1 s2.0 S0957417424010339 Main

Expert Systems With Applications 252 (2024) 124167
Contents lists available at ScienceDirect
Expert Systems With Applications

journal homepage: www.elsevier.com/locate/eswa
Artificial intelligence in education: A systematic literature review

Shan Wang a, 1, Fang Wang b, 2, *, Zhen Zhu c, 3, Jingxuan Wang c, Tam Tran a, Zhao Du d
a
Edwards School of Business, University of Saskatchewan, 25 Campus Drive, Saskatoon, SK, S7N 5A7, Canada
b
Lazaridis School of Business & Economics, Wilfrid Laurier University, Ontario, N2L 3C5, Canada
c
School of Economics and Management, China University of Geosciences, Wuhan, Hubei, 430078, China
d
Sport Business School, Beijing Sport University, Beijing, 100084, China
A R T I C L E I N F O A B S T R A C T
Keywords: Artificial intelligence (AI) in education (AIED) has evolved into a substantial body of literature with diverse
Artificial intelligence perspectives. In this review paper, we seek insights into three critical questions: (1) What are the primary cat
Education egories of AI applications explored in the education field? (2) What are the predominant research topics and their
Bibliometric analysis
key findings? (3) What is the status of major research design elements, including guiding theories, methodolo
Literature review
Content analysis
gies, and research contexts? A bibliometric analysis of 2,223 research articles followed by a content analysis of
selected 125 papers reveals a comprehensive conceptual structure of the existing literature. The extant AIED
research spans a wide spectrum of applications, encompassing those for adaptive learning and personalized
tutoring, intelligent assessment and management, profiling and prediction, and emerging products. Research
topics delve into both the technical design of education systems and the examination of the adoption, impacts,
and challenges associated with AIED. Furthermore, this review highlights the diverse range of theories applied in
the AIED literature, the multidisciplinary nature of publication venues, and underexplored research areas. In
sum, this research offers valuable insights for interested scholars to comprehend the current state of AIED
research and identify future research opportunities in this dynamic field.
1. Introduction as Absorb LMS and Docebo, deliver multiple AI capabilities to support

teaching and learning activities, such as intelligent content creation,
Information technologies, particularly artificial intelligence (AI), are administrative task automation, and personalized learning (Leh, 2022).
revolutionizing modern education. AI algorithms and educational robots In the realm of educational robots, SoftBank Robotics Nao and Pepper
are now integral to learning management and training systems, robots are developed to serve as language-teaching social robots (Bel
providing support for a wide array of teaching and learning activities paeme & Tanaka, 2022).
(Costa et al., 2017; García et al., 2007). Numerous applications of AI in The applications of AIED are rapidly evolving, reshaping the overall
education (AIED) have emerged. For example, Khan Academy offers teaching and learning landscape (Popenici & Kerr, 2017). The advent of
Khanmigo, an AI tutor harnessing GPT-4 capabilities, delivering generative AI technologies has introduced further opportunities,
personalized learning support and intelligent feedback across various attracting investment into and development of the AIED industry. The
subjects, including mathematics, programming, and language learning. global AIED market, valued at USD 1.82 billion in 2021, is projected to
Similarly, Duolingo, a language learning platform, uses sophisticated AI grow at a compound annual rate of 36 % from 2022 to 2030 (Grand
systems to improve learner experiences (Bicknell et al., 2023). iFlyTek ViewResearch, 2021). Learners, teachers, and educational institutions
offers intelligent assessment systems tailored for various grading sce are quickly embracing AIED. Recent statistics indicate that 43 % of
narios, including the national college entrance examination in China college students in the US use AI tools like ChatGPT and half of in
(iFlyTek, 2024). AI-powered learning management systems (LMS), such structors employ AI to develop their lessons (Businessolution.org, 2023).
* Corresponding author.
E-mail addresses: [email protected] (S. Wang), [email protected] (F. Wang), [email protected] (Z. Zhu), [email protected] (J. Wang), tmt143@
usask.ca (T. Tran), [email protected] (Z. Du).
1
ORCID: 0000-0002-0698-4341.
2
ORCID: 0000-0003-4263-1604.
3
ORCID: 0000-0001-7735-2192.
https://doi.org/10.1016/j.eswa.2024.124167
Received 15 January 2024; Received in revised form 14 April 2024; Accepted 4 May 2024
Available online 9 May 2024
0957-4174/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
S. Wang et al. Expert Systems With Applications 252 (2024) 124167
Moreover, AIED demonstrates its efficacy and effectiveness. Adaptive This research contributes to the literature of AIED in multiple ways.
learning enabled by AIED has been shown to improve student test results First, it offers a comprehensive understanding of the conceptual struc
by 62 %, while AI usage, in general, enhances student performance by ture of the AIED research, filling a gap in existing work. Moreover, in
30 % and reduces anxiety by 20 % (Businessolution.org, 2023). light of the recent trend of a substantial surge in AIED research articles
Concurrently, research on AIED has surged in recent years, yielding a and the review works on specific AIED domains, this study provides a
substantial body of work exploring various aspects of these applications, critical, up-to-date overview of the evolving research landscape, incor
including design, effectiveness, and outcomes (Chiu et al., 2023). This porating the latest articles. Additionally, the examination of the current
burgeoning research landscape has attracted review studies, which offer status of AIED research has unveiled underexplored research areas and
insights into the general AIED research field (Chassignol et al., 2018; highlighted essential future research directions. These include the
Goksel & Bozkurt, 2019; Guan et al., 2020; Hwang et al., 2020; Srini integration of new AI technologies, the elevation of theoretical contri
vasan, 2022) as well as specific topics such as learning analytics butions in research, and the enhancement of scientific rigor through
(Charitopoulos et al., 2020), machine learning and precision education theory-guided research design. These valuable insights may lend useful
(Luan & Chin-Chung, 2021), or educational AI within particular subject assistance in shaping the development of the AIED research field.
areas such as mathematics (Hwang & Tu, 2021) or STEM (Xu & Ouyang, This research is structured as follows. First, we provide a literature
2022). Nonetheless, few studies have systematically delineated the review of existing review studies in the AIED research field. Second, we
conceptual structure of the AIED research field and its theoretical un detail the process and results of the bibliometric analysis. We then
derpinnings, which are pivotal for understanding its current state and present a systematic literature review of a selected set of empirical
evolving prospects. research on AIED, offering insights into categories of AIED applications,
This review aims to provide an in-depth understanding of the con primary research topics, and common research design elements. We
ceptual structure of existing AIED research. Specifically, it addresses the conclude with a discussion of the major outcomes and the contributions
following research questions: of this study.
(1) What are the primary categories of AI applications explored in the 2. Literature review
education field?
(2) What are the predominant research topics and their key findings? AI is a subfield of computer science dedicated to understanding
(3) What is the status of major research design elements in the AIED human thought processes and recreating their effects through informa
field, including research methods, guiding theories, and research tion systems. The primary goal of AI is to create intelligent systems (i.e.,
contexts? computer programs or machines) that are capable of intelligent behav
iors (Rainer et al., 2016), including learning, reasoning, problem-
For these questions, this research employs a mixed research meth solving, perception, and creating. Typical examples of AI technologies
odology, combining a bibliometric analysis (Donthu et al., 2021) with a include expert systems, neural networks (including machine learning
systematic literature review (Snyder, 2019). Bibliometric analysis in and deep learning techniques), fuzzy logic, genetic algorithms, and
volves the quantitative summarization of metadata of extensive research intelligent agents (Rainer et al., 2016). Scholars often distinguish be
articles, including publication year, title, abstract, citations, authors, tween strong and weak AI (Wells, 2023). Strong AI, also known as
and institutions. It serves as an efficient method for grasping the state of artificial general intelligence, possesses a broad spectrum of human
a research field, particularly when the review scope is broad and the capabilities, including communication, reasoning, and emotional re
dataset is too extensive for manual examination (Donthu et al., 2021). In sponses, and is capable of multiple tasks. In contrast, weak AI, also
contrast, a systematic literature review, through content analysis of known as narrow AI, does not possess a full array of human capabilities
research articles, can delve into research nuances that are of interest to but can use algorithms to solve problems or reason for specific tasks,
researchers (Snyder, 2019). Together, these two complementary ap such as fraud detection and chess playing. AI applications that have been
proaches can provide a comprehensive view of the conceptual structure currently developed and brought into commercial use are categorized as
and emerging trends in the research field (Donthu et al., 2021). weak AI.
This research starts with a bibliometric analysis of 2,223 papers The field of education especially lends itself to AI technologies since
within the general topic of AIED. The descriptive analysis of the bib educational activities, including learning and teaching, are knowledge-
liometric metadata offers insights into publication trends, influential intensive cognitive activities, and AI applications, which are created for
journal sources, and key articles. To gain a comprehensive under cognition and problem-solving based on algorithms and knowledge
standing of emerging research concepts, we provide the co-occurrence base, can effectively support and augment educators’ and learners’
networks of two types of keywords that are associated with articles: abilities in teaching and learning. Since the advent of AI in the mid-1950
keywords plus and author keywords. Next, we selected and coded 125 s, AI technologies have been increasingly applied to facilitate education
empirical research articles for a systematic literature review, including and training in various subjects, including language, STEM, and medi
AIED applications, research topics, and other research design details, cine (Perrotta & Selwyn, 2020). To date, AIED applications are devel
such as research methodologies, background theories, and research oped to support teaching and learning activities such as content
contexts. preparation and dissemination, interactions and collaboration, and
The coding results show four primary categories of AI applications performance assessment (Chassignol et al., 2018; Perrotta & Selwyn,
within the AIED literature, including adaptive learning and personalized 2020).
tutoring, intelligent assessment and management, profiling and predic A substantial body of studies has examined AIED applications,
tion, and emerging products, with adaptive learning and personalized leading to review studies in the field. Table 1 provides a list of recent
tutoring being the most studied. The research topics range from system review articles. Several reviews pertain to the general field of AIED
design and implementation, adoption and use, AIED impacts, and its (Chassignol et al., 2018; Chen, Xie, & Hwang, 2020; Chen, Xie, Zou,
challenges, with system design and implementation being the most et al., 2020; Chiu et al., 2023; Goksel & Bozkurt, 2019; Guan et al.,
popular topic. The coding also reveals that experiments are the most 2020), while most focus on a specific application area such as chatbot
frequently used research methodology, and several learning theories, (Okonkwo & Ade-Ibijola, 2021), precision education (Luan & Chin-
including constructivist learning theory, learning style theory, cognitive Chung, 2021), mathematics education (Hwang & Tu, 2021), STEM
theories of learning, and item response theory, are among the most (Xu & Ouyang, 2022), or student assessment (González-Calatayud et al.,
employed theories that guide the research design. Higher education is 2021). Scholars have used bibliometric, systematic or simply narrative
the most frequent research context. reviews in their investigation of the field. For example, through a
2
Table 1
Literature Review: Major Review Studies in Recent Years.
Article Type Review content Time of articles Research domain
reviewed
Kulik & Fletcher, Systematic review of 50 papers Synthesized and analyzed the effect sizes of the effectiveness of Not specified Intelligent tutoring
2016 intelligent tutoring systems systems
Chassignol et al., Narrative review Developed a framework that classifies AIED applications by different Not specified General
2018 components of education process: content, teaching methods,
assessment and communication
Zhang et al., 2018 Bibliometric study of 1,579 Conducted descriptive analyses of bibliometric data, including top 1995–2008 Learning analytics
papers authors and journals; Summarized four methods in learning analytics
and their evolution patterns.
Hinojo-Lucena Bibliometric study of 132 papers Conducted a descriptive study of bibliometric data, including 2007–2017 Higher education
et al., 2019 publication trend, sources, authors, organizations, and countries
Zawacki-Richter Systematic review of 146 papers Conducted descriptive analyses of bibliometric data, including 2007–2018 Higher education
et al., 2019 publication trends, journals, countries, author affiliation and
methods; Summarized AIED applications.
Charitopoulos et al., Systematic review of 316 papers Coded education problems addressed, learning contexts, soft 2010–2018 Educational data mining
2020 computing methods employed and major journal outlets for each area and learning analytics
of educational data mining and learning analytics research
Chen, Xie, & Bibliometrics of 9,560 papers Conducted a descriptive study of bibliometric data, including grants, 1999–2019 General
Hwang, 2020 conferences, journals, software tools, institutions, and researchers
Chen, Xie, Zou, Systematic analysis of 45 papers Reviewed publication journals, citation counts, regions and 1990–2016 General
et al., 2020 institutions of authors, application scenario types, research issues,
adopted AI technologies, and definitions concerned
Guan et al., 2020 Computer-assisted text analysis Reviewed research themes over two stages (i.e., 2000–2009 and 2000–2019 General
and manual content analysis of 2010–2018).
425 papers
González-Calatayud Systematic review of 22 papers Reviewed AI definition, pedagogical models used, reasons for using 2010–2020 AI for student assessment
et al., 2021 AI, the use of automated scoring, and comparison of assessment
evaluation accuracy between AI use and non-use
Hwang & Tu, 2021 Bibliometric mapping analysis Reviewed publication journal, paper citations, cited authors, 1996–2020 AI in mathematics
and systematic review of 43 keywords, application domains, sample groups, research methods, education
papers roles of AI, adopted AI algorithms and research issues
Luan & Chin- Systematic review of 40 Reviewed multiple elements of research papers, including research 2016–2020 Machine-learning-based
Chung, 2021 empirical papers purpose, education context, data sources, learners’ individual precision education
differences, learning outcomes, learning algorithms, evaluation of
algorithms, and major research findings
Okonkwo & Ade- Systematic review of 53 papers Summarized types of chatbot applications, their benefits, 2015–2021 Chatbots in education
Ibijola, 2021 implementation challenges, and potential future areas
Celik et al., 2022 Systematic review of 44 papers Reviewed the role of teachers in AIED, the advantages that AI offers 2004–2020 Teachers’ perspective
teachers, the challenges teachers face when using AI, and AI methods
in AI-based research with teachers
Xu & Ouyang, 2022 Systematic review of 63 Summarized AI applications in STEM education and their associated 2011–2021 AI in STEM education
empirical AI-STEM studies elements such as educational information (content), subjects (leaners
and instructors), medium, and environment
Chiu et al., 2023 Systematic review of 92 papers Summarized AIED applications and outcomes, including applications 2012–2021 General
in the domains of learning, teaching, assessment and admin, and
outcomes related to teachers and learners
narrative review, Chassignol et al. (2018) summarized and presented the AIED research, which are critical in comprehending the current body
their major literature findings in a framework with four components of of studies and charting future research development.
the educational process: content, teaching method, assessment, and
communication. Goksel and Bozkurt (2019) conducted a co-word anal 3. Bibliometric analysis of AI in education research
ysis of the keywords in 393 papers between 1970 and 2018, summari
zing three key concepts in the AIED literature, including adaptive 3.1. Data collection
learning, personalization and learning styles, and expert systems and
intelligent tutoring systems. Xu and Ouyang (2022) conducted a sys This research uses the Web of Science (WoS) database to compile an
tematic review of 63 empirical AI-STEM research from 2011 to 2021, initial set of papers. The WoS database is a commonly employed
summarizing AI applications in STEM education, their characteristics resource for conducting systematic literature reviews. Following the
and effects. methodology outlined by Goksel and Bozkurt (2019), we conducted a
Despite existing review studies on AIED research, there is a need for a search in WoS in June 2022 to retrieve English publications that contain
comprehensive review of the up-to-date literature to gain insights into the terms “artificial intelligence” and “education” in their title, abstract,
the conceptual structure of the field. First, the majority of the existing or keywords. This initial search yielded a total of 3,690 articles. We then
review focuses on AIED applications and their characteristics (Chassi performed a manual screening to assess the relevance of these articles to
gnol et al., 2018; Chiu et al., 2023; Xu & Ouyang, 2022), missing a our focus on AIED. Any publications deemed irrelevant or lacking sub
higher-level comprehensive overview of research topics and methodol stantial content on AIED were removed from our dataset. Additionally,
ogies, which is key to scholarly interest. Second, existing review studies we retained only scholarly works with full-text access, encompassing
on the general AIED field are mainly based on the sample of articles journal articles and conference papers. The final dataset comprised
before 2019. However, the COVID-19 pandemic spurs the adoption of AI 2,223 articles published between 1984 and June 2022. Subsequently,
and the research of AIED. This up-to-date sample needs to be reviewed we performed a bibliometric analysis of these 2,223 articles utilizing the
and their insights aggregated. Finally, there is a lack of examination of R package “bibliometrix” and its interactive web version “biblioshiny”,
the foundational theories that are commonly employed in and steering as developed by Aria and Cuccurullo (2017).
3
3.2. Descriptive analysis of bibliometric data classic paper on user adoption and behavior towards information sys
tems, and Felder’s (1988) widely cited paper on teaching and learning
Table 2 summarizes the basic information of the articles in our styles. These cited references suggest three disciplines foundational to
dataset. The publication dates of the 2,223 articles span from 1984 to AIED research: Computer Science and AI, MIS, and Education.
June 2022. These articles were published across 1,247 journals and To further identify the impacts of the papers in our sample, Fig. 4
collectively cite 60,764 references. In June 2022, the average age of illustrates the average article citation per year (AACPR). AACPR repre
these articles was 5.62 years, indicating that over half of the AIED sents the total number of citations received by papers published in a
research papers were published after 2016. To further investigate the specific year, normalized by citable years (i.e., the number of years since
publication trends in AIED, Fig. 1 illustrates the growth of this field. publication). This normalization accounts for the fact that older publi
Notably, AIED did not emerge as a prominent research area until 2017. cations tend to accumulate more citations over time. Normalizing cita
The annual publication counts never exceeded 50 articles from 1984 to tions in this way can mitigate the age effect when assessing paper quality
2016. However, since 2017, this field has garnered considerable based on citations. As shown in Fig. 4, AACPR demonstrates an upward
research attention, experiencing a significant surge between 2019 and trend for papers published since 2014. This indicates that recent pub
2021. This growth can be attributed to the rapid advancement of AI lications tend to attract more citations than their older publications,
capabilities in recent years (Roser, 2022) and the transformation to suggesting a growing impact of recent research in the field.
online teaching during the COVID-19 pandemic (Du et al., 2022). In addition, Fig. 4 marks specific papers that contribute to notable
An examination of the top-cited journals and articles reveals further spikes in the AACPR trend line. The highest spike occurs in 1990. In our
insights. Fig. 2a lists the top 10 journals that publish the most numbers sample, only one paper, namely, Nwana (1990) review article on
of articles in our sample, and Fig. 2b shows the top 10 locally cited intelligent tutoring systems is from that year. This paper has garnered a
sources (i.e., journals cited by the articles in our sample). The two substantial number of citations, underscoring both its quality and the
journal lists demonstrate that AIED is a cross-disciplinary field. Research enduring interest in the topic of intelligent tutoring systems over the
is published in Computer Science journals (such as Journal of Intelligent past 28 years. Other heavily cited papers include Chou et al.’s (2003)
and Fuzzy Systems, Wireless Communications and Mobile Computing, IEEE research on learning companions and educational agents, García et al.’s
Access), Education journals (such as International Journal of Emerging (2007) study on students’ learning styles detection, and Dwivedi et al.’s
Technologies in Learning, and Computer and Education), and Management (2020) commentary on the impact of COVID-19 on education. These
Information Systems (MIS) journals (such as Computer in Human papers have significantly contributed to the scholarly discourse in AIED.
Behavior). Comparing Fig. 2a and 2b shows that open-access journals
have served as a major outlet for AIED research in terms of the number 3.3. Keyword co-occurrence analysis
of papers published—9 out of the 10 journals shown in Fig. 2a are open-
access journals (except for Journal of Intelligent and Fuzzy Systems), To understand the conceptual structure of the literature, we conduct
whereas traditional, established journals such as Computer and Education keyword co-occurrence analysis, also known as co-word analysis. In this
and Computers in Human Behavior are more impactful in terms of cita analysis, a co-word network is constructed, where nodes represent
tions they attract. Both open-access journals and traditional journals keywords, edges signify co-occurrence relationships, and edge weights
contribute to the dissemination of knowledge on AIED. indicate the frequency of co-occurrences within the literature body.
To gain further insights into the impactful work in AIED, we sum Keywords provide concise summaries of research works and are well-
marize the top 15 globally cited papers, the top 15 locally cited papers, suited for co-occurrence analyses, allowing us to discover structural
and the top 15 cited references in our sample in Fig. 3a, 3b, and 3c, patterns among core concepts in the literature. Our analysis utilizes both
respectively. The full information of these papers is provided in the keywords plus and author keywords available in WoS: keywords plus are
Appendix. The top globally and locally cited papers cover a range of standardized keywords provided by WoS, and author keywords are
themes in AIED, which can be roughly classified into three categories, provided by authors in their articles. Prior to the analysis, we preprocess
including (1) general opinion papers (Dwivedi et al., 2020; Gadanidis, and clean the keywords, making necessary adjustments such as replac
2017) and literature reviews (Chen et al., 2020; Hinojo-Lucena et al., ing “AI” with “artificial intelligence” and standardizing both “student”
2019; Zawacki-Richter et al., 2019); (2) research on popular AIED ap and “students” to “students”.
plications, including machine learning and precision education (Costa Fig. 5 illustrates the distributions of the top 50 keywords plus and
et al., 2017; Duong et al., 2019), intelligent tutoring (Nwana, 1990), author keywords in panels a and b, respectively. Notably, “artificial
learning companion agents (Chou et al., 2003), chatbots (Fryer et al., intelligence” and “education” occupy the top two positions in both
2017), and educational robotics (Murphy, 2001); (3) research on panels, as these were the primary search terms used to identify the
perception and attitude towards AI systems (Sit et al., 2020). literature body. Beyond these two keywords, the lists of keywords plus
An examination of the top 15 cited references by the papers in our and author keywords exhibit significant differences, with author key
sample (see Appendix 1c) reveals the disciplinary foundations of AIED words being more diverse and closely tied to the content of the articles.
research. Except for the above-mentioned AIED topics, classic work in First, keywords plus are more broadly descriptive, while author key
fields of computer science and AI, MIS, and education are cited, words are more specific to article content. For instance, the top key
including Turing’s (1950) impactful work on machine intelligence, words in keywords plus, excluding the top 2, consist of general terms
Russell and Norvig’s (2002) popular textbook on AI, David’s (1989) such as “system”, “students”, “performance”, “design”, “technology”,
“models”, and “sciences”. In contrast, author keywords delve into
specialized niche areas within AIED, including terms like “machine
Table 2
Article Information in the Sample. learning”, “higher education”, “e-learning”, “intelligent tutoring sys
tem”, and “robotics”. Second, the distribution of author keywords is
Description Results
more skewed than that of keywords plus. The distribution of author
Timespan 1984:2022(June) keywords is heavily biased towards the two search keywords, “artificial
Journals included 1,247
intelligence” and “education”. This skewness is probably due to the
Articles included 2,223
Average years from publication 5.62 greater diversity and less standardized nature of author keywords
Average citations per documents 4.09 compared to keywords plus.
Cited references 60,764
Keywords plus (ID) 1,336 3.3.1. Co-occurrence network of keywords plus
Author keywords (DE) 5,076
To further discover the conceptual patterns of the literature, we
4
Fig. 1. Annual Scientific Production (Note: bars with number of publications less than 20 are not labelled).
Fig. 2. Important Journal .

Sources
provide the co-occurrence network of the two types of keywords in The big data analytics cluster mainly focuses on constructing systems,
Fig. 6. The network analysis was carried out using the biblioNetwork() models, frameworks, and environments that utilize big data and algo
function within the R bibliometrix package. The network includes the rithms to make predictions within the educational context. Keywords
top 50 nodes, and clusters were identified using the Louvain clustering associated with this cluster consist of “model”, “framework”, “system”,
algorithm. Fig. 6a displays the results for the keywords plus co- “environment”, “algorithms”, “big data”, and “analytics”. Additionally,
occurrence network analysis, revealing four distinct concept clusters: the term “classification” is frequently co-occurring, suggesting that
user behaviors, design science, big data analytics, and AIED impacts. classification algorithms are commonly used within this research
The user behaviors cluster primarily revolves around user intentions domain.
and behaviors towards AIED systems, such as system adoption and use The AIED impacts cluster pertains to the influence of AI, particularly
behaviors. Typical keywords within this cluster include “behavior”, its impact on learner skills, learning quality, and experiences resulting
“perceptions”, “user acceptance”, “intention”, and “engagement”. Some from the use of new technologies or technology-supported simulated
keywords, such as “health”, “language”, “children”, and “higher edu learning environments. Relevant keywords encompass “experience”,
cation”, point to the contexts of studies. That is, health(care)-related “skills”, “impact”, “quality”, “simulation”, and “challenges”. The term
education, language education, children’s education, and higher edu “perspectives” emerges within this cluster, indicating that the impact is
cation are focal research contexts of AIED behavior research. viewed from multiple angles. This cluster aligns with the concept of
The design science cluster predominantly concerns the design and “21st-century skills”, as identified by (Chiu et al., 2023) in their sys
implementation of AIED systems and tools. Key terms associated with tematic review of the impacts of AIED. Their research suggests that AIED
this cluster encompass “design”, “science”, “implementation”, “feed can help students acquire problem-solving and online collaboration
back”, and “tools”. Some keywords indicate the current technological skills, enhancing learning quality.
trends upon which these systems are constructed, including “augmented
reality”, “virtual reality”, and “tutoring systems”. The term “knowledge” 3.3.2. Co-occurrence network of author keywords
frequently co-occurs, possibly signifying the importance of establishing The results of the author keyword co-occurrence network analysis
knowledge bases and knowledge mapping within AIED systems. are more varied and less coherent than those from the keywords plus co-
5
Fig. 3. Most Cited Papers and References.
occurrence analysis, primarily because author keywords are diverse in clusters: machine learning, educational technology, learning systems,
nature. To avoid potential bias from the dominance and significant emerging technologies, and AI education.
skewness of the top two search terms, “artificial intelligence” and “ed The machine learning cluster primarily focuses on the application of
ucation,” as shown in the above keyword distribution analysis, we data science and AI technologies for educational data mining and
excluded these terms from the network analysis. Fig. 6b presents the co- learning analytics. This cluster includes terms related to AI techniques
occurrence network analysis result, revealing five distinct concept such as “machine learning”, “deep learning”, “neural networks”,
6
Fig. 4. Average Article Citation Per Year (AACPY).
Fig. 5. Most Frequent Keywords.
“natural language processing”, and “intelligent systems”. It also in The educational technology cluster centers on the use of AI technology
corporates the term “K-12 education”, indicating a main context of study to enhance teaching and learning, with a particular emphasis on
in the machine learning cluster. implementing AI such as pedagogical agents and intelligent pedagogical
7
Fig. 6. Keyword Co-occurrence Networks.
assistance system to support creative and innovative pedagogical prac terms like “models” and “frameworks” in the big data analytics
tices. Two AI techniques, “chatbot” and “fuzzy logic”, are frequently cluster.
associated with research in this cluster. The educational technology, learning systems, and emerging tech
The learning systems cluster is primarily concerned with the devel nologies clusters in the author keyword network collectively corre
opment of innovative learning support systems, including intelligent spond to the broader design science cluster in the keywords plus
tutoring systems and learning management systems. Adaptive learning network, yet reflect different focuses in research. The educational
is a prominent direction of innovation within this cluster. Additionally, technology, learning systems, and emerging technologies clusters in
terms such as “MOOC”, “distance education”, “online learning”, and the author keyword network are concepts associated with AIED ap
“higher education” are frequently used as the contexts of studying plications, while the design science cluster in the keywords plus
learning systems. network seems to suggest the potential research topic on designing
The emerging technologies cluster contains a list of emerging technol these applications.
ogies relevant to AIED, including “big data analytics”, “augmented re The AIED impacts cluster in the keywords plus network is associated
ality”, “virtual reality”, “cloud computing”, “internet of things”, and with the AI education cluster in the author keyword network, as both
“industry 4.0″. Contextual terms like ”COVID-19″, “distance learning”, focus on studying the knowledge and skills associated with AIED.
and “engineering education” are associated with this cluster, indicating However, the AI education cluster delves into more detailed skills
that the COVID-19 pandemic and the resulting shift to distance learning and disciplinary knowledge for students to learn about AI, while the
have accelerated the applications and research related to emerging AIED impacts cluster focuses on skills learned as a result of using
technologies. AIED applications.
The AI education cluster primarily focuses on the education and Interestingly, the user behaviors cluster does not prominently
learning of AI-related knowledge and skills. These include STEM and emerge in the author keyword network. This may be attributed to the
computer science-related knowledge and skills, such as “learning arti less standardized nature of author keywords and the inclusion of
ficial intelligence”, “tools”, “computer science education”, “task anal only the top 50 nodes in the network for clarity. Behavior keywords
ysis”, “robotics”, and “computational thinking”. may be wide-ranged in the author keywords, and thus do not
The keyword co-occurrence analysis serves as a foundational tool for prominently appear in the top 50 nodes.
comprehending the central concepts within the literature. Both analyses
have identified common concept clusters, which are summarized on the These findings provide valuable insights into the distribution of
left side of Fig. 7. Notably: research topics and concepts within the AIED literature, highlighting the
prevalence and distinctiveness of several important research areas.
The big data analytics cluster in the keywords plus co-occurrence
network relates to the machine learning cluster in the author 4. Content analysis
keyword network. However, the machine learning cluster delves
deeper into machine learning concepts compared to the more general Bibliometric analysis is inherently data-driven and does not probe
8
Fig. 7. Conceptual Mapping between Co-occurrence Networks and Systematic Review.
into the content details of research articles. Consequently, we comple 4.1. Paper selection and analysis
ment the bibliometric analysis with a systematic literature review that
involves manual content analysis of a smaller sample of articles. The To select a subset of papers for content analysis from the initial pool
results obtained from the co-occurrence network analysis serve as a of 2,223 papers used in the bibliometric analysis, we applied the
foundation for coding the selected papers in the systematic review. The following criteria: (1) Papers that are published in journals of the cate
results pertaining to research designs from the content analysis are gory quartile Q1 in the Journal Citation Reports; (2) Papers that clearly
summarized on the right side of Fig. 7. Fig. 7 also illustrates a mapping describe the AI applications under study and report their impact on
between concepts derived from the co-occurrence network analysis and teaching and learning; and (3) Papers that contain an empirical study.
those derived from the systematic analysis of the selected papers. We focus on empirical studies because they go beyond conceptual un
derstanding and provide empirical evidence in addressing specific
research questions. Understanding their content and research methods
Fig. 8. Coding Schema.
9
can provide valuable insights into the development of the AIED research and create an adaptive learning environment for learners based on their
field and guide future research. A total of 125 papers were identified for knowledge level, learning style, emotional state, and interest prefer
content analysis. ences. These applications have evolved significantly in recent years,
As illustrated in Fig. 8, we coded common components of research transitioning from rule-based expert systems to more complex AI tech
design in each article. These components included research objects, niques and algorithms like neural networks and decision trees. The
objectives, methodologies, guiding theories, and educational contexts. design of these applications has become increasingly interactive and
learner-centered. Two sub-categories of adaptive learning and person
4.2. Results of content analysis alized tutoring applications include intelligent tutoring systems and
adaptive hypermedia learning systems.
Fig. 9 illustrates the distributions of various types of AIED applica Intelligent tutoring systems (ITS) are computer-assisted instructional
tions, research topics, research methods, guiding theories, and educa systems that harness the power of AI technologies to emulate the role of
tion contexts coded in the subset of papers in content analysis. a human tutor. These systems are designed to offer immediate and
personalized instruction or feedback to students under specific educa
4.2.1. Research objects: AI applications in education tional strategies (Hooshyar et al., 2015). ITS research frequently men
As depicted in Fig. 9a and Fig. 10, we have identified four primary tions about “intelligent tutor”, “intelligent tutoring system” (Aleven
categories of AIED applications, each with its corresponding sub- et al., 2009), and “intelligent agent” (Xu & Wang, 2006). The functions
categories: (1) Adaptive learning and personalized tutoring, (2) Intelli of ITS applications, as studied in the 34 articles within our sample, can
gent assessment and management, (3) Profiling and prediction, and (4) be categorized into three primary types: learner status diagnosis and
Emerging technologies or products. Among all applications (shown in adaptive feedback provision (n = 24), adaptive test and exercise pro
Fig. 9a), the most studied are adaptive learning and personalized vision (n = 5), and adaptive learning content recommendation (n = 5).
tutoring applications (40 % of papers in our sample), followed by Like a human tutor, ITS can monitor and diagnose learners’ learning
intelligent assessment and management applications (24.8 %), profiling progress and provide targeted feedback and guidance. For instance,
and prediction applications (20 %), and emergent products in education Gülcü (2009) proposed and tested “ZOSMAT”, a mathematical tutoring
(15.2 %). system that tracks a student’s learning journey and offers personalized
guidance based on their performance. Leveraging AI technology, edu
4.2.1.1. Adaptive learning and personalized tutoring applications. This cators can provide specialized support to learners experiencing diffi
category of AIED applications aims to customize the learning process culties, ultimately achieving precision education (Lin & Lai, 2021). In
Fig. 9. Distribution of Coding Categories.
10
Fig. 10. AIED Applications.
addition, ITS can provide exercises that align with learners’ cognitive 4.2.1.2. Intelligent assessment and management applications. Tracking
abilities. Craig et al. (2013) presented a system for mathematics after- and recording students’ learning progress and providing timely evalu
school intervention, which dynamically selects exercise content based ation are challenging for instructors, particularly in a large class setting.
on students’ mastery levels. Beyond tests and exercises, ITS excel at Intelligent assessment and management applications have been devel
recommending learning materials tailored to learners’ status. For oped to address these challenges by offering automatic grading and
instance, they can suggest English reading materials (Hsu et al., 2010) evaluation capabilities and support for collaborative learning and
and learning remediation materials (Lin et al., 2016) to minimize stu resource management. As a result, these applications can be categorized
dents’ knowledge disorientation. Due to its versatility and numerous into two primary types: intelligent assessment systems and learning
advantages, ITS find applications in various contexts. These include management systems.
mathematics, chemistry (Rau et al., 2015), sports training (Liu et al., Intelligent assessment systems (IAS) leverage AI technologies to
2021), medical guidance (Poitras et al., 2016), and even animation conduct assessment tasks with high accuracy and efficiency in educa
design teaching (Tang et al., 2022). These diverse applications showcase tional settings. They offer valuable feedback to both students and in
how ITS enhance learning experiences across a wide range of domains. structors. Their functions discussed in the 18 articles within our sample
Adaptive hypermedia learning systems (AHLS) place a heightened can be categorized into three primary types: assessing students’ learning
emphasis on accommodating learners’ learning styles and preferences, abilities and behaviors (n = 9), automatic grading (n = 7), and teaching
tailoring hypermedia-enabled presentations and navigation support to evaluation (n = 2).
individual students. This approach positions students at the center of the The student assessment function encompasses various elements,
learning environment. AHLS functions explored in the 16 articles from including gauging students’ knowledge levels (Cheng et al., 2022),
our sample can be categorized into three primary types: learning style assessing their overall competencies (Niu, 2022), and evaluating their
detection or analysis (n = 8), adaptive presentation (n = 4), and adap learning attitudes (Chen et al., 2007). Student assessment can aid
tive path navigation support (n = 6). learners in adjusting their learning strategies to enhance their learning
Students exhibit diverse learning styles, including variations like outcomes. It can also assist educators in tailoring their teaching methods
reflective or active learning, field-dependent or independent learning, to suit students’ aptitudes. The automatic grading function is fast
and intuitive or sensitive learning (García et al., 2007). Applications evolving with a growing trend in activities such as oral training and
focusing on learning style detection or analysis frequently employ highly creative writing. For instance, Fu et al. (2020) utilized AI-based
classification algorithms like Bayesian networks (García et al., 2008; digital automatic scoring tools to provide learners with real-time
Schiaffino et al., 2008), neural networks (Lo et al., 2012), and decision scores and pronunciation corrections, offering immediate feedback.
trees (Lin et al., 2013). These algorithms analyze and identify students’ Wilson et al. (2021) investigated an automated writing evaluation sys
learning styles within an e-learning system, enabling adaptive adjust tem called MI Write, which assesses students’ writing quality across six
ments to content presentation according to individual preferences. For dimensions, including critical thinking development, style, and other
instance, in the UZWEBMAT system designed by Özyurt et al. (2013), personalized traits. The teaching evaluation function employs data
learning objects were tailored in three distinct ways to align with visu mining technology to analyze students’ classroom evaluation question
al–auditory–kinesthetic (VAK) learning styles (Fleming, 2001) for each naires (Agaoglu, 2016).
subject. Lo et al. (2012) crafted an adaptive web-based system featuring Learning management systems (LMS) play a crucial role in facilitating
an adaptive network interface that aligns with students’ cognitive styles. teaching management tasks, such as delivering learning resources to
Additionally, AHLS offer students guidance and navigation assis students, overseeing and enhancing learner interactions, and stream
tance to expedite the discovery of learning materials or recommend the lining course administrative workflows (Şahin & Yurdugül, 2022). Some
most effective learning pathways. These pathways can adhere to the LMS platforms also support the responsibilities of school administrators.
originally planned learning path (Lin et al., 2013) or offer remedial As a central management platform, LMS serves as a hub connecting in
paths for corrective learning (Hsieh et al., 2013). This adaptability structors, students, and administrators, aggregating substantial volumes
empowers students to navigate their learning experiences more of activity data. AI is increasingly integrated into LMS design to enable
effectively. adaptive and intelligent management of learning and teaching activities.
The functions of LMS studied in the 13 articles within our sample can be
11
grouped into three primary categories: collaborative learning support (n instructional requirements of particular scenarios. Currently, educa
= 5), classroom and exam management (n = 4), and information and tional chatbots serve primarily four functions: (a) classroom teachers,
resource management (n = 4). (b) peer support, (c) companions to foster emotional connections, and
The collaborative learning support function in LMS encompasses two (d) telepresence robot teachers (Sharkey, 2016). For instance, Chen et al.
distinct types: promoting collaborative writing through document (2021) investigated a game-based intelligent robot designed for teach
visualization (Calvo et al., 2011) and facilitating remote group discus ing Chinese idioms. This robot utilizes sound effects and engaging vi
sions (Chen & Tsao, 2021). In remote group discussions, instructors can suals to pique children’s interest in learning content. Furthermore,
assign specific roles to students, monitor their participation and educational robots serve as conversational agents for language conver
collaboration, and receive alerts in case of conflicts among students, sation exercises (Zhang & Han, 2021) or engage in book discussions with
allowing timely intervention (Casamayor et al., 2009). The classroom students (Liu et al., 2022), promoting learner autonomy through
and exam management function relies on the integration of various personalized methods. A study evaluating instructional functions and
technologies to enhance classroom teaching. This includes features such effectiveness of chatbots identified three key advantages of integrating
as conducting remotely controlled experiments (Kong et al., 2009), chatbots into teaching: enhanced interaction with students, increased
providing online examinations (Tasci et al., 2014), delivering real-time feedback for students, and user-friendliness (Vázquez-Cano et al., 2021).
automated notifications on student performance in multi-tabletop However, some studies also suggest that while chatbots can yield useful
enabled classrooms (Martinez-Maldonado et al., 2015), and automati effects, maintaining long-term student interests and engagement is still
cally detecting students’ focus and attention in classroom settings (Chiu challenging (Fryer et al., 2017).
& Tseng, 2021). The information and resource management function VR and AR are cutting-edge immersive technologies that seamlessly
supports the exchange of educational materials between instructors and integrate the virtual world and the real world in real time, offering users
students (Lin et al., 2009), course content management by instructors experiences that the physical world alone cannot provide. Of these
(Peredo et al., 2011; Yaghmaie & Bahreininejad, 2011), and the technologies, AR has gained significant popularity in educational con
recommendation of online courses to students. Recent research proposes texts. Out of the 12 relevant articles in our sample, 8 focused on AR
integrating virtual assistants into LMS to assist students in navigating applications in education. For instance, Chen et al. (2022) proposed a
course content and providing notifications about various course activ children’s digital art ability training system with artificial intelligence-
ities stored in the LMS calendar (Wang & Park, 2021). assisted learning for contour recognition, tone color matching, and
color ratio calculation. Students can use smart glasses to view AR
4.2.1.3. Profiling and prediction applications. AIED applications focused paintings, enhancing their imagination and painting capabilities. Lin
on profiling and prediction leverage educational data mining and et al. (2021) examined the influence of AR-enabled AI of Things (AIoT)
learning analytics to identify learner characteristics, forecast their learning on computational thinking skills training, concluding that AIoT
learning outcomes, empower learners with greater control over their learning can increase students’ motivation to learn and has a positive
education, and enable educators to identify and assist at-risk students, impact on their problem-solving and comprehension. In addition, four
thereby reducing the likelihood of academic failure (López-Zambrano articles in our sample explored VR applications in various educational
et al., 2021). Depending on their predictive objectives, functions of contexts, including art education (Rong et al., 2022) and 360-degree
profiling and prediction applications in the 19 relevant articles within content presentation (Kim et al., 2022). The findings from these
our sample can be categorized into four types: failure and dropout studies confirm that AI-supported virtual learning modes are engaging
warning (n = 6), academic achievement prediction (n = 5), learning and appealing to students.
analysis and learner modeling (n = 4), and course arrangement and
teaching performance (n = 4). 4.2.2. Research topics
Many profiling and prediction applications rely on learner models, We identified four main categories of research topics within our
which gather and analyze learners’ behavioral data to offer effective sample of AIED studies: system and application design, adoption and
cognitive and management support. Various methods, such as support acceptance of AIED, impacts of AIED, and challenges of AIED. Fig. 9b
vector machines, decision trees, neural networks, and naive Bayes illustrates the distribution of these research topics. System and appli
models (Winkler et al., 2021), are employed to build learner models and cation design emerges as the most frequently studied topic, accounting
generate predictions regarding their performance, typically of a quan for 52.8 % of the sample, followed by impacts of AIED at 39.2 %.
titative nature. Nabizadeh et al. (2022) discussed AI applications that Adoption and acceptance of AIED represent a small proportion at 5.6 %,
not only deliver prediction results but also aid course instructors in while challenges of AIED make up 2.4 % of the sample.
identifying student types early in the course, enabling them to provide
more effective learning support. Furthermore, profiling and prediction 4.2.2.1. System and application design. This stream of research focuses
applications are designed to predict students’ course selections (Kardan on designing AI algorithms, analytical models, or frameworks, as well as
et al., 2013) and curriculum satisfaction (Guo, 2010), offering support intelligent systems for learning support and learning analytics. Some
for curriculum planning and teaching optimization. studies are dedicated to designing and comparing the effectiveness of
various machine learning algorithms for mining and predicting stu
4.2.1.4. Emerging products. In addition to the aforementioned cate dents’ learning behavior, learning styles, and performance (Costa et al.,
gories of AI applications, several emerging products that heavily rely on 2017; Waheed et al., 2020). Other studies revolve around the develop
AI have been discussed and studied. The literature highlights two major ment and testing of AI-based teaching support systems and auxiliary
sub-categories: (1) educational robots and (2) virtual reality (VR) and products, such as cloud collaborative writing support tools (Calvo et al.,
augmented reality (AR) applications. These emerging applications 2011), interactive intelligent physics teaching systems (Myneni et al.,
enhance the interaction between learners and learning environments or 2013), and trainer systems (Liu et al., 2021). The design and validation
resources, primarily within the context of online education (Du et al., of AI models, frameworks, or systems are fundamental to the field of
2023). AIED; consequently, this represents the largest category of research
Educational robots explored in the literature primarily revolve around topics.
chatbots. A chatbot is software that combines AI and natural language
processing to interact with a human user through text or voice (Pérez 4.2.2.2. Adoption and acceptance of AIED. This stream of research
et al., 2020). In the field of education, the use of chatbots varies studies factors that influence the adoption and acceptance of AIED ap
depending on the capabilities of AI technology and the specific plications by instructors as well as by learners. From the instructors’
12
perspective, Wang et al. (2020) identified relative advantage, compati Table 3

bility, perceived trust, and experience as the determinants of teachers’ Guiding Theories.
willingness to adopt ITS. From the learner’s perspective, it is reported Category Theories Count
that system quality, service quality, content quality, technical infra
Education Constructivist learning theory 6
structure, perceived usefulness, and perceived ease of use are key drivers Learning style theory 6
for adoption (Kreijns et al., 2007; Song & Kong, 2017). Cognitive theories of learning 5
The theory of multimedia learning 3
Self-regulated learning theory 2
4.2.2.3. Impacts of AIED. This stream of research investigates the
Bloom’s taxonomy of cognitive domain 1
impact of AIED applications on various aspects of learning, including Kolb’s experiential learning 1
academic performance, affective perception, learning behavior, and Learning-by-doing theory 1
learning ability. Several studies have highlighted the significant positive Reinforcement learning theory 1
impact of AI technology on students’ academic performance. For Scaffolding theory 1
Situational learning theory 1
example, research by (Wang, 2014) found that evaluation-centered e- The text comprehension theory 1
learning systems were effective in promoting students’ academic per The theory of cognitive knowledge 1
formance and correcting misconceptions, especially for students with acquisition
low levels of prior knowledge. Moreover, students have demonstrated Tinto’s theory of student integration 1
Zone of proximal development theory 1
positive attitudes toward the incorporation of AI into education (Özyurt
Mathematics Game theory 2
et al., 2013), expressing increased interest in learning (Liu et al., 2022) The knowledge space theory 2
and improved concentration (Rong et al., 2022) owing to these Fuzzy logic theory 1
applications. Fuzzy set theory 1
Graph theory 1
The mathematical problem-solving 1
4.2.2.4. Challenges of AIED. This research stream delves into the theory of Mayer
ongoing issues and challenges in AIED, emphasizing a balanced, social- Psychology Cognitive load theory 3
technical perspective. Several papers in this category discuss a range of Achievement goal theory 2
challenges related to AIED development. For instance, Maghsudi et al. Activity theory 2
Flow theory 2
(2021) highlighted issues encompassing technical aspects (e.g., content Meta-cognition theory 2
production and recommendation), personal aspects (e.g., lifelong Self-determination theory 2
learning, assessment and evaluation, incentives, and motivations), and The self-efficacy theory 2
social aspects (e.g., learning networks and diversity and fairness of al Appraisal theory 1
Human plausible reasoning theory 1
gorithms) in the development of AIED. Perrotta and Selwyn (2020)
Multiple intelligences theory 1
focused on the challenges related to the application of deep learning in Social learning theory 1
education, including concerns about data quality, the reductionist Sociocultural theory 1
approach of deep learning-based applications, and the integration of The investment theory of creativity 1
educational knowledge in application development. Luckin et al. (2016) The personal construct theory 1
The schema theory 1
advocated the importance of building inter-stakeholder partnerships
The systems model of creativity 1
between AI developers, educators, and researchers to develop effective Theory of reasoned action 1
AIED applications. These papers advocate placing humans at the center Psychometric Item response theory 5
of application development and considering stakeholders’ motivations, Classical testing theories 2
Sociology Discourse theory 1
involvement, and expertise to address the challenges and ensure
Innovation diffusion theory 1
meaningful and impactful AIED solutions. Design science/ engineering Affordance theory 1
design
4.2.3. Research methodology Structure-behavior-function theory 1
Six research methods are coded, including experiments (37.6 % of Communication The dialogue theory for critical thinking 1
the articles in our sample), statistical analysis (including econometric

analysis and machine learning model training; 22.4 %), survey research theory (6), cognitive theories of learning (5), and item response theory
(12 %), descriptive analysis (i.e., with simple illustrative examples; 10.4 (5). These theories are reviewed in the following.
%), qualitative studies (9.6 %), and mixed research methods (8 %). Their Constructivist learning theory views learning as an active process of
occurrence distribution is illustrated in Fig. 9c. Among these methods, knowledge construction, emphasizing the role of learners in actively
experiments are the most frequently used and qualitative research is less shaping their understanding of the world through direct experiences and
used. Nonetheless, qualitative research, such as case studies, plays a reflective practices (Bada, 2015; Piaget, 1964). This theory can guide the
crucial role in theory building. Enriching qualitative research offers understanding of the mechanism of AI-simulated or AI-enabled learning.
valuable opportunities to gain novel insights and develop useful theories For instance, Winkler et al. (2021) applied the constructivist learning
that can significantly contribute to the development and applications of paradigm to investigate whether interactions with scaffolding-based
AIED. smart personal assistant technology empower students to internalize
and independently apply problem-solving strategies. Similarly, Rong
4.2.4. Guiding theories (2022) employs constructivist learning theory and reinforcement
The AIED literature draws upon a diverse set of theories from various learning theory to elucidate the impact of AI and VR technology on
fields to inform its research and development. Table 3 summarizes a students’ levels of concentration and creativity.
total of 45 theories coded in our subsample of 125 articles. These the Learning style theory emphasizes the significance of individuals’
ories originate from the fields of psychology (17), education (15), learning styles and preferences, which shape how they “absorb, process,
mathematics (6), psychometrics (2), sociology (2), design science (2), and retain new information and skills” (Dantas & Cunha, 2020, p. 1).
and communication (1). The AIED field heavily builds upon psychology This theory encompasses a wide array of concepts and models aimed at
and education theories for research development. elucidating the variations in learners’ preference to learn. Some notable
Among these theories, the most applied in the literature are models include Kolb’s learning styles inventory (Kolb, 1976), Felder and
constructivist learning theory (6 articles in our sample), learning style Silverman’s learning style model (Felder, 1988), Honey and Mumford’s
13
learning style model and learning styles questionnaire (Honey & learning, educational technology, intelligent systems, emerging tech
Mumford, 1986), and Fleming’s VAK learning style model (Fleming, nologies, and AI education. A detailed examination uncovers both sim
2001). Learning style theory particularly pertains to the development of ilarities and disparities between these two sets of clusters. The clusters
adaptive learning systems. For instance, Özyurt et al. (2013) utilized the derived from author keywords offer a more detailed perspective than
VAK learning style model to tailor learning materials within an adaptive those identified in the keywords plus network. For instance, the
and intelligent individualized e-learning environment named UZWEB educational technology, learning systems, and emerging technologies
MAT. García et al. (2008) employed Felder and Silverman’s learning clusters within the author keywords network correspond to the design
style model (Felder, 1988) to identify the learning styles of engineering science cluster in the keywords plus network. However, it’s noteworthy
students in online courses. that the AIED user behavior cluster identified in the keywords plus
Cognitive theories of learning elucidate learners’ behaviors through network does not appear in the author keywords network.
the lens of their mental processes. For example, Mednick (1962) theory The subsequent content analysis provides insights into various
of creativity addresses the cognitive aspects of creativity, positing that research elements, including research objects (i.e., AIED applications),
creativity arises from an individual’s ability to synthesize various ele research objectives and topics, research methods, guiding theories, and
ments or ideas in novel and imaginative ways. Piaget’s theory of research contexts (i.e., education stages). The literature explored four
cognitive development (Piaget, 1936, 1971), another prominent primary categories of AIED applications, including adaptive learning
example, posits four stages in the construction of a mental model of the and personalized tutoring, profiling and prediction, intelligent assess
world: sensorimotor, preoperational, concrete operational, and formal ment and management, and emerging products, with adaptive learning
operational stages. Within the field of AIED, cognitive theories often and personalized tutoring being the most extensively studied. Among
serve as foundational frameworks for ITS research and other related the four research topics identified (i.e., system and application design,
studies (Aleven et al., 2009; Waalkens et al., 2013). adoption and acceptance of AIED, impacts of AIED, and challenges of
Item response theory originates from psychometric research and is “a AIED), system and application design emerged as the most frequently
psychometric technique used in the development, evaluation, investigated. Experiments prove to be the predominant research method
improvement, and scoring of multi-item scales” (Toland, 2013, p. 120). among the five categorized research methods, which include mixed
It provides a framework for assessing learners’ abilities, attitudes, or research methods, qualitative studies, experiments, statistical analysis of
unobservable characteristics based on their responses to observable secondary data, survey research, and descriptive studies. Among the 45
items, such as test items (Carlson & Davier, 2013), which are commonly theories identified from the literature, constructivist learning theory,
adopted in AIED applications (Yang & Li, 2018). Such assessments of learning style theory, and cognitive theories of learning are the most
student mastery level in learning is essential for the development of commonly employed theories guiding or supporting the theoretical
systems like personalized education (Maghsudi et al., 2021), intelligent development of research. Higher education and K12 education are the
assessment systems (Csapó & Molnár, 2019), and adaptive learning top two research contexts that receive significant attention in the AIED
systems (Yang et al., 2013). literature.
In summary, AIED research applies a wide range of theories, This research contributes to the AIED literature in several ways. First,
reflecting its interdisciplinary nature. This diversity in theory applica it adds to the body of AIED reviews by offering a systematic exploration
tion mirrors the multifaceted nature of AIED research and its evolving of the literature’s conceptual structure. While prior reviews have
landscape of topics. Furthermore, the literature demonstrates varying explored the general AIED research field or specific domains AIED ap
degrees of theory application. Fig. 9d illustrates that approximately 24 plications (Chassignol et al., 2018; Goksel & Bozkurt, 2019; Guan et al.,
% of articles use one or more primary theories to systematically guide 2020; Hwang et al., 2020; Srinivasan, 2022), few have systematically
the research design (e.g., developing research framework or hypothe examined the conceptual underpinnings of the literature. In this study,
ses), while 20.8 % of articles use theories marginally (e.g., referencing we employ both bibliometric analysis and content analysis to unveil the
theories without much elaboration or in the explanations of the research concepts associated with research elements of central interest to
results). The remaining articles (55.2 %) do not prominently feature scholars. This approach yields a comprehensive understanding of the
theories in their discussions. literature, going beyond the exploration of research objects and AIED
applications to include guiding theories, research topics, and research
4.3. Research context: Education stages methodologies. Their occurrence distributions summarized in Fig. 9
inform the current status of the research design elements.
For use contexts, AIED research spans various stages of the educa Second, the study enriches our comprehension of the AIED research
tional process: higher education, K12 education, preschool education, or landscape and spotlights several areas for research attention. For
general education (without specifying a particular stage). Fig. 9e illus instance, there are clear opportunities for research on the integration of
trates the distribution of these education stages within the 125 papers the latest advancements in AI technologies. While our review encom
we reviewed. Approximately half of the research (45.6 %) centers on passes a broad spectrum of AIED applications, certain latest de
higher education, followed by K12 education (32 %) and general edu velopments, such as generative AI, are absent from our sample articles.
cation (19.2 %). Preschool education garners the least attention (3.2 %). Generative AI benefits from substantial human involvement for superior
This discrepancy may stem from the perception that AI-supported edu results, making it a promising domain for scholarly exploration. Recent
cation is better suited for adult learners who are more autonomous and research in AIED has increasingly emphasized the role of humans in AI
self-regulated, while preschool education requires greater human care application design, with a shift toward paradigms that emphasize
and attention. It also offers significant opportunities for AIED research. learner collaboration and leadership (Andersen et al., 2022; Ouyang &
Jiao, 2021; Xu & Ouyang, 2022). Developing AI systems for learners in
5. Discussion leadership roles remains an ongoing and intricate task (Ouyang & Jiao,
2021), and generative AI offers the potential to assist and engage users,
This research employs bibliometric analysis and content analysis for as either leaders or collaborators, in task completion.
a systematic review of the AIED literature. In the bibliometric analysis, Another area deserving of attention is AI in preschool education,
the keywords co-occurrence analysis reveals two sets of concept clusters which is underrepresented in our sample. AI applications in preschool
explored in the literature. The keywords plus co-occurrence network education may necessitate more engaging design and greater involve
highlights four concept clusters: user behaviors, design science, big data ment of parents. Designing applications that are captivating for both
analytics, and AIED impacts. Conversely, the author keywords co- parents and children presents a promising avenue for exploration. User
occurrence network illustrates concept clusters related to machine emotion is another area offering prosperous opportunities for further
14
research. While emotions are crucial to learning in IT-driven environ studies (Ågerfalk & Karlsson, 2020), qualitative research methods,
ment (Li et al., 2023), existing AIED applications are mostly weak AI which are instrumental for theory generation, are underrepresented.
with limited ability for emotional connections with users. Consequently, However, in the age of AI, a plethora of foundational questions resur
studies on user emotion are lacking in our sample. Future research can face, demanding fresh perspectives and answers. These questions
systematically study user emotions and their roles in the AI-empowered encompass topics like potential disparities in learning styles in the era of
learning environment. For example, flow theory (Csikszentmihalyi, AI, the evolving nature of teaching in the presence of AI, the develop
1990) stands out as a promising theoretical framework to guide future ment of design science in AIED (Goldkuhl & Sjöström, 2021), and the
research on AIED applications for preschool educations. Flow represents exploration of digital experiences for learners and other stakeholders
a state of heightened focus, concentration, and enjoyment. AIED system (Kreps, 2021). Addressing these questions through qualitative studies,
designs that stimulate a state of flow can exert a long-lasting impact on such as case studies and expert panel interviews, can lead to novel
learning. theoretical frameworks and innovative insights, thereby contributing to
Additionally, ethical considerations are another important field of the academic advancement of AIED (Myers, 2009; Yin, 2009).
future study. In our review, AIED ethical studies did not emerge as a Our research highlights the cross-disciplinary nature of the AIED
major research theme in the keyword co-occurrence analysis and con research. Through a granular analysis of the top 15 cited references, we
tent analysis. However, some ethics concerns have been generally dis identified Computer Science and AI, MIS, and Education as three major
cussed in “challenges of AIED”. The increasing integration of AI disciplines foundational to AIED research. Additionally, the summarized
technologies in education has led to growing ethical risks and concerns, theories suggest that education and psychology serve as two primary
including issues related to personal data privacy, algorithm biases, and disciplinary categories. Consequently, fostering increased collaboration
learner and educator autonomy (Akgun & Greenhow, 2022; Boulay, among computer scientists, psychologists, educators, and MIS experts is
2023; Wells, 2023). For example, learning analytics may incentivize a viable path towards higher calibre and more innovative AIED research.
aggressive collection of personal and surveillance data, students may For example, scholars from psychology and education can provide
learn biased knowledge from ChatGPT or other AI models, and teachers theoretical guidance for research design, while computer scientists can
may develop reliance on analytics results to make decisions on students bring their programming expertise to research collaboration. Moreover,
who are in difficulty and require additional assistance (Boulay, 2023). successful interdisciplinary collaboration requires integrative contribu
Biased algorithms can also perpetuate problematic social values. For tions from multiple disciplines and overcoming functional silos (Turner
example, it is recognized that AI-driv & Baker, 2020). MIS scholars can play an important role in facilitating
en language translation tools routinely introduce gender stereotypes such integration, given that MIS is inherently a cross-disciplinary field
when translating from gender-neutral languages, thereby influencing and its scholars often possess experience in integrative research.
language learners’ social perceptions regarding gender (Miller et al.,
2018). Therefore, designing AIED applications that adhere to ethical 6. Conclusions
standards is vital for the well-being of humanity. Future research can
expand to include topics such as the ramifications of ethical risks, users’ This research employs a mixed research method, combining biblio
perception of AIED ethical risks, and how such perceptions affect their metric analysis and content analysis, to uncover and comprehend the
behaviors regarding the adoption and usage of AIED applications. core concepts within the field of AIED. The findings from both ap
Design science researchers can incorporate ethical criteria as one of the proaches converge, providing a comprehensive understanding of AIED
performance metrics in their experiment design—criteria for evaluating concepts. This study contributes to the body of AIED literature reviews
AIED applications encompass not only learning effectiveness and algo by emphasizing the importance of grasping the conceptual structure of
rithm accuracy but also fairness, algorithm transparency, and trust. the field. Additionally, the research suggests several future directions,
The analysis of research methodologies highlights opportunities for including the need to incorporate latest AI technologies, strengthen
developing high-quality research. Notably, mixed research methods, AIED research in the preschool education context, enhance research
which have the potential to significantly enhance result robustness and, quality through mixed methods, prioritize theoretical contributions and
consequently, research quality, are underutilized in the field. Moreover, enhance collaboration among computer scientists, psychologists, edu
over half of the empirical studies in the AIED field do not integrate any cators, and MIS experts.
theoretical frameworks into their research development. Future
research endeavors can strive for more rigorous research designs and
methodologies to further enhance the overall quality of research in the Declaration of competing interest
domain. For instance, design science research in AIED can benefit from
the application of design theories such as affordance theory. These The authors declare that they have no known competing financial
theories can systematically guide scholars and developers in identifying interests or personal relationships that could have appeared to influence
and designing functionalities to achieve important affordances for AIED the work reported in this paper.
(Crompton et al., 2022; Wang et al., 2024).
This research underscores the importance of qualitative studies for Data availability
the advancement of the AIED research field. As the existing research has
primarily focused on AIED application development and empirical Data will be made available on request.
Appendix:. Impactful papers
1a. Top Globally Cited Papers.
15
1 García, P., Amandi, A., Schiaffino, S., & Campo, M. (2007). Evaluating Bayesian networks’ precision for detecting students’ learning styles. Computers & Education, 49(3),
794–808.
2 Dwivedi, Y. K., Hughes, D. L., Coombs, C., Constantiou, I., Duan, Y., Edwards, J. S., Gupta, B., Lal, B., Misra, S., Prashant, P., Raman, R., Rana, N. P., Sharma, S. K., & Upadhyay,
N. (2020). Impact of COVID-19 Pandemic on Information Management Research and Practice: Transforming Education, Work and Life. International Journal of Information
Management 55, Article 102211.
3 Zawacki-Richter, O., Marín, V. I., Bond, M., & 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, 16(39), 1–27.
4 Costa, E. B., Fonseca, B., Santana, M. A., Araújo, F. F. d., & Rego, J. (2017). Evaluating the effectiveness of educational data mining techniques for early prediction of students’
academic failure in introductory programming courses. Computers in Human Behavior, 73, 247–256.
5 Chou, C.-Y., Chan, T.-W., & Lin, C.-J. (2003). Redefining the learning companion: The past, present, and future of educational agents. Computers & Education 40(3), 255–269.
6 Santos, D. P. d., Giese, D., Brodehl, S., Chon, S. H., Staab, W., Kleinert, R., Maintz, D., & Baeßler, B. (2019). Medical students’ attitude towards artificial intelligence: A
multicentre survey. European Radiology 29, 1640–1646
7 Nwana, H. S. (1990). Intelligent tutoring systems: An overview. Artificial Intelligence Review, 4, 251–277.
8 Geirhos, R., Jacobsen, J.-H., Michaelis, C., Zemel, R., Brendel, W., Bethge, M., & Wichmann, F. A. (2020). Shortcut learning in deep neural networks. Nature Machine Intelligence,
2(11), 665–673.
9 Mehmood, R., Alam, F., Albogami, N. N., Katib, I., Albeshri, A., & Altowaijri, S. M. (2017). UTiLearn: A personalised ubiquitous teaching and learning system for smart societies.
IEEE Access, 5, 2615–––2635.
10 Vattam, S. S., Goel, A. K., Rugaber, S., Hmelo-Silver, C. E., Jordan, R., Gray, S., & Sinha, S. (2011). Understanding complex natural systems by articulating structure-behavior-
function models. Educational Technology & Society, 14(1), 66–81.
11 Fryer, L. K., Ainley, M., Thompson, A., Gibson, A., & Sherlock, Z. (2017). Stimulating and sustaining interest in a language course: An experimental comparison of Chatbot and
Human task partners. Computers in Human Behavior, 75, 461–468.
12 Murphy, R. R. (2001). “Competing” for a robotics education. IEEE Robotics & Automation Magazine, 8(2), 44–55.
13 Gong, B., Nugent, J. P., Guest, W., Parker, W., Chang, P. J., Khosa, F., & Nicolaou, S. (2019). Influence of artificial intelligence on Canadian medical students’ preference for
radiology specialty: A national survey study Academic Radiology, 26(4), 566–577.
14 Custers, E. J. F. M. (2015). Thirty years of illness scripts: Theoretical origins and practical applications. Medical Teacher, 37(5), 457–462.
15 Ravyse, W. S., Blignaut, A. S., Leendertz, V., & Woolner, A. (2017). Success factors for serious games to enhance learning: A systematic review. Virtual Reality, 21(1), 31–58.
1b. Top Locally Cited Papers.

794–808.
multicentre survey. European Radiology 29, 1640–1646.
3 Chassignol, M., Khoroshavin, A., Klimova, A., & Bilyatdinova, A. (2018). Artificial intelligence trends in education: A narrative overview. Procedia Computer Science 136, 16–24.
4 Hinojo-Lucena, F.-J., Aznar-Díaz, I., Cáceres-Reche, M.-P., & Romero-Rodríguez, J.-M. (2019). Artificial intelligence in higher education: A bibliometric study on its impact in
the scientific literature. Education Science, 9(51), 1–9.
5 Chen, L., Chen, P., & Lin, Z. (2020). Artificial intelligence in education: A review. IEEE Access, 8, 75264–––75278.
6 Touretzky, D., Gardner-McCune, C., Martin, F., & Seehorn, D. (2019). Envisioning AI for K-12: What should every child know about AI? The Thirty-Third AAAI Conference on
Artificial Intelligence (AAAI-19), Honolulu, Hawaii, USA.
7 Kolachalama, V. B., & Garg, P. S. (2018). Machine learning and medical education. NPJ Digital Medicine, 1(54), 1–3.
8 Masters, K. (2019). Artificial intelligence in medical education. Medical Teacher, 41(9), 976–980.
9 Knox, J. (2020). Artificial intelligence and education in China. Learning, Media and Technology, 45(3), 298–311.
10 Sit, C., Srinivasan, R., Amlani, A., Muthuswamy, K., Azam, A., Monzon, L., & Poon, D. S. (2020). Attitudes and perceptions of UK medical students towards artificial intelligence
and radiology: A multicentre survey. Insights into Imaging, 11(14), 1–6.
794–808.
12 Gadanidis, G. (2017). Artificial intelligence, computational thinking, and mathematics education. The International Journal of Information and Learning Technology, 34(2),
133–139.
13 Luckin, R., & Cukurova, M. (2019). Designing educational technologies in the age of AI: A learning sciences driven approach. British Journal of Educational Technology, 50(6),
2824–2838.
14 Duong, M. T., Rauschecker, A. M., Rudie, J. D., Chen, P.-H., Cook, T. S., Bryan, R. N., & Mohan, S. (2019). Artificial intelligence for precision education in radiology. The British
Journal of Radiology, 92(1103), 20190389.
15 Cope, B., Kalantzis, M., & Searsmith, D. (2021). Artificial intelligence for education: Knowledge and its assessment in AI-enabled learning ecologies. Educational Philosophy and
Theory, 53(12), 1229–1245.
1c. Top 15 Cited References.

1 Zawacki-Richter, O., Marín, V. I., Bond, M., & 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 16(39), 1–27.
2 Popenici, S. A., & 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), 1–13.
3 Luckin, R., & Holmes, W. (2016). Intelligence Unleashed: An argument for AI in Education. UCL Knowledge Lab.
4 Roll, I., & Wylie, R. (2016). Evolution and revolution in artificial intelligence in education International Journal of Artificial Intelligence in Education, 26, 582–599.
5 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, 701–712.
6* Turing, A. M. (1950). Computing machinery and intelligence. Mind, 49, 433–460.
7 Papert, S. (1980). Mindstorms: Children, Computers, and Powerful Ideas. Basic Books, Inc.
multicentre survey. European Radiology, 29, 1640–1646.
9 VanLehn, K. (2011). The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems. Educational Psychologist, 46(4), 197–221.
10* Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340.
11 Wenger, E. (1987). Artificial Intelligence and Tutoring Systems: Computational and Cognitive Approaches to the Communication of Knowledge (1st ed.). Elsevier Science.
12* Russell, S., & Norvig, P. (2002). Artificial Intelligence: A Modern Approach (2nd edition ed.). Prentice Hall.
13 Chassignol, M., Khoroshavin, A., Klimova, A., & Bilyatdinova, A. (2018). Artificial Intelligence trends in education: A narrative overview. Procedia Computer Science 136, 16–24.
14 Carbonell, J. R. (1970). AIinCAI: An artificial-intelligence approach to computer-assisted instruction. IEEE Transactions on Man-Machine Systems, MMS-11(4), 190–202.
15* Felder, R. M. (1988). Learning and teaching styles in engineering education. Journal of Engineering Education, 78(7), 674–681.
Note: Star (*) indicates papers from disciplines other than the research field of AI in education.
16
References artificial intelligence in education. Computers and Education: Artificial Intelligence, 4,

Article 100118. https://doi.org/10.1016/j.caeai.2022.100118
Chou, C.-Y., Chan, T.-W., & Lin, C.-J. (2003). Redefining the learning companion: The
Agaoglu, M. (2016). Predicting instructor performance using data mining techniques in
past, present, and future of educational agents. Computers & Education, 40, 255–269.
higher education. IEEE Access, 4, 2379–2387. https://doi.org/10.1109/
https://doi.org/10.1016/S0360-1315(02)00130-6
ACCESS.2016.2568756
Costa, E. B., Fonseca, B., Santana, M. A., Araújo, F. F. D., & Rego, J. (2017). Evaluating
Ågerfalk, P. J., & Karlsson, F. (2020). Artefactual and empirical contributions in
the effectiveness of educational data mining techniques for early prediction of
information systems research. European Journal of Information Systems, 29, 109–113.
students’ academic failure in introductory programming courses. Computers in
https://doi.org/10.1080/0960085X.2020.1743051
Human Behavior, 73, 247–256. https://doi.org/10.1016/j.chb.2017.01.047
Akgun, S., & Greenhow, C. (2022). Artificial intelligence in education: Addressing ethical
Craig, S. D., Hu, X., Graesser, A. C., Bargagliott, A. E., Sterbinsky, A., Cheney, K. R., &
challenges in K-12 settings. AI and Ethics, 2, 431–440. https://doi.org/10.1007/
Okwumabua, T. (2013). The impact of a technology-based mathematics after-school
s43681-021-00096-7
program using ALEKS on student’s knowledge and behaviors. Computers & Education,
Aleven, V., McLaren, B. M., & Sewall, J. (2009). Scaling up programming by
68, 495–504. https://doi.org/10.1016/j.compedu.2013.06.010
demonstration for intelligent tutoring systems development: An open-access web site
Crompton, H., Jones, M. V., & Burke, D. (2022). Affordances and challenges of artificial
for middle school mathematics learning. IEEE Transactions on Learning Technologies,
intelligence in K-12 education: A systematic review. Journal of Research on
2, 64–78. https://doi.org/10.1109/TLT.2009.22
Technology in Education, 1–21. https://doi.org/10.1080/15391523.2022.2121344
Andersen, R., Mørch, A. I., & Litherland, K. T. (2022). Collaborative learning with block-
Csapó, B., & Molnár, G. (2019). Online diagnostic assessment in support of personalized
based programming: Investigating human-centered artificial intelligence in
teaching and learning: The eDia system. Frontiers in Psychology, 10, 443648. https://
education. Behaviour & Information Technology, 41, 1830–1847. https://doi.org/
doi.org/10.3389/fpsyg.2019.01522
10.1080/0144929X.2022.2083981
Csikszentmihalyi, M. (1990). Flow: The Psychology of Optimal Experience. Harper & Row.
Aria, M., & Cuccurullo, C. (2017). Bibliometrix: An R-tool for comprehensive science
Dantas, L. A., & Cunha, A. (2020). An integrative debate on learning styles and the
mapping analysis. Journal of Informetrics, 11, 959–975. https://doi.org/10.1016/j.
learning process. Social Sciences & Humanities Open, 2, Article 100017. https://doi.
joi.2017.08.007
org/10.1016/j.ssaho.2020.100017
Bada, S. O. (2015). Constructivism learning theory : A paradigm for teaching and
Donthu, N., Kumar, S., Mukherjee, D., Pandey, N., & Lim, W. M. (2021). How to conduct
learning. IOSR Journal of Research & Method in Education, 5, 66–70. https://api.
a bibliometric analysis: An overview and guidelines. Journal of Business Research,
semanticscholar.org/CorpusID:37780480.
133, 285–296. https://doi.org/10.1016/j.jbusres.2021.04.070
Belpaeme, T., & Tanaka, F. (2022). Social Robots as Educators. Retrieved April 10, 2024
Du, Z., Wang, F., & Wang, S. (2022). Enhancing learner participation in online discussion
from https://www.oecd-ilibrary.org/sites/1c3b1d56-en/index.html?
forums in massive open online courses: The role of mandatory participation. Frontiers
itemId=/content/component/1c3b1d56-en#section-d1e17138-3fa1249ab7.
in Psychology, 13, 819640. https://doi.org/10.3389/fpsyg.2022.819640
Bicknell, K., Brust, C., & Settles, B. (2023). How Duolingo’s AI Learns What You Need to
Du, Z., Wang, F., Wang, S., & Xiao, X. (2023). Online listening responses and e-learning
Learn. Retrieved April 10, 2024 from https://spectrum.ieee.org/duolingo.
performance. Information Technology & People, 36(4), 1509–1532. https://doi.org/
Boulay, B. D. (2023). Artificial Intelligence in Education and Ethics. In O. Zawacki-
10.1108/ITP-09-2021-0687
Richter, & I. Jung (Eds.), Handbook of Open, Distance and Digital Education (pp.
Duong, M. T., Rauschecker, A. M., Rudie, J. D., Chen, P.-H., Cook, T. S., Bryan, R. N., &
93–108). Springer Singapore. https://doi.org/10.1007/978-981-19-2080-6_6.
Mohan, S. (2019). Artificial intelligence for precision education in radiology. The
Businessolution.org. AI in Education Statistics 2023 [Adoption, Benefits, Challenges]
British Journal of Radiology, 92(1103), 20190389. https://doi.org/10.1259/
2023 Kingdom%20use%20AI%20for%20their%20projects%20and%20homework
bjr.20190389
Retrieved April 10, 2024 from.
Dwivedi, Y. K., Hughes, D. L., Coombs, C., Constantiou, I., Duan, Y., Edwards, J. S.,
Calvo, R. A., O’Rourke, S. T., Jones, J., Yacef, K., & Reimann, P. (2011). Collaborative
Gupta, B., Lal, B., Misra, S., Prashant, P., Raman, R., Rana, N. P., Sharma, S. K., &
writing support tools on the cloud. IEEE Transactions on Learning Technologies, 4,
Upadhyay, N. (2020). Impact of COVID-19 pandemic on information management
88–97. https://doi.org/10.1109/tlt.2010.43
research and practice: Transforming education, work and life. International Journal of
Carlson, J., & Davier, M. V. (2013). Item response theory. In Advancing Human Assessment
Information Management, 55, Article 102211. https://doi.org/10.1016/j.
(pp. 133–178). Princeton: Educational Testing Service. https://doi.org/10.1007/
ijinfomgt.2020.102211
978-3-319-58689-2_5.
Felder, R. M. (1988). Learning and teaching styles in engineering education. Journal of
Casamayor, A., Amandi, A., & Campo, M. (2009). Intelligent assistance for teachers in
Engineering Education, 78, 674–681. https://api.semanticscholar.org/CorpusID:
collaborative e-learning environments. Computers & Education, 53, 1147–1154.
140475379.
https://doi.org/10.1016/j.compedu.2009.05.025
Fleming, N. D. (2001). Teaching and Learning Styles: VARK Strategies. Fleming: N.D.
Celik, I., Dindar, M., Muukkonen, H., & Järvelä, S. (2022). The promises and challenges
Fryer, L. K., Ainley, M., Thompson, A., Gibson, A., & Sherlock, Z. (2017). Stimulating and
of artificial intelligence for teachers: A systematic review of research. TechTrends, 66,
sustaining interest in a language course: An experimental comparison of chatbot and
616–630. https://doi.org/10.1007/s11528-022-00715-y
human task partners. Computers in Human Behavior, 75, 461–468. https://doi.org/
Charitopoulos, A., Rangoussi, M., & Koulouriotis, D. (2020). On the use of soft computing
10.1016/j.chb.2017.05.045
methods in educational data mining and learning analytics research: A review of
Fu, S., Gu, H., & Yang, B. (2020). The affordances of AI-enabled automatic scoring
years 2010–2018. International Journal of Artificial Intelligence in Education, 30,
applications on learners’ continuous learning intention: An empirical study in China.
371–430. https://doi.org/10.1007/s40593-020-00200-8
British Journal of Educational Technology, 51, 1674–1692. https://doi.org/10.1111/
Chassignol, M., Khoroshavin, A., Klimova, A., & Bilyatdinova, A. (2018). Artificial
bjet.12995
intelligence trends in education: A narrative overview. Procedia Computer Science,
Gadanidis, G. (2017). Artificial intelligence, computational thinking, and mathematics
136, 16–24. https://doi.org/10.1016/j.procs.2018.08.233
education. The International Journal of Information and Learning Technology, 34,
Chen, B., Hwang, G.-H., & Wang, S.-H. (2021). Gender differences in cognitive load when
133–139. https://doi.org/10.1108/IJILT-09-2016-0048
applying game-based learning with intelligent robots. Educational Technology &
García, P., Amandi, A., Schiaffino, S., & Campo, M. (2007). Evaluating Bayesian
Society, 24, 102–115.
networks’ precision for detecting students’ learning styles. Computers & Education,
Chen, C.-M., & Tsao, H.-W. (2021). An instant perspective comparison system to facilitate
49, 794–808. https://doi.org/10.1016/j.compedu.2005.11.017
learners’ discussion effectiveness in an online discussion process. Computers &
García, P., Schiaffino, S., & Amandi, A. (2008). An enhanced Bayesian model to detect
Education, 164, Article 104037. https://doi.org/10.1016/j.compedu.2020.104037
students’ learning styles in Web-based courses. Journal of Computer Assisted Learning,
Chen, H.-M., Yu, C., & Chang, C.-S. (2007). E-Homebook system: A web-based interactive
24, 305–315. https://doi.org/10.1111/j.1365-2729.2007.00262.x
education interface. Computers & Education, 49, 160–175. https://doi.org/10.1016/j.
Goksel, N., & Bozkurt, A. (2019). Artificial intelligence in education: Current insights and
compedu.2005.05.003
future perspectives In S. Sisman-Ugur & G. Kurubacak (Eds.), Handbook of Research
Chen, L., Chen, P., & Lin, Z. (2020). Artificial intelligence in education: A review. IEEE
on Learning in the Age of Transhumanism (pp. 224-236). IGI Global. doi: 10.4018/978-
Access, 8, 75264–75278. https://doi.org/10.1109/ACCESS.2020.2988510
1-5225-8431-5.ch014.
Chen, S.-Y., Lin, P.-H., & Chien, W.-C. (2022). Children’s digital art ability training
Goldkuhl, G., & Sjöström, J. (2021). Design science theorizing: The contribution of
system based on AI-assisted learning: A case study of drawing color perception.
practical theory. In N. R. Hassan, & L. P. Willcocks (Eds.), Advancing Information
Frontiers in Psychology, 13, Article 823078. https://doi.org/10.3389/
Systems Theories (pp. 239–273). Palgrave Macmillan. https://doi.org/10.1007/978-
fpsyg.2022.823078
3-030-64884-8_7.
Chen, X., Xie, H., & Hwang, G.-J. (2020). A multi-perspective study on Artificial
González-Calatayud, V., Prendes-Espinosa, P., & Roig-Vila, R. (2021). Artificial
Intelligence in Education: Grants, conferences, journals, software tools, institutions,
intelligence for student assessment: A systematic review. Applied Sciences, 11(12),
and researchers. Computers and Education: Artificial Intelligence, 1, Article 100005.
5467. https://doi.org/10.3390/app11125467
https://doi.org/10.1016/j.caeai.2020.100005
GrandViewResearch. (2021). AI In education market size, share & trends analysis report.
Chen, X., Xie, H., Zou, D., & Hwang, G.-J. (2020). Application and theory gaps during the
Grand View Research. Retrieved January 14, 2023 from https://www.grandviewre
rise of Artificial Intelligence in Education. Computers and Education: Artificial
search.com/industry-analysis/artificial-intelligence-ai-education-market-report.
Intelligence, 1, Article 100002. https://doi.org/10.1016/j.caeai.2020.100002
Guan, C., Mou, J., & Jiang, Z. (2020). Artificial intelligence innovation in education: A
Cheng, Y., Wu, G., Zou, H., Luo, P., & Cai, Z. (2022). A knowledge query network model
twenty-year data-driven historical analysis. International Journal of Innovation
based on rasch model embedding for personalized online learning. Frontiers in
Studies, 4, 134–147. https://doi.org/10.1016/j.ijis.2020.09.001
Psychology, 13, Article 846621. https://doi.org/10.3389/fpsyg.2022.846621
Gülcü, A. (2009). ZOSMAT: Web-based intelligent tutoring system for teaching–learning
Chiu, C.-K., & Tseng, J. C. R. (2021). A Bayesian classification network-based learning
process. Expert Systems with Applications, 36, 1229–1239. https://doi.org/10.1016/j.
status management system in an intelligent classroom. Educational Technology &
eswa.2007.11.064
Society, 24, 274–285. https://doi.org/10.1016/S0360-1315(01)00073-2
Guo, W. W. (2010). Incorporating statistical and neural network approaches for student
Chiu, T. K. F., Xia, Q., Zhou, X., Chai, C. S., & Cheng, M. (2023). Systematic literature
course satisfaction analysis and prediction. Expert Systems with Applications, 37,
review on opportunities, challenges, and future research recommendations of
3358–3365. https://doi.org/10.1016/j.eswa.2009.10.014
17
Hinojo-Lucena, F.-J., Aznar-Díaz, I., Cáceres-Reche, M.-P., & Romero-Rodríguez, J.-M. Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. B. (2016). Intelligence unleashed: An
(2019). Artificial intelligence in higher education: A bibliometric study on its impact argument for AI in education. Pearson Education.
in the scientific literature. Education Science, 9, 1–9. https://doi.org/10.3390/ Maghsudi, S., Lan, A., Xu, J., & van Der Schaar, M. (2021). Personalized education in the
educsci9010051 artificial intelligence era: What to expect next. IEEE Signal Processing Magazine, 38,
Honey, P., & Mumford, A. (1986). Using our learning styles. Peter Honey. 37–50. https://doi.org/10.1109/Msp.2021.3055032
Hooshyar, D., Ahmad, R. B., Yousefi, M., Yusop, F. D., & Horng, S. J. (2015). A flowchart- Martinez-Maldonado, R., Clayphan, A., Yacef, K., & Kay, J. (2015). MTfeedback:
based intelligent tutoring system for improving problem-solving skills of novice Providing notifications to enhance teacher awareness of small group work in the
programmers. Journal of Computer Assisted Learning, 31, 345–361. https://doi.org/ classroom. IEEE Transactions on Learning Technologies, 8, 187–200. https://doi.org/
10.1111/jcal.12099 10.1109/Tlt.2014.2365027
Hsieh, T.-C., Lee, M. C., & Su, C.-Y. (2013). Designing and implementing a personalized Mednick, S. (1962). The associative basis of the creative process. Psychological Review,
remedial learning system for enhancing the programming learning. Educational 69, 220–232. https://doi.org/10.1037/h0048850
Technology & Society, 16, 32–46. Miller, F. A., Katz, J. H., & Gans, R. (2018). AI+I=AI2: The OD imperative to add
Hsu, C.-K., Hwang, G.-J., & Chang, C.-K. (2010). Development of a reading material inclusion to the algorithms of artificial intelligence. OD Practitioner, 50, 6–12.
recommendation system based on a knowledge engineering approach. Computers & Murphy, R. R. (2001). “Competing” for a robotics education. IEEE Robotics & Automation
Education, 55, 76–83. https://doi.org/10.1016/j.compedu.2009.12.004 Magazine, 8, 44–55. https://doi.org/10.1109/100.932757
Hwang, G.-J., & Tu, Y.-F. (2021). Roles and research trends of artificial intelligence in Myers, M. D. (2009). Qualitative Research in Business & Management. Sage.
mathematics education: A bibliometric mapping analysis and systematic review. Publications.
Mathematics, 9, 594. https://doi.org/10.3390/math9060584 Myneni, L. S., Narayanan, N. H., Rebello, S., Rouinfar, A., & Pumtambekar, S. (2013). An
Hwang, G.-J., Xie, H., Wah, B. W., & Gašević, D. (2020). Vision, challenges, roles and interactive and intelligent learning system for physics education. IEEE Transactions
research issues of Artificial Intelligence in Education. Computers & Education: on Learning Technologies, 6, 228–239. https://doi.org/10.1109/tlt.2013.26
Artificial Intelligence, 1, Article 100001. https://doi.org/10.1016/j. Nabizadeh, A. H., Goncalves, D., Gama, S., & Jorge, J. (2022). Early prediction of
caeai.2020.100001 students’ final grades in a gamified course. IEEE Transactions on Learning
iFlyTek. (2024). From holding the “red pen” to holding the “mouse”, the technological Technologies, 15, 311–325. https://doi.org/10.1109/tlt.2022.3170494
revolution behind the college entrance examination marking. Retrieved April 04, Niu, P. (2022). An artificial intelligence method for comprehensive evaluation of
2024 from https://edu.iflytek.com/solution/examination. preschool education quality. Frontiers in Psychology, 13, Article 955870. https://doi.
Kardan, A. A., Sadeghi, H., Ghidary, S. S., & Sani, M. R. F. (2013). Prediction of student org/10.3389/fpsyg.2022.955870
course selection in online higher education institutes using neural network. Nwana, H. S. (1990). Intelligent tutoring systems: An overview. Artificial Intelligence
Computers & Education, 65, 1–11. https://doi.org/10.1016/j.compedu.2013.01.015 Review, 4, 251–277. https://doi.org/10.1007/BF00168958
Kim, J., Kim, K., & Kim, W. (2022). Impact of immersive virtual reality content using Okonkwo, C. W., & Ade-Ibijola, A. (2021). Chatbots applications in education: A
360-degree videos in undergraduate education. IEEE Transactions on Learning systematic review. Computers and Education: Artificial Intelligence, 2, Article 100033.
Technologies, 15, 137–149. https://doi.org/10.1109/Tlt.2022.3157250 https://doi.org/10.1016/j.caeai.2021.100033
Kolb, D. A. (1976). Learning style inventory. McBer & Company. Ouyang, F., & Jiao, P. (2021). Artificial intelligence in education: The three paradigms.
Kong, S. C., Yeung, Y. Y., & Wu, X. Q. (2009). An experience of teaching for learning by Computers and Education: Artificial Intelligence, 2, Article 100020. https://doi.org/
observation: Remote-controlled experiments on electrical circuits. Computers & 10.1016/j.caeai.2021.100020
Education, 52, 702–717. https://doi.org/10.1016/j.compedu.2008.11.011 Özyurt, Ö., Özyurt, H., Baki, A., & Güven, B. (2013). Integration into mathematics
Kreijns, K., Kirschner, P. A., Jochems, W., & Buuren, H. V. (2007). Measuring perceived classrooms of an adaptive and intelligent individualized e-learning environment:
sociability of computer-supported collaborative learning environments. Computer & Implementation and evaluation of UZWEBMAT. Computers in Human Behavior, 29,
Education, 49, 176–192. https://doi.org/10.1016/j.compedu.2005.05.004 726–738. https://doi.org/10.1016/j.chb.2012.11.013
Kreps, D. (2021). Theorizing Digital Experience: Four Aspects of the Infomaterial. In N. Peredo, R., Canales, A., Menchaca, A., & Peredo, I. (2011). Intelligent Web-based
R. Hassan & L. P. Willcocks (Eds.), Advancing Information Systems Theories (pp. education system for adaptive learning. Expert Systems with Applications, 38,
219–237). 14690–14702. https://doi.org/10.1016/j.eswa.2011.05.013
Kulik, J. A., & Fletcher, J. D. (2016). Effectiveness of intelligent tutoring systems: A meta- Pérez, J. Q., Daradoumis, T., & Puig, J. M. M. (2020). Rediscovering the use of chatbots
analytic review. Review of Educational Research, 86, 42–78. https://doi.org/10.3102/ in education: A systematic literature review. Computer Applications in Engineering
0034654315581420 Education, 28, 1549–1565. https://doi.org/10.1002/cae.22326
Leh, J. (2022). AI in LMS: 10 must-see innovations for learning professionals. Retrieved Perrotta, C., & Selwyn, N. (2020). Deep learning goes to school: Toward a relational
April 05, 2024 from https://talentedlearning.com/ai-in-lms-innovations-learning-p understanding of AI in education. Learning, Media and Technology, 45, 251–269.
rofessionals-must-see/. https://doi.org/10.1080/17439884.2020.1686017
Li, Y., Chang, M., Zhao, H., Jiang, C., & Xu, S. (2023). Anxiety only makes it worse: Piaget, J. (1936). Origins of intelligence in the child. Routledge & Kegan Paul.
Exploring the impact mechanisms of app-based learning on performance progress. Piaget, J. (1964). Cognitive development in children: Development and learning. Journal
Journal of Computer Assisted Learning, 39, 63–76. https://doi.org/10.1111/ of Research in Science Teaching, 2, 176–186. https://doi.org/10.1002/
jcal.12727 tea.3660020306
Lin, C.-C., Guo, K., & Lin, Y.-C. (2016). A simple and effective remedial learning system Piaget, J. (1971). The Theory of Stages in Cognitive Development. In D. Green,
with a fuzzy expert system. Journal of Computer Assisted Learning, 32, 647–662. M. P. Ford, & G. B. Flamer (Eds.), Measurement and Piaget (pp. 1–11). McGraw-Hill.
https://doi.org/10.1111/jcal.12160 Poitras, E. G., Lajoie, S. P., Doleck, T., & Jarrell, A. (2016). Subgroup discovery with user
Lin, C. F., Yeh, Y.-C., Hung, Y. H., & Chang, R. I. (2013). Data mining for providing a interaction data: An empirically guided approach to improving intelligent tutoring
personalized learning path in creativity: An application of decision trees. Computers systems. Educational Technology & Society, 19, 204–214.
& Education, 68, 199–210. https://doi.org/10.1016/j.compedu.2013.05.009 Popenici, S. A., & Kerr, S. (2017). Exploring the impact of artificial intelligence on
Lin, F.-R., Hsieh, L.-S., & Chuang, F.-T. (2009). Discovering genres of online discussion teaching and learning in higher education. Research and Practice in Technology
threads via text mining. Computers & Education, 52, 481–495. https://doi.org/ Enhanced Learning, 12, 1–13. https://doi.org/10.1186/S41039-017-0062-8
10.1016/j.compedu.2008.10.005 Rainer, K., Prince, B., Splettstoesser-Hogeterp, I., & SanchezRodriguez, C. (2016).
Lin, Y.-S., Chen, S.-Y., Tsai, C.-W., & Lai, Y.-H. (2021). Exploring computational thinking Introduction to Information Systems ((4th ed.).). John Wiley & Sons Inc.
skills training through augmented reality and AIoT learning. Frontiers in Psychology, Rau, M. A., Michaelis, J. E., & Fay, N. (2015). Connection making between multiple
12, Article 640115. https://doi.org/10.3389/fpsyg.2021.640115 graphical representations: A multi-methods approach for domain-specific grounding
Lin, Y.-S., & Lai, Y.-H. (2021). Analysis of AI precision education strategy for small of an intelligent tutoring system for chemistry. Computers & Education, 82, 460–485.
private online courses. Frontiers in Psychology, 12, Article 749629. https://doi.org/ https://doi.org/10.1016/j.compedu.2014.12.009
10.3389/fpsyg.2021.749629 Rong, Q., Lian, Q., & Tang, T. (2022). Research on the influence of AI and VR technology
Liu, C.-C., Liao, M.-G., Chang, C.-H., & Lin, H.-M. (2022). An analysis of children’ for students’ concentration and creativity. Frontiers in Psychology, 13, Article 767689.
interaction with an AI chatbot and its impact on their interest in reading. Computers https://doi.org/10.3389/fpsyg.2022.767689
& Education, 189, Article 104576. https://doi.org/10.1016/j.compedu.2022.104576 Roser, M. (2022). The brief history of artificial intelligence: The world has changed fast –
Liu, J., Wang, L., & Zhou, H. (2021). The application of human-computer interaction what might be next? Our World in Data. Retrieved August 25, 2023 from https://
technology fused with artificial intelligence in sports moving target detection ourworldindata.org/brief-history-of-ai.
education for college athlete. Frontiers in Psychology, 12, Article 677590. https://doi. Şahin, M., & Yurdugül, H. (2022). Learners’ needs in online learning environments and
org/10.3389/fpsyg.2021.677590 third generation learning management systems (LMS 3.0). Technology Knowledge and
Liu, Z., Kong, X., Liu, S., Yang, Z., & Zhang, C. (2022). Automated detection of emotional Learning, 27, 33–48. https://doi.org/10.1007/s10758-020-09479-x
and cognitive engagement in MOOC discussions to predict learning achievement. Schiaffino, S., Garcia, P., & Amandi, A. (2008). eTeacher: Providing personalized
Computers & Education, 181, Article 104461. https://doi.org/10.1016/j. assistance to e-learning students. Computers & Education, 51, 1744–1754. https://
compedu.2022.104461 doi.org/10.1016/j.compedu.2008.05.008
Lo, J.-J., Chan, Y.-C., & Yeh, S.-W. (2012). Designing an adaptive web-based learning Sharkey, A. J. C. (2016). Should we welcome robot teachers? Ethics and Informatoin
system based on students’ cognitive styles identified online. Computers & Education, Technology, 18, 283–297. https://doi.org/10.1007/s10676-016-9387-z
58, 209–222. https://doi.org/10.1016/j.compedu.2011.08.018 Sit, C., Srinivasan, R., Amlani, A., Muthuswamy, K., Azam, A., Monzon, L., & Poon, D. S.
López-Zambrano, J., Lara Torralbo, J. A., & Romero Morales, C. (2021). Early prediction (2020). Attitudes and perceptions of UK medical students towards artificial
of student learning performance through data mining: A systematic review. intelligence and radiology: A multicentre survey. Insights into Imaging, 11, 1–6.
Psicothema, 33, 456–465. https://doi.org/10.7334/psicothema2021.62 https://doi.org/10.1186/s13244-019-0830-7
Luan, H., & Chin-Chung, T. (2021). A Review of using machine learning approaches for Snyder, H. (2019). Literature review as a research methodology: An overview and
precision education. Journal of Educational Technology & Society, 24, 250–266. guidelines. Journal of Business Research, 104, 333–339. https://doi.org/10.1016/j.
jbusres.2019.07.039
18
Song, Y., & Kong, S.-C. (2017). Investigating students’ acceptance of a statistics learning Wang, T., & Park, J. (2021). Design and implementation of intelligent sports training
platform using technology acceptance model. Journal of Educational Computing system for college students’ mental health education. Frontiers in Psychology, 12,
Research, 55, 865–897. https://doi.org/10.1177/0735633116688320 Article 634978. https://doi.org/10.3389/fpsyg.2021.634978
Srinivasan, V. (2022). AI & learning: A preferred future. Computers and Education: Wells, R.E. (2023). Strong AI vs. weak AI: What’s the difference? Strong AI can do anything a
Artificial Intelligence, 3, Article 100062. https://doi.org/10.1016/j. human can do, while weak AI is limited to a specific task. LifeWire. Retrieved August 22,
caeai.2022.100062 2023 from https://www.lifewire.com/strong-ai-vs-weak-ai-7508012.
Tang, T., Li, P., & Tang, Q. (2022). New strategies and practices of design education Wilson, J., Ahrendt, C., Fudge, E. A., Raiche, A., Beard, G., & MacArthur, C. (2021).
under the background of artificial intelligence technology: Online animation design Elementary teachers’ perceptions of automated feedback and automated scoring:
studio. Frontiers in Psychology, 13, Article 767295. https://doi.org/10.3389/ Transforming the teaching and learning of writing using automated writing
fpsyg.2022.767295 evaluation. Computers & Education, 168, Article 104208. https://doi.org/10.1016/j.
Tasci, T., Parlak, Z., Kibar, A., Tasbasi, N., & Cebeci, H. (2014). A novel agent-supported compedu.2021.104208
academic online examination system. Educational Technology & Society, 17, 154–168. Winkler, R., Söllner, M., & Leimeister, J. M. (2021). Enhancing problem-solving skills
Toland, M. D. (2013). Practical guide to conducting an item response theory analysis. The with smart personal assistant technology. Computers & Education, 165, Article
Journal of Early Adolescence, 34, 120–151. https://doi.org/10.1177/ 104148. https://doi.org/10.1016/j.compedu.2021.104148
0272431613511332 Xu, D., & Wang, H. (2006). Intelligent agent supported personalization for virtual
Turner, J. R., & Baker, R. (2020). Collaborative research: Techniques for conducting learning environments. Decision Support Systems, 42, 825–843. https://doi.org/
collaborative research from the science of team science (SciTS). Advances in 10.1016/j.dss.2005.05.033
Developing Human Resources, 22, 72–86. https://doi.org/10.1177/ Xu, W., & Ouyang, F. (2022). The application of AI technologies in STEM education: A
1523422319886300 systematic review from 2011 to 2021. International Journal of STEM Education, 9,
Vázquez-Cano, E., Mengual-Andrés, S., & López-Meneses, E. (2021). Chatbot to improve 1–20. https://doi.org/10.1186/s40594-022-00377-5
learning punctuation in Spanish and to enhance open and flexible learning Yaghmaie, M., & Bahreininejad, A. (2011). A context-aware adaptive learning system
environments. International Journal of Educational Technology in Higher Education, 18, using agents. Expert Systems with Applications, 38, 3280–3286. https://doi.org/
1–20. https://doi.org/10.1186/s41239-021-00269-8 10.1016/j.eswa.2010.08.113
Waalkens, M., Aleven, V., & Taatgen, N. (2013). Does supporting multiple student Yang, F., & Li, F. W. B. (2018). Study on student performance estimation, student
strategies lead to greater learning and motivation? Investigating a source of progress analysis, and student potential prediction based on data mining. Computers
complexity in the architecture of intelligent tutoring systems. Computers & Education, & Education, 123, 97–108. https://doi.org/10.1016/j.compedu.2018.04.006
60, 159–171. https://doi.org/10.1016/j.compedu.2012.07.016 Yang, Y., Leung, H., Yue, L., & Deng, L. (2013). Generating a two-phase lesson for
Waheed, H., Hassan, S.-U., Aljohani, N. R., Hardman, J., Alelyani, S., & Nawaz, R. guiding beginners to learn basic dance movements. Computers & Education, 61, 1–20.
(2020). Predicting academic performance of students from VLE big data using deep https://doi.org/10.1016/j.compedu.2012.09.006
learning models. Computers in Human Behavior, 104, Article 106189. https://doi.org/ Yin, R. K. (2009). Case study research: Design and methods (Vol. 5). SAGE Publications Inc.
10.1016/j.chb.2019.106189 Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019). Systematic review
Wang, N., Wang, X., & Su, Y.-S. (2024). Critical analysis of the technological affordances, of research on artificial intelligence applications in higher education – where are the
challenges and future directions of Generative AI in education: A systematic review. educators? International Journal of Educational Technology in Higher Education, 16, 39.
Asia Pacific Journal of Education, 44, 139–155. https://doi.org/10.1080/ https://doi.org/10.1186/s41239-019-0171-0
02188791.2024.2305156 Zhang, H., & Han, X. (2021). Influence of vocalized reading practice on English learning
Wang, S., Yu, H., Hu, X., & Li, J. (2020). Participant or spectator? Comprehending the and psychological problems of middle school students. Frontiers in Psychology, 12,
willingness of faculty to use intelligent tutoring systems in the artificial intelligence Article 709023. https://doi.org/10.3389/fpsyg.2021.709023
era. British Journal of Educational Technology, 51, 1657–1673. https://doi.org/ Zhang, J., Zhang, X., Jiang, S., Ordonez de Pablos, P., & Sun, Y. (2018). Mapping the
10.1111/bjet.12998 study of learning analytics in higher education. Behaviour & Information Technology,
Wang, T.-H. (2014). Developing an assessment-centered e-Learning system for improving 37, 1142–1155. https://doi.org/10.1080/0144929X.2018.1529198
student learning effectiveness. Computers & Education, 73, 189–203. https://doi.org/
10.1016/j.compedu.2013.12.002
19

1 s2.0 S0957417424010339 Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S0957417424010339 Main

Uploaded by

Document Information

Original Description:

Original Title

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

1 s2.0 S0957417424010339 Main

Uploaded by

Copyright:

Available Formats

Expert Systems With Applications 252 (2024) 124167

Contents lists available at ScienceDirect

Expert Systems With Applications

Artificial intelligence in education: A systematic literature review

1. Introduction as Absorb LMS and Docebo, deliver multiple AI capabilities to support

Fig. 2. Important Journal .

Fig. 3. Most Cited Papers and References.

Fig. 4. Average Article Citation Per Year (AACPY).

Fig. 5. Most Frequent Keywords.

Fig. 6. Keyword Co-occurrence Networks.

Fig. 7. Conceptual Mapping between Co-occurrence Networks and Systematic Review.

Fig. 8. Coding Schema.

Fig. 9. Distribution of Coding Categories.

Fig. 10. AIED Applications.

perspective, Wang et al. (2020) identified relative advantage, compati­ Table 3

the articles in our sample), statistical analysis (including econometric

Appendix:. Impactful papers

1a. Top Globally Cited Papers.

1b. Top Locally Cited Papers.

1c. Top 15 Cited References.

References artificial intelligence in education. Computers and Education: Artificial Intelligence, 4,

You might also like

perspective, Wang et al. (2020) identified relative advantage, compati Table 3