the “cultural revolution”, came to the forefront of
political leadership.
But the collaboration with the professor did not stop there. At his request, I first accepted to
be his redactor for a book, even though it was not part of my duties, namely Agriculture at
the Edge of Dryness, an occasion that allowed me to reinforce my previous convictions
about his qualities as a man and as an author. But the moment that simply connected our
relationship forever was an extremely important, relevant publishing project: the capital
work of the greatest Romanian geographer, Simion Mehedinți, also one of the most
representative in the world – Terra. Introduction to Geography as a Science – which was
republished. A Sisyphean work! More than half a century had passed since the publication
of the work (in 1930), a period in which the world had experienced radical political,
Vol. 26 • No. 65• February 2024 383

scientific, and other events and changes. Professor Tufescu has produced an exceptional
critical edition, with a royal respect for the scientific contribution of his great predecessor
and with an incredible critical apparatus, designed not to “destroy” the work in question, as
we have seen in some cases, but to explain it in the perspective of such alert modern times.
In fact, he has succeeded in further accentuating the merits of the great Mehedinți and not
in diminishing them, as some of his fellow countrymen have done.
The storyteller Victor Tufescu. During the meetings, at first more sporadic and then
regular, I liked to challenge Professor Tufescu to tell stories, as he was charming. This
happened especially in his apartment on Titulescu Street, where I used to go with Dan
Bălteanu, but not infrequently, neither alone. When we arrived, either with Dan or alone,
Mrs. Liliana would offer us/me a sweet and a glass of cold water and sometimes she would
ask us if we/I would like a tea or a coffee. On those occasions, I learnt, earlier than many
others, some of the stories he told in his two works: The Tufescus (Neamul Tufeștilor) and
Family Portraits (Portrete de familie). In fact, when the first one was published in 1999, I
played a certain role: together with Professor Florina Bran, I asked her husband, none other
than Paul Bran, the rector of the ASE, a wonderful man, to have it photocopied in the
institution's printing house. It was an ideal solution, giving the work a special charm, thanks
to the professor's exceptional writing. His Excellency was also delighted with the chosen
solution. The previous year we had organised a celebration for his 90th birthday at the ASE
rector's office. It was a particularly cold day, so when the professor appeared, we were all
speechless: he was wearing a light overcoat and his familiar small brimmed hat on his head,
giving us another sportsmanship lesson, if there was one. I
go back to what the professor said. He had, among other
things, an extraordinary gift for storytelling, whether it
concerned events from his childhood and adolescence,
especially, or from later. The modulations of his voice were
accompanied by facial expressions, depending on the
subject, funny or, on the contrary, sad or sometimes even
tragic. He did not victimise himself and he rarely named
directly a person who had hurt him. The teacher also had
an “elephant's memory”, as they say: even at 90 years of
age he would recount details, details, descriptions of the
atmosphere, descriptions of people (their physical and
intellectual characteristics, qualities and faults, etc.), and in
the end, in almost all cases, he would have a comforting
word or a meaningful smile.
I remember now, as if I had heard him yesterday, that Dan
Homage by acad. Victor Balteanu, the closest and most appreciated man by the
Tufescu to the founder professor in the last part of his life, was also present – the
of Romanian university story of Patriarch Theoctist, a poor boy from a large family
geography, Simion on the neighbouring estate of Tufescu family who, as a
Mehedinți, also small child, “ran away” to the monastery, from where he
a complement to his was “recovered” and brought home, being subjected to the
fundamental work Terra. necessary “corrections”, the teacher intervening with the
Introduction to friendly neighbour with the argument that perhaps this was
Geography as a Science the child's inclination and destiny, which was perfectly true
(1930) in time! Not to mention the stories about his family,

whether parents, brothers, sisters, other relatives, or about his two children, Astrid and
Mircea, or about his grandchildren...
I can also tell what he did not have. He did not envy those who, in order to prepare their
own pedestal for a future statue, dismantled their ancestors or contemporaries. A telling
example was the reedition of the fundamental Romanian geographical work: Terra.
Introduction to Geography as a Science by Simion Mehedinți (1994). More than five
decades had passed since its publication, and the publisher proposed to re-edit it, but there
were discussions, on the one hand, whether this was appropriate and, on the other hand,
who should be the “editor”. The choice of Professor Victor Tufescu was indeed a happy
one, the edition he edited being a true model of reedition over time, not diminishing from
its value and, moreover, through the footnotes, constantly reinforcing the extraordinary
thought and modernity of the great forward. In addition, he has added a significant volume:
Simion Mehedinți. Life and work as an author.
Music lover. I remember even now the professor's stories about his passion for music, and
this was before he had written his memoirs. An extraordinary continuous flight into the
field, from the piper's box (with the parrot who, at a signal, “would draw out a note with a
curved beak, which he offered to anyone who deposited a small coin”) to the military band
(“which enchanted me”) and finally to an orchestra (“which seemed to me to be something
else altogether”). This is how he relates, in retrospect, that first orchestra audition, at the
age of only three and a half, when he accompanied his mother to a celebration at the high
school where his father was principal: “The impression made on me by those vocal and
instrumental songs was divine; I seemed to rise into a world of angels and floated lightly in
the air like dandelion flakes. I was unable to express with the poor vocabulary of my age
what I was feeling, but my mother told me that I was pulling her hand and I was ready to
cry. I never forgot for a long time that wonderful harmony of sounds, which transfigured
me.” But the professor was not only a keen music lover; he took lessons and played the
violin himself, and was quite talented, according to the testimony of some contemporaries.
He kept that violin with sanctity, which I saw at his house in Jean Texier and at his
apartment on Titulescu Street.
An incredible life and destiny. The great portraitist. Professor Victor Tufescu had a
more than an interesting life and destiny, which he did not take notice of, only at an old age
giving us some elements, and not in public. What a pity! Who would have guessed that this
rather small, well connected, polite and deferential gentleman, who was equal to himself,
was descended from one of the oldest families of Moldavian boyars (he called them
“boiernași”), that one of his great-grandmothers was the second wife of the ruler Constantin
Cantemir, father of the famous scholar Dimitrie Cantemir, that some of his ancestors held
important dignities and other positions, especially professors (including university
professors), MPs, generals, etc.
His original name, on his father's line, was Dolhici, from his great-grandfather Vlad
Dolhici, from whom the name of the village where on May 1, 1406 it was mentioned that
he had his residence, namely the village of Dolhești. At the end of the 16th century and the
beginning of the 17th century, his descendants split into three branches, one, the basic one,
remaining in Dolhești, being mentioned in documents from the end of the 17th century
under the name of Tufescu, bearing the mention “great-grandchildren of old Vlad Dolhici”,
the first, mentioned in documents, being the starost Mihu, whose daughter became the
second wife of the mentioned Constantin Cantemir.
Vol. 26 • No. 65• February 2024 385

On his mother's side, he was also descended from an old family who came from Bucovina
after the region came under Austrian rule and founded the village of Dumbrăveni, whose
name comes from the fact that the first houses were built in an oak forest grove. That
ancestor was called Matei Huzdup, and his grandson, Gheorghe Filip (Filip Huzdup's son),
would become the professor's great grandfather, with the nickname of Old Dad (“Tata
Bătrân”). The latter's son Niculai, a priest, was interested in social issues and laid the
foundations of cooperation in Romania, together with the great economist Ion Răducanu,
future rector of the Commercial Academy (today ASE).
The professor was, among many other things, an extraordinarily delicate portraitist. Here,
as a first example, is the image of his great-grandfather:
“During the holidays spent in Dumbrăveni, I met many relatives of my mother's family.
Among the most representative was my grandfather's father, whom everyone called Old
Dad. His name was Gheorghe Filip and he was a man of small height, very lively even
when he was old, with a handsome face, ruddy in the cheeks, with a moustache à la Stephen
the Great. He was of cheerful nature, a good storyteller, with a humorous speech”. Good
storyteller! This is who the professor inherited, who was also, as I said, an excellent
storyteller. Returning, let us see a new sample of the professor's gift for observation and
storytelling: “What also impressed me as a child was the ceremonial that took place when
Old Dad came to see my grandfather («Bunelu»), his son. My grandfather would kiss his
hand because he was his father and, in turn, Old Dad would kiss his son's hand because he
was a priest. This sign of mutual respect, this worship of ancient customs, was very moving
to me”. But let us also see the portrait of his mother, whom he particularly appreciated and
loved, simply adored. I understood him perfectly because I too had such feelings for my
mother: “The first image that emerged, bright in the blackness of memory and lingering in
outlines I have never forgotten, was of my mother leaning over my little bed with red string
netting on the sides and pictures of angels and beautifully coloured toys on the boards at
either end. Some nights, I could hardly be persuaded to leave my toys on the carpet to be
placed in the crib, but very quickly my eyelids would close and I would fall asleep
dreaming that I was still playing, but with special toys, sometimes fantastic, as far as the
fantasy of the limited horizon of images of my early childhood (...) My mother used to
move my crib into the middle of the room so she could keep a closer eye on my sleep, with
its childlike anxieties. Her gentle figure bent over my blinking and her caresses passed
unbidden in my dream. I was convinced at the time that she was the most beautiful mother
in the world, the gentlest and sweetest.”
And another delicate and moving scene. The professor tells us that when he was about three
and a half years old, he accompanied his father to the old house (it was a hundred years old
at the time!) and its garden, which he intended to renovate. The child broke away from his
father and, after passing through a small gate in the neglected garden, walked “along a path
overgrown with knotweed”, seeing first, to his great surprise, “a lone red tulip” and then...
But let us leave it to the wonderful storyteller: “I stood still for a few moments, as if
bewitched, for from under a bush a spiky hedgehog appeared, moving with small steps. I
had no idea what this could be: only from a book with colourful drawings did I understand
that it could be a hedgehog... I was discovering nature with my own senses. I was overcome
with excitement and stunned. After a while, which seemed a long time to me, the hedgehog
unfurled again and took off quickly, losing itself in the tall grass.” “Small steps”, “quick”...
how wonderful!

When, after a long search, his father found him and said “I have been looking everywhere
for you. What have you been doing in these weeds?”, the child's response was astonishing:
“I was in a story!”
The way to geography. Although the professor has been attracted to nature since he was a
child, choosing geography was not easy or natural. His mother, whom he loved and
respected enormously, wanted him to become a physician. And he took the exam at the
Faculty of Medicine (one of the few faculties where there was such a test), succeeded, but
very quickly found that he could not accommodate himself to the field (death, smells, etc.)
and had to drop out. He applied to “Geography and Natural Sciences”, to which he already
felt a certain attraction (I will come back to this), but at the suggestion of an uncle (Jean, his
father's elder brother), he also studied law (which was to offer him a sure chance of earning
an income, as the aforementioned relative had been pestering him), graduating with an
excellent degree in both (in 1930). But he chose the path of geography and he did well,
because he became one of the greatest and most representative Romanian geographers.
Some of his professors, including Gheorghe Năstase and Ion Simionescu, the latter a great
scientific personality (he was also president of the Romanian Academy) and, above all, one
of the most talented popularisers of nature, contributed to the consolidation of his decision.
He took his doctorate at the age of only 28, in 1936, with his thesis on Dealul Mare –
Hârlău. Observations on the evolution of relief and human settlements (published in the
following year). But in fact, everything had started at the age of 14, when he took part in a
school trip to the Bucovina Mountains.
This is what Victor Tufescu as a pupil wrote in his notebook: “In the morning, in the bright
sunlight, when I saw the mountains for the first time, it seemed to me like something
wonderful, like a fairy-tale dream. The mountains looked like curtains of darkly coloured
fir trees, closing off the horizon in close proximity. The rivers and even the small streams
flowed with their clear waters, leaping over the boulders; the towering spruces seemed to
me to come from the legend of the Nibelungs, they were different from our deciduous
forests pierced by the arrows of the sun's rays... And I began to write and draw in my
pupil's notebook, forgetting the pupils who threw stones into the water or played with the
ball”.
Professor at the Bucharest University of Economic Studies (ASE). Among the many
geographers who have taught, over the years, in this prestigious institution of higher
education was the university professor Victor Tufescu, who worked at the beginning (1941-
1946) at the same time as another great geographer and professor, Nicolae Al. Rădulescu
(1905-1989).
Victor Tufescu entered higher education coming from high school – almost a rule at that
time in Romania – before being a teacher at a traditional high school in Brasov (“Dr.
Meșotă”) between 1935 and 1938 (interrupted in 1936, when he was transferred to the
“Vemiamin” Theological Seminary in Iași, in order to prepare and defend his doctoral
thesis at the renowned university in the city), then followed a scholarship at Sorbonne
(1938-1939), where, newly married, he also took his wife, remaining with a pleasant image
for the rest of his life: “The stay in Paris was serene, a beautiful year of my life; I was
looking after my studies, Liliana was looking after French language courses”.
He was then a professor at the Higher Normal School in Bucharest (now the “I.L.
Caragiale” National College), with postgraduate rank. At the Academy of High
Vol. 26 • No. 65• February 2024 387

Commercial and Industrial Studies (A.I.S.C.I.) he was appointed substitute associated

professor in 1941, and when N.Al. Rădulescu left for the University of Bucharest, where he
became a full professor by competitive examination and, at the same time, head of the
Geography Department at A.I.S.C.I. (1946).
However, the professor began his academic career in a bad period, during the Second World
War, and in the one that followed, not less bad, which meant the sovietisation of Romania,
with major interference in higher education in general and in the economic one in particular.
His academic career, to which he had devoted himself with all his passion and skill, being
“the right man in the right place”, was, however, brutally cut short for five years (1951-
1956), when he was forced to work as a designer at the Institute of Forestry Research and
Management in Bucharest/Institutul de Cercetări și Amenajări Silvice București (1952-1955)
and, unfortunately, for nine months (November 1955-July 1956), he was arrested and sent to
prison. He was released and returned to his department in the ASE (then called the “Institute
of Economic Sciences and Planning V.I. Lenin”), where he remained until 1968, when he
moved to the Faculty of Geology and Geography (Geography section) of the University of
Bucharest, at the initiative (and insistence, as the professor told me in a discussion) of the
great mathematician Gheorghe Mihoc, then rector of the famous university (who insisted on
the idea of restoring the greatness of geography in this higher education institution).
Unfortunately, he would only teach for five years, and in the autumn of 1973, he was simply
removed from his professorship, quite unnaturally, after the start of the academic year. In
fact, it was political interference: a younger gentleman (I.P.), a relative of Ceaușescu's
political figurehead Manea Mănescu, needed a vacant teaching post to advance, and another
professor had his eye on one of the most significant offices in the faculty.
I have unwittingly been a direct witness to this machination. On that early October day, I
had an appointment with the professor in his office. When I got there, at the faculty, on the
third floor, on the right, in the last office on the left, I found that the office was closed,
which surprised me: I knew that not only me, but also the professor, was extremely
punctual. It jumped out at me that in the corridor in front of the desk was a large, massive
desk (an old-style desk), which I recognised as the professor's. As I sat there in total
bewilderment, I was approached by a young, burly assistant whom I knew, although he had
not been my “seminarian”, M.I., who said to me, in his frustrated style that I would become
accustomed to later, when we became good friends, that “you’re sitting there for nothing,
Tufescu has been dismissed, you wait for him for nothing. From now on, you worship
someone else and chances are that he is one of yours!” (I understood that, by the way, he
was from my native region, but his reaction was totally different).
By tradition, since its foundation in 1913, the Academy of Economic Studies has taught
subjects grouped under the name of Economic Geography. Thus, Professor Victor Tufescu
gave several courses in the period immediately following the Great World War, all of
which were published textbooks (and, in fact, lithographed at first).
a) Geography of Romania and neighbouring countries, taught in the academic year 1942-
1943;
b) Geography of Europe and Asia, taught in the academic year 1943-1944;

c) Regional Economic Geography, a two-year course:

- Large extra-European geo-economic areas, taught to first-year students (academic year
1946-1947);
- Large European geoeconomic areas, taught to second-year students (the same academic
year 1946-1947);
d) General economic geography, taught to first-year students (the same academic year
1946-1947);
e) Economic Geography of Romania, taught in the academic year 1948-1949;
f) Economic Romania and its geo-economic relations with neighbouring countries, course
introduced in the academic year 1948-1949 (it is not clear from the title page which year
students were taught).
General economic geography course. I found in the ASE Central Library and was able to
study, among other things, the lithographed course for students of the 1946-1947 academic
year, a real treatise in the field, having no less than 758 pages. This textbook, too, follows
the general rule of the professor's books: extremely logical structure, accuracy of treatment
(clear sentences, in a natural sequence), punctuation of essential elements, without
abdicating one's own scientific convictions. Although he was an extremely polite man, as
many of us knew him, he was not afraid to be very direct, often cutting interventions when
he did not agree with something and was convinced that things were different. In this
respect, I would give an example from the above-mentioned General Economic Geography
course. When he talks about economic geography (in the preamble to the course, entitled
Introduction), he declares himself dissatisfied with the definitions given to it by two of his
contemporaries, not just any two, but Simion Mehedinți, who was already recognised as the
“Patriarch of Romanian Geography”, and Nicolae Al. Rădulescu, his former colleague and
head of department at the Academy of Higher Commercial and Industrial Studies.
After the generous Introduction, in fact a more than
welcome theoretical approach, the work is structured in two
parts:
● Part One: Economic Factors, with two chapters (Nature
and Man respectively);
● Part Two: The Geography of Production, with six
chapters (Food Plants, Fruit and Beverages, Spices,
Narcotics and Medicinal Plants, Textile and Rubber Plants,
Animal Products, Mineral Products), plus the Agricultural
Regions of the Globe and the Major Domestic Animal
Regions.
Returning to the professorship will be beneficial both for
him and for the mass of students. Because, as Veselina
Urucu points out in the volume dedicated to the professor
Homage volume in 2000, “Victor Tufescu was, first and foremost, a
dedicated to Acad. professor. Being a professor, like for other young people in
Victor Tufescu the first half of the twentieth century, was an ideal, a noble
Vol. 26 • No. 65• February 2024 389

profession, a full achievement on the scale of social values. Becoming a university

professor was the culmination of a brilliant career for any Romanian intellectual”.1 This
appreciation in no way diminishes the exceptional scientific contribution of Professor
Victor Tufescu to the heritage of geography, which will reward him, on merit, although a
little late, with the title of academician, in 1992 (two years earlier he had been elected
corresponding member).
Professor's work. Professor Victor Tufescu has written not only a lot (more than 300
scientific studies, of which about 20 are books), but also well, and his personal imprint can
be felt. He was – without exaggeration of completeness – a perfect collaborator: he
delivered materials (book manuscripts or articles) on time, elaborated with an enviable
accuracy, with an outstanding logic and topicality of phrase. His handwriting was of perfect
elegance that reminded you of the old professional calligraphers. As for the content of his
writing, you had to be very well prepared, informed, armed with really strong arguments to
get him to change anything.
I also express the appreciation of Professor Mihai Ielenicz, former Dean of the Faculty of
Geography and President of the Romanian Geographical Society: “The scientific
achievements, the perfect pedagogical style based on perfect information, captivating
presentations and simple but complete and easy-to-follow
schemes, the oratorical art hard to put into words: chosen
phrases and intonation that made the rooms where he
exhibited always full; the proverbial calmness in any
dialogue or dispute, the depth of discussions and analyses of
any kind, the correct relations of understanding and support
for all those who wanted them, the masked passivity,
although his soul cried, in the face of the injustices created
by the regime or the open immobility of some
colleagues...”.
Of his many books, two stand out in terms of size, which of
course add to their value. Natural Modelling and
Accelerated Erosion (Academia Publishing House, 1966,
618 pages) and Romania. Nature, Man, Economy (Scientific
Publishing House, 1974, 530 pages). The former is the first
Geography of the RSR – treatise on dynamic geomorphology to appear in our
a textbook for the 12th country, and is the culmination of his research, thought, and
grade, a textbook that elaboration in studies in the physical field of geography.
so many generations The second is a unique, unparalleled geographical synthesis
of students, some of that offers an all-encompassing view of our country.
whom have become
Professor Victor Tufescu was, I believe, the most complex
great geographers, have
Romanian geographer, the only one who has dealt equally
learned from, written
with the two main branches of geography, physical and
together with Prof.
human, and their subbranches, combining them in the most
Claudiu Giurcăneanu,
blissful way in the treatise Romania. Nature, Man,
a fellow professor at ASE
1
Victor Tufescu, Romanian Geographers Collection, Institute of Geography, Romanian Academy, Bucharest,
2000 (volume coordinator Veselina Urucu); the coordinator's study is entitled Victor Tufescu. The Man. The
Professor. The Geographer.

Economy. There is, however, an inclination towards geomorphology, in the case of physical
geography, and towards the geography of human population and settlement and the
geography of agriculture, in the case of human geography. As regards geomorphology, in
addition to the treatise already mentioned and the numerous articles and studies, books such
as Subcarpathians (1962, 70 pages), Natural relief modelling and accelerated erosion
(1966, 618 pages), Subcarpathians and marginal hills of Transylvania (1966, 254 pages),
Torrents (1967, 68 pages) are worth mentioning.
I cannot also fail to note the love with which he has bent down and revealed the merits of
great geographical personalities (Dimitrie Cantemir, Emmanuel de Martonne, and Simion
Mehedinţi, including the critical edition: Terra. Introduction to Geography as a Science
and the volume Simion Mehedinți. Life and Works, 1994, 139 pages – or George Vâlsan,
Vintilă Mihăilescu, Tiberiu Morariu). He was also co-author and coordinator of several
preuniversity geography textbooks (The Geography of R.S. Romania, Geography of the
environment, Earth's fundamental geographical issues), and one of the most
knowledgeable and talented authors of geographical descriptions: In the Valley of Moldova.
Views and evocations (1970, 216 pages), Stopovers in the countryside (1976, 195 pages),
People of the Carpathians (1982, 207 pages).
These are just a few glimpses of memories and appreciations of Professor Victor Tufescu.
When I became a close collaborator of the professor and afterwards, for more than two
decades, I never suspected that one day I would become a university professor in the
department that he served, so beautifully and well, for almost three decades. For me it was
truly a great honour.
Professor Silviu Neguț was editor-in-chief of Encyclopaedic Publishing House

(1990-1993), and also vice-dean (1993-2008) and dean (2008-2012) of the Faculty of
International Business and Economics (REI) at ASE. His contributions are plenary.
Following in the footsteps of his predecessor, Professor Victor Tufescu, he is the author of
geography textbooks for preuniversity education (grades 5-12). His 2005 work Introduction
to Geopolitics, published by Meteor Press, will mark a decade since the introduction in the
undergraduate curriculum of the discipline of Geopolitics (1994), for the first time in post-
communist Romanian university education. At the same time, in 2004, it will found a
dedicated Master's programme. Also, under his signature, Geopolitics. The Universe of
Power, 2008 and Geopolitics, 2015 will follow. Thus, the “school of geopolitics at ASE” –
whose founder can be considered Professor Silviu Neguț – was taking shape. Through more
than 1300 radio broadcasts, he contributed to spreading knowledge of geography and
geopolitics, as no one else has. He was awarded the “Simion Mehedinți” Prize of the
Romanian Academy in 1997 for Mathematical Modelling in Human Geography, published
by Scientific Publishing House. In 2004, he also published The geography of tourism, in
2011, Human Geography, the first Romanian treatise in the field, and in 2022, Geostrategy
(with Marius-Cristian Neacșu). The library of his native commune, Vadu Pașii, Buzău
County, bears his name. A volume of ‘Writings in Honour of Professor Silviu Neguţ’ was
published in 2010 by CD Press. Another work, ‘Tireless Explorer of the Planetary and
Human Universe’, was dedicated to him in 2014, when the title of Doctor Honoris Causa
was awarded.
Vol. 26 • No. 65• February 2024 391

Journal No 65

Uploaded by

Copyright:

Available Formats

Journal No 65

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Journal No 65

Uploaded by

Copyright:

Available Formats

Amfiteatru Economic Journal

The Bucharest University of Economic Studies

Amfiteatru Economic Journal

The theme of the next issue

Issue no. 66/2024 – Innovative Application of AI in Business Impacting Socio-

The Journal is indexed/abstracted by the following international economic

Clarivate Analytics (2008)

EBSCO Publishing (2009)

ProQuest LLC (2012)

DOAJ – Directory of Open Access Journals (2010)

EconLIT – Journal of Economic Literature (2006)

SCOPUS – Elsevier B.V. Bibliographic Databases (2008)

IBSS – International Bibliography of the Social Sciences (2006)

RePEc – Research Papers in Economics (2004)

Cabell’s Directory of Publishing Opportunities (2006)

ISSN 1582-9146 www.amfiteatrueconomic.ro

Editorial Advisory Board

Challenges for Competence-Oriented Education in the Context of the Development

Identifying Sufficient and Necessary Competencies in the Effective Use

Challenges of Artificial Intelligence on the Learning Process in Higher Education .... 53

Exploring University Students’ Perceptions of Generative Artificial Intelligence

The Social Impact of Using Artificial Intelligence in Education .................................... 89

Exploring the Antecedents of Artificial Intelligence Products’ Usage.

Artificial Intelligence Adoption in the Workplace and Its Impact

Artificial Intelligence in Education: Next-Gen Teacher Perspectives ......................... 145

Generation Z Students’ Perceptions on the Abilities, Skills and Competencies

Artificial Intelligence and the Modelling of Teachers’ Competencies ......................... 181

Artificial Intelligence and Competency-Based Education: A Bibliometric Analysis . 220

Artificial Intelligence in Business Education: Benefits and Tools ................................ 241

Quantitative Evaluation of Willingness to Use Artificial Intelligence

Redesign of Accounting Education to Meet the Challenges of Artificial

Research on Tax Compliance Incentive Effects of Platform Companies

Amfiteatru Economic recommends

Architecture to Transform Classic Academic Courses into Adaptive Learning

Vol. 26 • No. 65• February 2024 5

Please cite this article as:

Vol. 26 • No. 65 • February 2024 7

Vol. 26 • No. 65 • February 2024 9

Vol. 26 • No. 65 • February 2024 11

Identifying Sufficient and Necessary Competencies in the Effective Use

Challenges of Artificial Intelligence on the Learning Process in Higher Education .... 53

Exploring University Students’ Perceptions of Generative Artificial Intelligence

The Social Impact of Using Artificial Intelligence in Education .................................... 89

Exploring the Antecedents of Artificial Intelligence Products’ Usage.

Artificial Intelligence Adoption in the Workplace and Its Impact

Artificial Intelligence in Education: Next-Gen Teacher Perspectives ......................... 145

Generation Z Students’ Perceptions on the Abilities, Skills and Competencies

Artificial Intelligence and the Modelling of Teachers’ Competencies ......................... 181

A Systematic Analysis of New Approaches to Digital Economic Education Based

Artificial Intelligence in Business Education: Benefits and Tools................................ 241

Quantitative Evaluation of Willingness to Use Artificial Intelligence

Redesign of Accounting Education to Meet the Challenges of Artificial

Vol. 26 • No. 65• February 2024 13

THE DEVELOPMENT OF EDUCATIONAL COMPETENCES FOR ROMANIAN

Please cite this article as: Article History

 Corresponding author: Giani Ionel Grădinaru – e-mail: [email protected]

Vol. 26 • No. 65 • February 2024 15

1. Review of the scientific literature

Vol. 26 • No. 65 • February 2024 17

couple in information and communication technology and in the disciplines of science,