REVIEW

published: 19 October 2020

doi: 10.3389/fpsyg.2020.580820

Challenges and Future Directions of

Big Data and Artificial Intelligence in
Education
Hui Luan 1 , Peter Geczy 2 , Hollis Lai 3 , Janice Gobert 4,5 , Stephen J. H. Yang 6 ,
Hiroaki Ogata 7 , Jacky Baltes 8 , Rodrigo Guerra 9 , Ping Li 10 and Chin-Chung Tsai 1,11*
1
Institute for Research Excellence in Learning Sciences, National Taiwan Normal University, Taipei, Taiwan, 2 National
Institute of Advanced Industrial Science and Technology, Tsukuba, Japan, 3 School of Dentistry, Faculty of Medicine &
Dentistry, University of Alberta, Edmonton, AB, Canada, 4 Graduate School of Education, Rutgers – The State University
of New Jersey, New Brunswick, NJ, United States, 5 Apprendis, LLC, Berlin, MA, United States, 6 Department of Computer
Science and Information Engineering, College of Electrical Engineering and Computer Science, National Central University,
Taoyuan City, Taiwan, 7 Graduate School of Informatics, Kyoto University, Kyoto, Japan, 8 Department of Electrical
Engineering, College of Technology and Engineering, National Taiwan Normal University, Taipei, Taiwan, 9 Centro
de Tecnologia, Universidade Federal de Santa Maria, Santa Maria, Brazil, 10 Department of Chinese and Bilingual Studies,
Faculty of Humanities, The Hong Kong Polytechnic University, Kowloon, Hong Kong, 11 Program of Learning Sciences,
National Taiwan Normal University, Taipei, Taiwan

We discuss the new challenges and directions facing the use of big data and
Edited by: artificial intelligence (AI) in education research, policy-making, and industry. In recent
Ronnel B. King, years, applications of big data and AI in education have made significant headways.
University of Macau, China
This highlights a novel trend in leading-edge educational research. The convenience
Reviewed by:
Hannele Niemi, and embeddedness of data collection within educational technologies, paired with
University of Helsinki, Finland computational techniques have made the analyses of big data a reality. We are moving
Ze Wang,
University of Missouri, United States
beyond proof-of-concept demonstrations and applications of techniques, and are
*Correspondence:
beginning to see substantial adoption in many areas of education. The key research
Chin-Chung Tsai trends in the domains of big data and AI are associated with assessment, individualized
[email protected] learning, and precision education. Model-driven data analytics approaches will grow
Specialty section:
quickly to guide the development, interpretation, and validation of the algorithms.
This article was submitted to However, conclusions from educational analytics should, of course, be applied with
Educational Psychology,
caution. At the education policy level, the government should be devoted to supporting
a section of the journal
Frontiers in Psychology lifelong learning, offering teacher education programs, and protecting personal data.
Received: 07 July 2020 With regard to the education industry, reciprocal and mutually beneficial relationships
Accepted: 22 September 2020 should be developed in order to enhance academia-industry collaboration. Furthermore,
Published: 19 October 2020
it is important to make sure that technologies are guided by relevant theoretical
Citation:
Luan H, Geczy P, Lai H, Gobert J,
frameworks and are empirically tested. Lastly, in this paper we advocate an in-depth
Yang SJH, Ogata H, Baltes J, dialog between supporters of “cold” technology and “warm” humanity so that it can
Guerra R, Li P and Tsai C-C (2020)
lead to greater understanding among teachers and students about how technology,
Challenges and Future Directions
of Big Data and Artificial Intelligence and specifically, the big data explosion and AI revolution can bring new opportunities
in Education. (and challenges) that can be best leveraged for pedagogical practices and learning.
Front. Psychol. 11:580820.
doi: 10.3389/fpsyg.2020.580820 Keywords: big data, artificial intelligence, education, learning, teaching

Frontiers in Psychology | www.frontiersin.org 1 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

INTRODUCTION positioned at the intersection of big data and AI, and

are capable of improving the services and productivity of
The purpose of this position paper is to present current education, as well as many other fields including commerce,
status, opportunities, and challenges of big data and AI in science, and government.
education. The work has originated from the opinions and Regarding education, our main area of interest here,
panel discussion minutes of an international conference on big the application of AI technologies can be traced back to
data and AI in education (The International Learning Sciences approximately 50 years ago. The first Intelligent Tutoring System
Forum, 2019), where prominent researchers and experts from “SCHOLAR” was designed to support geography learning, and
different disciplines such as education, psychology, data science, was capable of generating interactive responses to student
AI, and cognitive neuroscience, etc., exchanged their knowledge statements (Carbonell, 1970). While the amount of data was
and ideas. This article is organized as follows: we start with an relatively small at that time, it was comparable to the amount of
overview of recent progress of big data and AI in education. data collected in other traditional educational and psychological
Then we present the major challenges and emerging trends. studies. Research on AI in education over the past few decades has
Finally, based on our discussions of big data and AI in education, been dedicated to advancing intelligent computing technologies
conclusion and future scope are suggested. such as intelligent tutoring systems (Graesser et al., 2005; Gobert
Rapid advancements in big data and artificial intelligence et al., 2013; Nye, 2015), robotic systems (Toh et al., 2016;
(AI) technologies have had a profound impact on all areas of Anwar et al., 2019), and chatbots (Smutny and Schreiberova,
human society including the economy, politics, science, and 2020). With the breakthroughs in information technologies
education. Thanks in large part to these developments, we are in the last decade, educational psychologists have had greater
able to continue many of our social activities under the COVID- access to big data. Concretely speaking, social media (e.g.,
19 pandemic. Digital tools, platforms, applications, and the Facebook, Twitter), online learning environments [e.g., Massive
communications among people have generated vast amounts of Open Online Courses (MOOCs)], intelligent tutoring systems
data (‘big data’) across disparate locations. Big data technologies (e.g., AutoTutor), learning management systems (LMSs), sensors,
aim at harnessing the power of extensive data in real-time or and mobile devices are generating ever-growing amounts of
otherwise (Daniel, 2019). The characteristic attributes of big dynamic and complex data containing students’ personal records,
data are often referred to as the four V’s. That is, volume physiological data, learning logs and activities, as well as their
(amount of data), variety (diversity of sources and types of learning performance and outcomes (Daniel, 2015). Learning
data), velocity (speed of data transmission and generation), analytics, described as “the measurement, collection, analysis,
and veracity (the accuracy and trustworthiness of data) (Laney, and reporting of data about learners and their contexts, for
2001; Schroeck et al., 2012; Geczy, 2014). Recently, a 5th V purposes of understanding and optimizing learning and the
was added, namely value (i.e., that data could be monetized; environments in which it occurs” (Long and Siemens, 2011,
Dijcks, 2013). Because of intrinsic big data characteristics (the p. 34), are often implemented to analyze these huge amounts of
five Vs), large and complex datasets are impossible to process data (Aldowah et al., 2019). Machine learning and AI techniques
and utilize by using traditional data management techniques. further expand the capabilities of learning analytics (Zawacki-
Hence, novel and innovative computational technologies are Richter et al., 2019). The essential information extracted from
required for the acquisition, storage, distribution, analysis, and big data could be utilized to optimize learning, teaching,
management of big data (Lazer et al., 2014; Geczy, 2015). and administration (Daniel, 2015). Hence, research on big
Big data analytics commonly encompasses the processes of data and AI is gaining increasing significance in education
gathering, analyzing, and evaluating large datasets. Extraction of (Johnson et al., 2011; Becker et al., 2017; Hwang et al.,
actionable knowledge and viable patterns from data are often 2018) and psychology (Harlow and Oswald, 2016; Yarkoni
viewed as the core benefits of the big data revolution (Mayer- and Westfall, 2017; Adjerid and Kelley, 2018; Cheung and
Schönberger and Cukier, 2013; Jagadish et al., 2014). Big data Jak, 2018). Recently, the adoption of big data and AI in the
analytics employ a variety of technologies and tools, such as psychology of learning and teaching has been trending as a
statistical analysis, data mining, data visualization, text analytics, novel method in cutting-edge educational research (Daniel,
social network analysis, signal processing, and machine learning 2015; Starcic, 2019).
(Chen and Zhang, 2014).
As a subset of AI, machine learning focuses on building
computer systems that can learn from and adapt to data THE POSITION FORMULATION
automatically without explicit programming (Jordan and
Mitchell, 2015). Machine learning algorithms can provide A growing body of literature has attempted to uncover the
new insights, predictions, and solutions to customize the value of big data at different education levels, from preschool to
needs and circumstances of each individual. With the higher education (Chen N.-S. et al., 2020). Several journal articles
availability of large quantity and high-quality input training and book chapters have presented retrospective descriptions
data, machine learning processes can achieve accurate results and the latest advances in the rapidly expanding research
and facilitate informed decision making (Manyika et al., area from different angles, including systematic literature
2011; Gobert et al., 2012, 2013; Gobert and Sao Pedro, 2017). review (Zawacki-Richter et al., 2019; Quadir et al., 2020),
These data-intensive, machine learning methods are bibliometric study (Hinojo-Lucena et al., 2019), qualitative

Frontiers in Psychology | www.frontiersin.org 2 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

analysis (Malik et al., 2019; Chen L. et al., 2020), and social AN INTERDISCIPLINARY APPROACH TO
network analysis (Goksel and Bozkurt, 2019). More details EDUCATIONAL ADOPTION OF BIG DATA
can be found in the previously mentioned reviews. In this
paper, we aim at presenting the current progress of the AND AI
application of big data and AI in education. By and large, the
In response to the new opportunities and challenges that the
research on the learner side is devoted to identifying students’
big data explosion and AI revolution are bringing, academics,
learning and affective behavior patterns and profiles, improving
educators, policy-makers, and professionals need to engage
methods of assessment and evaluation, predicting individual
in productive collaboration. They must work together to
students’ learning performance or dropouts, and providing
cultivate our learners’ necessary competencies and essential skills
adaptive systems for personalized support (Papamitsiou and
important for the 21st century work, driven by the knowledge
Economides, 2014; Zawacki-Richter et al., 2019). On the teacher
economy (Bereiter, 2002). Collaboration across diverse
side, numerous studies have attempted to enhance course
disciplines and sectors is a demanding task—particularly when
planning and curriculum development, evaluation of teaching,
individual sides lack a clear vision of their mutually beneficial
and teaching support (Zawacki-Richter et al., 2019; Quadir
interests and the necessary knowledge and skills to realize that
et al., 2020). Additionally, teacher dashboards, such as Inq-
vision. We highlight several overlapping spheres of interest
Blotter, driven by big data techniques are being used to inform
at the intersection of research, policy-making, and industry
teachers’ instruction in real time while students simultaneously
engagements. Researchers and the industry would benefit from
work in Inq-ITS (Gobert and Sao Pedro, 2017; Mislevy et al.,
targeted educational technology development and its efficient
2020). Big data technologies employing learning analytics and
transfer to commercial products. Businesses and governments
machine learning have demonstrated high predictive accuracy
would benefit from legislature that stimulates technology markets
of students’ academic performance (Huang et al., 2020). Only
while suitably protecting data and users’ privacy. Academics
a small number of studies have focused on the effectiveness
and policy makers would benefit from prioritizing educational
of learning analytics programs and AI applications. However,
reforms enabling greater adoption of technology-enhanced
recent findings have revealed encouraging results in terms of
curricula. The recent developments and evolving future trends at
improving students’ academic performance and retention, as well
intersections between researchers, policy-makers, and industry
as supporting teachers in learning design and teaching strategy
stakeholders arising from advancements and deployments of big
refinement (Viberg et al., 2018; Li et al., 2019; Sonderlund et al.,
data and AI technologies in education are illustrated in Figure 1.
2019; Mislevy et al., 2020).
The constructive domains among stakeholders progressively
Despite the growing number of reports and methods
evolve along with scientific and technological developments.
outlining implementations of big data and AI technologies
Therefore, it is important to reflect on longer-term projections
in educational environments, we see a notable gap between
and challenges. The following sections highlight the
contemporary technological capabilities and their utilization for
novel challenges and future directions of big data and
education. The fast-growing education industry has developed
AI technologies at the intersection of education research,
numerous data processing techniques and AI applications, which
policy-making, and industry.
may not be guided by current theoretical frameworks and
research findings from psychology of learning and teaching.
The rapid pace of technological progress and relatively slow BIG DATA AND AI IN EDUCATION:
educational adoption have contributed to the widening gap
between technology readiness and its application in education RESEARCH
(Macfadyen, 2017). There is a pressing need to reduce this
An understanding of individual differences is critical for
gap and stimulate technological adoption in education. This
developing pedagogical tools to target specific students and to
work presents varying viewpoints and their controversial issues,
tailor education to individual needs at different stages. Intelligent
contemporary research, and prospective future developments
educational systems employing big data and AI techniques are
in adoption of big data and AI in education. We advocate
capable of collecting accurate and rich personal data. Data
an interdisciplinary approach that encompasses educational,
analytics can reveal students’ learning patterns and identify their
technological, and governmental spheres of influence. In the
specific needs (Gobert and Sao Pedro, 2017; Mislevy et al.,
educational domain, there is a relative lack of knowledge and
2020). Hence, big data and AI have the potential to realize
skills in AI and big data applications. On the technological
individualized learning to achieve precision education (Lu et al.,
side, few data scientists and AI developers are familiar
2018). We see the following emerging trends, research gaps,
with the advancements in education psychology, though this
and controversies in integrating big data and AI into education
is changing with the advent of graduate programs at the
research so that there is a deep and rigorous understanding of
intersection of Learning Sciences and Computer Science. Finally,
individual differences that can be used to personalize learning in
in terms of government policies, the main challenges faced
real time and at scale.
are the regulatory and ethical dilemmas between support of
educational reforms and restrictions on adoptions of data- (1) Education is progressively moving from a one-size-fits-all
oriented technologies. approach to precision education or personalized learning

Frontiers in Psychology | www.frontiersin.org 3 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

FIGURE 1 | Contemporary developments and future trends at the intersections between research, policy, and industry driven by big data and AI advances in
education.

(Lu et al., 2018; Tsai et al., 2020). The one-size-fits-all focus that incorporates cognition in its design, such
approach was designed for average students, whereas as perception (VanRullen, 2017), emotion (Song et al.,
precision education takes into consideration the individual 2016), and cognitive thinking (Bramley et al., 2017).
differences of learners in their learning environments, Moreover, it is also shifting from a single domain (e.g.,
along with their learning strategies. The main idea of domain expertise, or expert systems) to a cross-disciplinary
precision education is analogous to “precision medicine,” approach through collaboration (Spikol et al., 2018;
where researchers harvest big data to identify patterns Krouska et al., 2019) and domain transfers (L’heureux et al.,
relevant to specific patients such that prevention and 2017). These controversial shifts are facilitating transitions
treatment can be customized. Based on the analysis from the knowing of the unknown (gaining insights
of student learning profiles and patterns, precision through reasoning) to the unknown of the unknown
education predicts students’ performance and provides (figuring out hidden values and unknown results through
timely interventions to optimize learning. The goal of algorithms) (Abed Ibrahim and Fekete, 2019; Cutumisu
precision education is to improve the diagnosis, prediction, and Guo, 2019). In other words, deterministic learning,
treatment, and prevention of learning outcomes (Lu et al., aimed at deductive/inductive reasoning and inference
2018). Contemporary research gaps related to adaptive engines, predominated in traditional expert systems and
tools and personalized educational experiences are old AI. Whereas, today, dynamic and stochastic learning,
impeding the transition to precision education. Adaptive the outcome of which involves some randomness and
educational tools and flexible learning systems are needed uncertainty, is gradually becoming the trend in modern
to accommodate individual learners’ interaction, pace, machine learning techniques.
and learning progress, and to fit the specific needs of (3) The format of machine-generated data and the purpose of
the individual learners, such as students with learning machine learning algorithms should be carefully designed.
disabilities (Xie et al., 2019; Zawacki-Richter et al., 2019). There is a notable gap between theoretical design and its
Hence, as personalized learning is customized for different applicability. A theoretical model is needed to guide the
people, researchers are able to focus on individualized development, interpretation, and validation of algorithms
learning that is adaptive to individual needs in real time (Gobert et al., 2013; Hew et al., 2019). The outcomes
(Gobert and Sao Pedro, 2017; Lu et al., 2018). of data analytics and algorithmically generated evidence
(2) The research focus on deploying AI in education is must be shared with educators and applied with caution.
gradually shifting from a computational focus that For instance, efforts to algorithmically detect mental
demonstrates use cases of new technology to cognitive states such as boredom, frustration, and confusion (Baker

Frontiers in Psychology | www.frontiersin.org 4 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

et al., 2010) must be supported by the operational A huge amount of learning data is created accordingly.
definitions and constructs that have been prudently The extraction of meaningful patterns and the discovery
evaluated. Additionally, the affective data collected by AI of knowledge from these data is expected to be carried out
systems should take into account the cultural differences through learning analytics and AI techniques. Inevitably,
combined with contextual factors, teachers’ observations, the current learning cultures, learning experiences,
and students’ opinions (Yadegaridehkordi et al., 2019). and classroom dynamics are changing as “we live
Data need to be informatively and qualitatively balanced, algorithmic lives” (Bucher, 2018). Thus, there is a critical
in order to avoid implicit biases that may propagate into need to adopt proper learning theories of educational
algorithms trained on such data (Staats, 2016). psychology and to encourage our learners to be active
(4) There are ethical and algorithmic challenges when participants rather than passive recipients or merely
balancing human provided learning and machine assisted tracked objects (Loftus and Madden, 2020). For example,
learning. The significant influence of AI and contemporary under the constructionist framework (Tsai, 2000), the
technologies is a double-edged sword (Khechine and technology-enhanced or AI-powered education may
Lakhal, 2018). On the one hand, it facilitates better empower students to know their learning activities and
usability and drives progress. On the other, it might lead patterns, predict their possible learning outcomes, and
to the algorithmic bias and loss of certain essential skills strategically regulate their learning behavior (Koh et al.,
among students who are extensively relying on technology. 2014; Loftus and Madden, 2020). On the other hand,
For instance, in creativity- or experience-based learning, in the era of information explosion and AI revolution,
technology may even become an obstacle to learning, the disadvantaged students and developing countries
since it may hinder students from attaining first-hand are indeed facing a wider digital divide. To reduce the
experiences and participating in the learning activities inequalities and bring more opportunities, cultivating
(Cuthbertson et al., 2004). Appropriately balancing the young people’s competencies is seemed like one of the
technology adoption and human involvement in various most promising means (UNESCO, 2015). Meanwhile,
educational contexts will be a challenge in the foreseeable overseas support from international organizations
future. Nonetheless, the convergence of human and such as World Bank and UNESCO are imperative
machine learning has the potential for highly effective for developing countries in their communication
teaching and learning beyond the simple “sum of the parts infrastructure establishment (e.g., hardware, software,
of human and artificial intelligence” (Topol, 2019). connectivity, electricity). Naturally, technology will not
(5) Algorithmic bias is another controversial issue (Obermeyer replace or hinder human learning; rather, a smart use of
et al., 2019). Since modern AI algorithms extensively rely new technologies will facilitate transfer and acquisition of
on data, their performance is governed solely by data. knowledge (Azevedo et al., 2019).
Algorithms adapt to inherent qualitative and quantitative
characteristics of data. For example, if data is unbalanced An overarching theme from the above trends of research is
and contains disproportionately better information that we need theories of cognitive and educational psychology
on students from general population in comparison to guide our understanding of the individual learner (and
to minorities, the algorithms may produce systematic individual differences), in order to develop best tools, algorithms,
and repeatable errors disadvantaging minorities. These and practices for personalized learning. Take, for example, VR
controversial issues need to be addressed before its wide (virtual reality) or AR (augmented reality) as a fast-developing
implementation in education practice since every single technology for education. The industry has developed many
student is precious. More rigorous studies and validation different types of VR/AR applications (e.g., Google Expeditions
in real learning environments are required though work with over 100 virtual field trips), but these have typically
along these lines is being done (Sao Pedro et al., 2013). been developed in the views of the industry (see further
(6) The fast expansion of technology and inequalities of discussion below) and may not be informed by theories and
learning opportunities has aroused great controversies. data from educational psychology about how students actually
Due to the exponential nature of technological progress, learn. To make VR/AR effective learning tools, we must
particularly big data and AI revolution, a fresh paradigm separate the technological features from the human experiences
and new learning landscape are on the horizon. For and abilities (e.g., cognitive, linguistic, spatial abilities of the
instance, the elite smartphone 10 years ago, in 2010, was learner; see Li et al., 2020). For example, VR provides a high-
BlackBerry. Today, 10 years later, even in sub-Saharan fidelity 3D real-life virtual environment, and the technological
Africa, 75% of the population has mobile phones several tools are built on the assumption that 3D realism enables
generations more advanced (GSMA Intelligence, 2020). the learner to gain ‘perceptual grounding’ during learning
Hence, the entry barriers are shifting from the technical (e.g., having access to visual, auditory, tactile experiences as
requirements to the willingness of and/or need for in real world). Following the ‘embodied cognition’ theory
adoption. This has been clearly demonstrated during (Barsalou, 2008), we should expect VR learning to yield better
the COVID-19 pandemic. The need for social distancing learning outcomes compared with traditional classroom learning.
and continuing education has led to online/e-learning However, empirical data suggest that there are significant
deployments within months (United Nations, 2020). individual differences in that some students benefit more than

Frontiers in Psychology | www.frontiersin.org 5 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

others from VR learning. It may be that the individuals with (Boyd and Crawford, 2012; Pardo and Siemens, 2014).
higher cognitive and perceptual abilities need no additional Education records and personal data are highly sensitive.
visuospatial information (provided in VR) to succeed in There are significant risks associated with students’
learning. In any case, we need to understand how embodied educational profiles, records, and other personal data.
experiences (provided by the technology) interact with different Appropriate security measures must be adopted by
learners’ inherent abilities (as well as their prior knowledge educational institutions. Commercial educational system
and background) for the best application of the relevant providers are actively exploiting both legislative gaps
technology in education. and concealed data acquisition channels. Increasing
numbers of industry players are implementing data-
oriented business models (Geczy, 2018). There is a vital
BIG DATA AND AI IN EDUCATION: role to play for legislative, regulatory, and enforcing bodies
at both the national and local levels. It is pertinent
POLICY-MAKING
that governments enact, implement, and enforce privacy
Following the revolution triggered by breakthroughs in big and personal data protection legislation and measures.
data and AI technology, policy-makers have attempted to In doing so, there is a need to strike a proper balance
formulate strategies and policies regarding how to incorporate between desirable use of personal data for educational
AI and emerging technologies into primary, secondary, and purposes and undesirable commercial monetization and
tertiary education (Pedró et al., 2019). Major challenges must abuse of personal data.
be overcome in order to suitably integrate big data and AI
into educational practice. The following three segments highlight
pertinent policy-oriented challenges, gaps, and evolving trends. BIG DATA AND AI IN EDUCATION:
(1) In digitally-driven knowledge economies, traditional INDUSTRY
formal education systems are undergoing drastic changes
As scientific and academic aspects of big data and AI in education
or even a paradigm shift (Peters, 2018). Lifelong learning
have their unique challenges, so does the commercialization of
is quickly being adopted and implemented through
educational tools and systems (Renz et al., 2020). Numerous
online or project-based learning schemes that incorporate
countries have attempted to stimulate innovation-based growth
multiple ways of teaching (Lenschow, 1998; Sharples,
through enhancing technology transfer and fostering academia-
2000; Field, 2001; Koper and Tattersall, 2004). This new
industry collaboration (Huggins and Thompson, 2015). In the
concept of continual education will require micro-credits
United States, this was initiated by the Bayh-Dole Act (Mowery
or micro-degrees to sustain learners’ efforts (Manuel
et al., 2001). Building a reciprocal and sustained partnership
Moreno-Marcos et al., 2019). The need to change the
is strongly encouraged. It facilitates technology transfers and
scope and role of education will become evident in the
strengthens the links between academia and the education
near future (Williams, 2019). For example, in the next
industry. There are several points to be considered when
few years, new instruction methods, engagement, and
approaching academia-industry collaboration. It is important
assessment will need to be developed in formal education
that collaboration is mutually beneficial. The following points
to support lifelong education. The system should be based
highlight the overlapping spheres of benefits for both educational
on micro-credits or micro-degrees.
commerce and academia. They also expose existing gaps and
(2) Solutions for integrating cutting-edge research findings,
future prospects.
innovative theory-driven curricula, and emerging
technologies into students’ learning are evidently (1) Commercializing intelligent educational tools and systems
beneficial, and perhaps even ready for adoption. However, that include the latest scientific and technological advances
there is an apparent divergence between a large number of can provide educators with tools for developing more
pre-service and in-service teachers and their willingness effective curricula, pedagogical frameworks, assessments,
to support and adopt these emerging technologies (Pedró and programs. Timely release of educational research
et al., 2019). Pre-service teachers have greater exposure advances onto commercial platforms is desirable by
to modern technologies and, in general, are more willing vendors from development, marketing, and revenue
to adopt them. In-service teachers have greater practical perspectives (Renz and Hilbig, 2020). Implementation of
experience and tend to more rely on it. To bridge the the latest research enables progressive development of
gap, effective teacher education programs and continuing commercial products and distinctive differentiation for
education programs have to be developed and offered to marketing purposes. This could also potentially solve
support the adoption of these new technologies so that the significant gap between what the industry knows
they can be implemented with fidelity (O’Donnell, 2008). and develops and what the academic research says with
This issue could become even more pressing to tackle in regard to student learning. Novel features may also be
light of the extended period of the COVID-19 pandemic. suitably monetized—hence, expanding revenue streams.
(3) A suitable legislative framework is needed to protect The gaps between availability of the latest research and
personal data from unscrupulous collection, unauthorized its practical adoption are slowing progress and negatively
disclosure, commercial exploitation, and other abuses impacting commercial vendors. A viable solution is a

Frontiers in Psychology | www.frontiersin.org 6 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

closer alignment and/or direct collaboration between (3) Vocational and practical education provides numerous
academia and industry. opportunities for fruitful academia-industry collaboration.
(2) A greater spectrum of commercially and freely available With the changing nature of work and growing technology
tools helps maintain healthy market competition. It adoption, there is an increasing demand for radical
also helps to avoid monopolies and oligopolies that changes in vocational education—for both teachers
stifle innovation, limit choices, and damage markets for and students (World Development and Report, 2019).
educational tools. Some well-stablished or free-of-charge Domain knowledge provided by teachers is beneficially
platforms (e.g., Moodle, LMS) might show such potential supplemented by AI-assisted learning environments
of oligopolies during the COVID-19 pandemic. With more in academia. Practical skills are enhanced in industrial
tools available on the market, educators and academics environments with hands-on experience and feedback
may explore novel avenues for improving education and from both trainers and technology tools. Hence, students
research. New and more effective forms of education may benefit from acquiring domain knowledge and enhancing
be devised. For instance, multimodal virtual educational their skills via interactions with human teachers and
environments have high potential future prospects. These trainers. Equally, they benefit from gaining the practical
skills via interactions with simulated and real-world
are environments that would otherwise be impossible in
technological environments. Effective vocational training
conventional physical settings (see previous discussion of
demands teachers and trainers on the human-learning
VR/AR). Expanding educational markets and commerce
side, and AI environments and actual technology tools on
should inevitably lead to expanding resources for research
machine-learning side. Collaboration between academia
and development funding (Popenici and Kerr, 2017).
and industry, as well as balanced human and machine
Collaborative research projects sponsored by the industry
learning approaches are pertinent for vocational education.
should provide support and opportunities for academics
to advance educational research. Controversially, in
numerous geographies there is a decreasing trend in
collaborative research. To reverse the trend, it is desirable DISCUSSION AND CONCLUSION
that academic researchers and industry practitioners
increase their engagements via mutual presentations, Big data and AI have enormous potential to realize highly
educations, and even government initiatives. All three effective learning and teaching. They stimulate new research
stakeholders (i.e., academia, industry, and government) questions and designs, exploit innovative technologies and
should play more active roles. tools in data collection and analysis, and ultimately become

TABLE 1 | Major challenges and possible solutions for integrating big data and AI into education.

Aspect Major challenges Possible solutions

Research • The mode of education is progressively moving from a one-size-fits-all • Adaptive educational tools and flexible learning systems will be needed
approach to precision education and individualized learning. to accommodate individual learners’ needs.
• AI research in education is currently focused on intelligent computing • The research focus on deploying AI in education needs to incorporate
technologies in a single domain. theories of cognition and knowledge about individual differences in
• The format, purpose, and meaning of machine-generated data should student learning.
be carefully designed. • A theoretical model is needed to guide the development, interpretation,
• The significant influence of AI and big data technologies is a and validation of algorithms. The data analytics must be applied with
double-edged sword. caution.
• Future studies should be aimed at using educational technologies in the
appropriate context tailored to the characteristics of individual learners.
Policy-making • In digitally-driven knowledge economies, traditional formal education • New methods of instruction, engagement, and assessment will need to
systems are undergoing drastic changes or even a paradigm shift. be developed in formal education to support lifelong education systems
• A large number of pre-service and in-service teachers are not ready to based on micro-credits or micro-degrees.
support and adopt new technologies. • Effective teacher education and continuing education programs have to
• There is a pressing need for privacy and personal data protections be designed and offered to support the adoption of these new
against unauthorized disclosure, commercial exploitation, and other technologies.
abuses. • The government must seek an optimal balance between personal data
collection and personal data protection in policy-making,
implementation, and enforcement.
Industry • The commercialization of intelligent educational tools and systems • Building a reciprocal and sustained partnership between academia and
presents a set of difficult challenges. the education industry is strongly encouraged.
• Expanding spectrum of commercially and freely available tools is • Collaborative research projects sponsored by the industry should
necessary to maintain healthy market competition. provide support for academics to advance applied research and its
• Vocational and practical trainings need radical changes to remain commercialization.
relevant and prudent. • Closer academia-industry collaboration with balanced human-oriented
and machine-assisted learning.

Frontiers in Psychology | www.frontiersin.org 7 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

a mainstream research paradigm (Daniel, 2019). Nonetheless, data. Such data-intensive and technology-driven analysis of
they are still fairly novel and unfamiliar to many researchers human behavior, in real-world and simulated environments, may
and educators. In this paper, we have described the general assist teachers in identifying students’ learning trajectories and
background, core concepts, and recent progress of this rapidly patterns, developing corresponding lesson plans, and adopting
growing domain. Along with the arising opportunities, we have effective teaching strategies (Klašnja-Milicevic et al., 2017; Gierl
highlighted the crucial challenges and emerging trends of big data and Lai, 2018). It may also support teachers in tackling students’
and AI in education, which are reflected in educational research, more complex problems and cultivating students’ higher-order
policy-making, and industry. Table 1 concisely summarizes the thinking skills by freeing the teachers from their monotonous
major challenges and possible solutions of big data and AI in and routine tasks (Li, 2007; Belpaeme et al., 2018). Hence, it is
education. In summary, future studies should be aimed at theory- now imperative for us to embrace AI and technology and prepare
based precision education, incorporating cross-disciplinary our teachers and students for the future of AI-enhanced and
application, and appropriately using educational technologies. technology-supported education.
The government should be devoted to supporting lifelong The adoption of big data and AI in learning and teaching
learning, offering teacher education programs, and protecting is still in its infancy and limited by technological and mindset
personal data. With regard to the education industry, reciprocal challenges for now; however, the convergence of developments
and mutually beneficial relationships should be developed in in psychology, data science, and computer science shows great
order to enhance academia-industry collaboration. promise in revolutionizing educational research, practice, and
Regarding the future development of big data and AI, we industry. We hope that the latest achievements and future
advocate an in-depth dialog between the supporters of “cold” directions presented in this paper will advance our shared goal
technology and “warm” humanity so that users of technology of helping learners and teachers pursue sustainable development.
can benefit from its capacity and not see it as a threat to their
livelihood. An equally important issue is that overreliance on
technology may lead to an underestimation of the role of humans AUTHOR CONTRIBUTIONS
in education. Remember the fundamental role of schooling: the
school is a great equalizer as well as a central socialization agent. HLu wrote the initial draft of the manuscript. PG, HLa, JG,
We need to better understand the role of social and affective and PL revised the drafts and provided theoretical background.
processing (e.g., emotion, motivation) in addition to cognitive SY, HO, JB, and RG contributed content for the original
processing in student learning successes (or failures). After all, draft preparation of the manuscript. C-CT provided theoretical
human learning is a social behavior, and a number of key regions focus, design, draft feedback, and supervised throughout the
in our brains are wired to be socially engaged (see Li and Jeong, research. All authors contributed to the article and approved the
2020 for a discussion). submitted version.
It has been estimated that approximately half of the
current routine jobs might be automated in the near future
(Frey and Osborne, 2017; World Development and Report, FUNDING
2019). However, the teacher’s job could not be replaced.
The teacher-student relationship is indispensable in students’ This work was financially supported by the Institute for Research
learning, and inspirational in students’ personal growth (Roorda Excellence in Learning Sciences of National Taiwan Normal
et al., 2011; Cheng and Tsai, 2019). On the other hand, University (NTNU) from the Featured Areas Research Center
new developments in technologies will enable us to collect Program within the framework of the Higher Education Sprout
and analyze large-scale, multimodal, and continuous real-time Project by the Ministry of Education (MOE) in Taiwan.

REFERENCES Baker, R. S. J. D., D’Mello, S. K., Rodrigo, M. M. T., and Graesser, A. C. (2010).
Better to be frustrated than bored: the incidence, persistence, and impact
Abed Ibrahim, L., and Fekete, I. (2019). What machine learning can tell us about of learners’ cognitive-affective states during interactions with three different
the role of language dominance in the diagnostic accuracy of german litmus computer-based learning environments. Int. J. Human-Comp. Stud. 68, 223–
non-word and sentence repetition tasks. Front. Psychol. 9:2757. doi: 10.3389/ 241. doi: 10.1016/j.ijhcs.2009.12.003
fpsyg.2018.02757 Barsalou, L. W. (2008). “Grounding symbolic operations in the brain’s
Adjerid, I., and Kelley, K. (2018). Big data in psychology: a framework for research modal systems,” in Embodied Grounding: Social, Cognitive, Affective, and
advancement. Am. Psychol. 73, 899–917. doi: 10.1037/amp0000190 Neuroscientific Approaches, eds G. R. Semin and E. R. Smith (Cambridge:
Aldowah, H., Al-Samarraie, H., and Fauzy, W. M. (2019). Educational data mining Cambridge University Press), 9–42. doi: 10.1017/cbo9780511805837.002
and learning analytics for 21st century higher education: a review and synthesis. Becker, S. A., Cummins, M., Davis, A., Freeman, A., Hall, C. G., and
Telemat. Inform. 37, 13–49. doi: 10.1016/j.tele.2019.01.007 Ananthanarayanan, V. (2017). NMC Horizon Report: 2017 Higher Education
Anwar, S., Bascou, N. A., Menekse, M., and Kardgar, A. (2019). A systematic review Edition. Austin, TX: The New Media Consortium.
of studies on educational robotics. J. Pre-College Eng. Educ. Res. (J-PEER) 9, Belpaeme, T., Kennedy, J., Ramachandran, A., Scassellati, B., and Tanaka, F. (2018).
19–42. doi: 10.7771/2157-9288.1223 Social robots for education: a review. Sci. Robot. 3:eaat5954. doi: 10.1126/
Azevedo, J. P. W. D., Crawford, M. F., Nayar, R., Rogers, F. H., Barron Rodriguez, scirobotics.aat5954
M. R., Ding, E. Y. Z., et al. (2019). Ending Learning Poverty: What Will It Take?. Bereiter, C. (2002). Education and MIND in the Knowledge Age. Mahwah,
Washington, D.C: The World Bank. NJ: LEA.

Frontiers in Psychology | www.frontiersin.org 8 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

Boyd, D., and Crawford, K. (2012). Critical questions for big data: provocations for Graesser, A. C., Chipman, P., Haynes, B. C., and Olney, A. (2005). AutoTutor: an
a cultural, technological, and scholarly phenomenon. Inform. Commun. Soc. 15, intelligent tutoring system with mixed-initiative dialogue. IEEE Trans. Educ. 48,
662–679. doi: 10.1080/1369118x.2012.678878 612–618. doi: 10.1109/te.2005.856149
Bramley, N. R., Dayan, P., Griffiths, T. L., and Lagnado, D. A. (2017). Formalizing GSMA Intelligence (2020). The Mobile Economy 2020. London: GSM Association.
Neurath’s ship: approximate algorithms for online causal learning. Psychol. Rev. Harlow, L. L., and Oswald, F. L. (2016). Big data in psychology: introduction to the
124, 301–338. doi: 10.1037/rev0000061 special issue. Psychol. Methods 21, 447–457. doi: 10.1037/met0000120
Bucher, T. (2018). If Then: Algorithmic Power and Politics. New York, NY: Oxford Hew, K. F., Lan, M., Tang, Y., Jia, C., and Lo, C. K. (2019). Where is the “theory”
University Press. within the field of educational technology research? Br. J. Educ. Technol. 50,
Carbonell, J. R. (1970). AI in CAI: an artificial-intelligence approach to computer- 956–971. doi: 10.1111/bjet.12770
assisted instruction. IEEE Trans. Man-Machine Sys. 11, 190–202. doi: 10.1109/ Hinojo-Lucena, F. J., Aznar-Díaz, I., Cáceres-Reche, M. P., and Romero-Rodríguez,
TMMS.1970.299942 J. M. (2019). Artificial intelligence in higher education: a bibliometric study
Chen, C. P., and Zhang, C. Y. (2014). Data-intensive applications, challenges, on its impact in the scientific literature. Educ. Sci. 9:51. doi: 10.3390/
techniques and technologies: a survey on big data. Inform. Sci. 275, 314–347. educsci9010051
doi: 10.1016/j.ins.2014.01.015 Huang, A. Y., Lu, O. H., Huang, J. C., Yin, C., and Yang, S. J. (2020). Predicting
Chen, L., Chen, P., and Lin, Z. (2020). Artificial intelligence in education: a review. students’ academic performance by using educational big data and learning
IEEE Access 8, 75264–75278. doi: 10.1109/ACCESS.2020.2988510 analytics: evaluation of classification methods and learning logs. Int. Learn.
Chen, N.-S., Yin, C., Isaias, P., and Psotka, J. (2020). Educational big data: Environ. 28, 206–230. doi: 10.1080/10494820.2019.1636086
extracting meaning from data for smart education. Interact. Learn. Environ. 28, Huggins, R., and Thompson, P. (2015). Entrepreneurship, innovation and regional
142–147. doi: 10.1080/10494820.2019.1635395 growth: a network theory. Small Bus. Econ. 45, 103–128. doi: 10.1007/s11187-
Cheng, K.-H., and Tsai, C.-C. (2019). A case study of immersive virtual field trips 015-9643-3
in an elementary classroom: students’ learning experience and teacher-student Hwang, G.-J., Spikol, D., and Li, K.-C. (2018). Guest editorial: trends and research
interaction behaviors. Comp. Educ. 140:103600. doi: 10.1016/j.compedu.2019. issues of learning analytics and educational big data. Educ. Technol. Soc. 21,
103600 134–136.
Cheung, M. W.-L., and Jak, S. (2018). Challenges of big data analyses and Jagadish, H. V., Gehrke, J., Labrinidis, A., Papakonstantinou, Y., Patel, J. M.,
applications in psychology. Zeitschrift Fur Psychol. J. Psychol. 226, 209–211. Ramakrishnan, R., et al. (2014). Big data and its technical challenges. Commun.
doi: 10.1027/2151-2604/a000348 ACM. 57, 86–94. doi: 10.1145/2611567
Cuthbertson, B., Socha, T. L., and Potter, T. G. (2004). The double-edged sword: Johnson, L., Smith, R., Willis, H., Levine, A., and Haywood, K. (2011). The 2011
critical reflections on traditional and modern technology in outdoor education. Horizon Report. Austin, TX: The New Media Consortium.
J. Adv. Educ. Outdoor Learn. 4, 133–144. doi: 10.1080/14729670485200491 Jordan, M. I., and Mitchell, T. M. (2015). Machine learning: trends, perspectives,
Cutumisu, M., and Guo, Q. (2019). Using topic modeling to extract pre- and prospects. Science 349, 255–260. doi: 10.1126/science.aaa8415
service teachers’ understandings of computational thinking from their coding Khechine, H., and Lakhal, S. (2018). Technology as a double-edged sword:
reflections. IEEE Trans. Educ. 62, 325–332. doi: 10.1109/te.2019.2925253 from behavior prediction with UTAUT to students’ outcomes considering
Daniel, B. (2015). Big data and analytics in higher education: opportunities and personal characteristics. J. Inform. Technol. Educ. Res. 17, 63–102. doi: 10.28945/
challenges. Br. J. Educ. Technol. 46, 904–920. doi: 10.1111/bjet.12230 4022
Daniel, B. K. (2019). Big data and data science: a critical review of issues for Klašnja-Milicevic, A., Ivanovic, M., and Budimac, Z. (2017). Data science in
educational research. Br. J. Educ. Technol. 50, 101–113. doi: 10.1111/bjet.12595 education: big data and learning analytics. Comput. Applicat. Eng. Educ. 25,
Dijcks, J. (2013). Oracle: Big data for the enterprise. Oracle White Paper. Redwood 1066–1078. doi: 10.1002/cae.21844
Shores, CA: Oracle Corporation. Koh, J. H. L., Chai, C. S., and Tsai, C. C. (2014). Demographic factors, TPACK
Field, J. (2001). Lifelong education. Int. J. Lifelong Educ. 20, 3–15. doi: 10.1080/ constructs, and teachers’ perceptions of constructivist-oriented TPACK. J. Educ.
09638280010008291 Technol. Soc. 17, 185–196.
Frey, C. B., and Osborne, M. A. (2017). The future of employment: how susceptible Koper, R., and Tattersall, C. (2004). New directions for lifelong learning using
are jobs to computerisation? Technol. Forecast. Soc. Change 114, 254–280. doi: network technologies. Br. J. Educ. Technol. 35, 689–700. doi: 10.1111/j.1467-
10.1016/j.techfore.2016.08.019 8535.2004.00427.x
Geczy, P. (2014). Big data characteristics. Macrotheme Rev. 3, 94–104. Krouska, A., Troussas, C., and Virvou, M. (2019). SN-Learning: an exploratory
Geczy, P. (2015). Big data management: relational framework. Rev. Bus. Finance study beyond e-learning and evaluation of its applications using EV-SNL
Stud. 6, 21–30. framework. J. Comp. Ass. Learn. 35, 168–177. doi: 10.1111/jcal.12330
Geczy, P. (2018). Data-Oriented business models: gaining competitive advantage. Laney, D. (2001). 3D data management: controlling data volume, velocity and
Global J. Bus. Res. 12, 25–36. variety. META Group Res. Note 6, 70–73.
Gierl, M. J., and Lai, H. (2018). Using automatic item generation to create solutions Lazer, D., Kennedy, R., King, G., and Vespignani, A. (2014). The parable of Google
and rationales for computerized formative testing. Appl. Psychol. Measurement Flu: traps in big data analysis. Science 343, 1203–1205. doi: 10.1126/science.
42, 42–57. doi: 10.1177/0146621617726788 1248506
Gobert, J., Sao Pedro, M., Raziuddin, J., and Baker, R. S. (2013). From log files to Lenschow, R. J. (1998). From teaching to learning: a paradigm shift in engineering
assessment metrics for science inquiry using educational data mining. J. Learn. education and lifelong learning. Eur. J. Eng. Educ. 23, 155–161. doi: 10.1080/
Sci. 22, 521–563. doi: 10.1080/10508406.2013.837391 03043799808923494
Gobert, J. D., and Sao Pedro, M. A. (2017). “Digital assessment environments L’heureux, A., Grolinger, K., Elyamany, H. F., and Capretz, M. A. (2017). Machine
for scientific inquiry practices,” in The Wiley Handbook of Cognition and learning with big data: challenges and approaches. IEEE Access 5, 7776–7797.
Assessment, eds A. A. Rupp and J. P. Leighton (West Sussex: Frameworks, doi: 10.1109/ACCESS.2017.2696365
Methodologies, and Applications), 508–534. doi: 10.1002/9781118956588. Li, H., Gobert, J., and Dickler, R. (2019). “Evaluating the transfer of scaffolded
ch21 inquiry: what sticks and does it last?,” in Artificial Intelligence in Education, eds
Gobert, J. D., Sao Pedro, M. A., Baker, R. S., Toto, E., and Montalvo, O. (2012). S. Isotani, E. Millán, A. Ogan, P. Hastings, B. McLaren, and R. Luckin (Cham:
Leveraging educational data mining for real-time performance assessment of Springer), 163–168. doi: 10.1007/978-3-030-23207-8_31
scientific inquiry skills within microworlds. J. Educ. Data Min. 4, 104–143. Li, P., and Jeong, H. (2020). The social brain of language: grounding second
doi: 10.5281/zenodo.3554645 language learning in social interaction. npj Sci. Learn. 5:8. doi: 10.1038/s41539-
Goksel, N., and Bozkurt, A. (2019). “Artificial intelligence in education: current 020-0068-7
insights and future perspectives,” in Handbook of Research on Learning in the Li, P., Legault, J., Klippel, A., and Zhao, J. (2020). Virtual reality for student
Age of Transhumanism, eds S. Sisman-Ugur and G. Kurubacak (Hershey, PA: learning: understanding individual differences. Hum. Behav. Brain 1, 28–36.
IGI Global), 224–236 doi: 10.4018/978-1-5225-8431-5.ch014 doi: 10.37716/HBAB.2020010105

Frontiers in Psychology | www.frontiersin.org 9 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

Li, X. (2007). Intelligent agent–supported online education. Dec. Sci. J. Innovat. Roorda, D. L., Koomen, H. M. Y., Spilt, J. L., and Oort, F. J. (2011). The influence
Educ. 5, 311–331. doi: 10.1111/j.1540-4609.2007.00143.x of affective teacher-student relationships on students’ school engagement and
Loftus, M., and Madden, M. G. (2020). A pedagogy of data and Artificial achievement: a meta-analytic approach. Rev. Educ. Res. 81, 493–529. doi: 10.
intelligence for student subjectification. Teach. Higher Educ. 25, 456–475. doi: 3102/0034654311421793
10.1080/13562517.2020.1748593 Sao Pedro, M., Baker, R., and Gobert, J. (2013). “What different kinds
Long, P., and Siemens, G. (2011). Penetrating the fog: analytics in learning and of stratification can reveal about the generalizability of data-mined skill
education. Educ. Rev. 46, 31–40. doi: 10.1007/978-3-319-38956-1_4 assessment models,” in Proceedings of the 3rd Conference on Learning Analytics
Lu, O. H. T., Huang, A. Y. Q., Huang, J. C. H., Lin, A. J. Q., Ogata, H., and Yang, and Knowledge (Leuven), 190–194.
S. J. H. (2018). Applying learning analytics for the early prediction of students’ Schroeck, M., Shockley, R., Smart, J., Romero-Morales, D., and Tufano, P. (2012).
academic performance in blended learning. Educ. Technol. Soc. 21, 220–232. Analytics: the real-world use of big data. IBM Global Bus. Serv. 12, 1–20.
Macfadyen, L. P. (2017). Overcoming barriers to educational analytics: how doi: 10.1002/9781119204183.ch1
systems thinking and pragmatism can help. Educ. Technol. 57, 31–39. Sharples, M. (2000). The design of personal mobile technologies for lifelong
Malik, G., Tayal, D. K., and Vij, S. (2019). “An analysis of the role of artificial learning. Comp. Educ. 34, 177–193. doi: 10.1016/s0360-1315(99)0
intelligence in education and teaching,” in Recent Findings in Intelligent 0044-5
Computing Techniques. Advances in Intelligent Systems and Computing, eds P. Smutny, P., and Schreiberova, P. (2020). Chatbots for learning: a review of
Sa, S. Bakshi, I. Hatzilygeroudis, and M. Sahoo (Singapore: Springer), 407–417. educational chatbots for the facebook messenger. Comp. Educ. 151:103862.
Manuel Moreno-Marcos, P., Alario-Hoyos, C., Munoz-Merino, P. J., and Delgado doi: 10.1016/j.compedu.2020.103862
Kloos, C. (2019). Prediction in MOOCs: a review and future research Sonderlund, A. L., Hughes, E., and Smith, J. (2019). The efficacy of learning
directions. IEEE Trans. Learn. Technol. 12, 384–401. doi: 10.1109/TLT.2018.285 analytics interventions in higher education: a systematic review. Br. J. Educ.
6808 Technol. 50, 2594–2618. doi: 10.1111/bjet.12720
Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C., et al. (2011). Song, Y., Dai, X.-Y., and Wang, J. (2016). Not all emotions are created
Big data: The Next Frontier for Innovation, Competition and Productivity. equal: expressive behavior of the networked public on China’s social
New York, NY: McKinsey Global Institute. media site. Comp. Hum. Behav. 60, 525–533. doi: 10.1016/j.chb.2016.
Mayer-Schönberger, V., and Cukier, K. (2013). Big data: A Revolution That Will 02.086
Transform How we live, Work, and Think. Boston, MA: Houghton Mifflin Spikol, D., Ruffaldi, E., Dabisias, G., and Cukurova, M. (2018). Supervised
Harcourt. machine learning in multimodal learning analytics for estimating success in
Mislevy, R. J., Yan, D., Gobert, J., and Sao Pedro, M. (2020). “Automated scoring in project-based learning. J. Comp. Ass. Learn. 34, 366–377. doi: 10.1111/jcal.
intelligent tutoring systems,” in Handbook of Automated Scoring, eds D. Yan, 12263
A. A. Rupp, and P. W. Foltz (London: Chapman and Hall/CRC), 403–422. Staats, C. (2016). Understanding implicit bias: what educators should know. Am.
doi: 10.1201/9781351264808-22 Educ. 39, 29–33. doi: 10.2307/3396655
Mowery, D. C., Nelson, R. R., Sampat, B. N., and Ziedonis, A. A. (2001). The growth Starcic, A. I. (2019). Human learning and learning analytics in the age of
of patenting and licensing by US universities: an assessment of the effects of artificial intelligence. Br. J. Educ. Technol. 50, 2974–2976. doi: 10.1111/bjet.
the Bayh–Dole act of 1980. Res. Pol. 30, 99–119. doi: 10.1515/97808047963 12879
61-008 The International Learning Sciences Forum (2019). The International Learning
Nye, B. D. (2015). Intelligent tutoring systems by and for the developing world: a Sciences Forum: International Trends for Ai and Big Data in Learning Sciences.
review of trends and approaches for educational technology in a global context. Taipei: National Taiwan Normal University.
Int. J. Art. Intell. Educ. 25, 177–203. doi: 10.1007/s40593-014-0028-6 Toh, L. P. E., Causo, A., Tzuo, P. W., Chen, I. M., and Yeo, S. H. (2016). A review
Obermeyer, Z., Powers, B., Vogeli, C., and Mullainathan, S. (2019). Dissecting racial on the use of robots in education and young children. J. Educ. Technol. Soc. 19,
bias in an algorithm used to manage the health of populations. Science 366, 148–163.
447–453. doi: 10.1126/science.aax2342 Topol, E. J. (2019). High-performance medicine: the convergence of human and
O’Donnell, C. (2008). Defining, conceptualizing, and measuring fidelity of artificial intelligence. Nat. Med. 25, 44–56. doi: 10.1038/s41591-018-0300-7
implementation and its relationship to outcomes in K-12 curriculum Tsai, C. C. (2000). Relationships between student scientific epistemological beliefs
intervention research. Rev. Educ. Res. 78, 33–84. doi: 10.3102/0034654 and perceptions of constructivist learning environments. Educ. Res. 42, 193–
307313793 205. doi: 10.1080/001318800363836
Papamitsiou, Z., and Economides, A. A. (2014). Learning analytics and educational Tsai, S. C., Chen, C. H., Shiao, Y. T., Ciou, J. S., and Wu, T. N. (2020).
data mining in practice: a systematic literature review of empirical evidence. Precision education with statistical learning and deep learning: a case study in
Educ. Technol. Soc. 17, 49–64. Taiwan. Int. J. Educ. Technol. Higher Educ. 17, 1–13. doi: 10.1186/s41239-020-
Pardo, A., and Siemens, G. (2014). Ethical and privacy principles for learning 00186-2
analytics. Br. J. Educ. Technol. 45, 438–450. doi: 10.1111/bjet.12152 UNESCO (2015). SDG4-Education 2030, Incheon Declaration (ID) and Framework
Pedró, F., Subosa, M., Rivas, A., and Valverde, P. (2019). Artificial Intelligence in for Action. For the Implementation of Sustainable Development Goal 4, Ensure
Education: Challenges and Opportunities for Sustainable Development. Paris: Inclusive and Equitable Quality Education and Promote Lifelong Learning
UNESCO. Opportunities for All, ED-2016/WS/28. London: UNESCO
Peters, M. A. (2018). Deep learning, education and the final stage of automation. United Nations (2020). Policy Brief: Education During Covid-19 and Beyond.
Educ. Phil. Theory 50, 549–553. doi: 10.1080/00131857.2017.1348928 New York, NY: United Nations
Popenici, S. A., and Kerr, S. (2017). Exploring the impact of artificial intelligence on VanRullen, R. (2017). Perception science in the age of deep neural networks. Front.
teaching and learning in higher education. Res. Pract. Technol. Enhanced Learn. Psychol. 8:142. doi: 10.3389/fpsyg.2017.00142
12:22. doi: 10.1186/s41039-017-0062-8 Viberg, O., Hatakka, M., Bälter, O., and Mavroudi, A. (2018). The current landscape
Quadir, B., Chen, N.-S., and Isaias, P. (2020). Analyzing the educational goals, of learning analytics in higher education. Comput. Human Behav. 89, 98–110.
problems and techniques used in educational big data research from 2010 to doi: 10.1016/j.chb.2018.07.027
2018. Int. Learn. Environ. 1–17. doi: 10.1080/10494820.2020.1712427 Williams, P. (2019). Does competency-based education with blockchain signal a
Renz, A., and Hilbig, R. (2020). Prerequisites for artificial intelligence in further new mission for universities? J. Higher Educ. Pol. Manag. 41, 104–117. doi:
education: identification of drivers, barriers, and business models of educational 10.1080/1360080x.2018.1520491
technology companies. Int. J. Educ. Technol. Higher Educ. 17:14. doi: 10.1186/ World Development and Report (2019). The Changing Nature of Work.
s41239-020-00193-3 Washington, DC: The World Bank/International Bank for Reconstruction and
Renz, A., Krishnaraja, S., and Gronau, E. (2020). Demystification of artificial Development.
intelligence in education–how much ai is really in the educational technology? Xie, H., Chu, H.-C., Hwang, G.-J., and Wang, C.-C. (2019). Trends and
Int. J. Learn. Anal. Art. Intell. Educ. (IJAI). 2, 4–30. doi: 10.3991/ijai.v2i1.12675 development in technology-enhanced adaptive/personalized learning: a

Frontiers in Psychology | www.frontiersin.org 10 October 2020 | Volume 11 | Article 580820

Luan et al. Education Big data and AI

systematic review of journal publications from 2007 to 2017. Comp. Educ. Conflict of Interest: JG was employed by company Apprendis, LLC, Berlin.
140:103599. doi: 10.1016/j.compedu.2019.103599
Yadegaridehkordi, E., Noor, N. F. B. M., Ayub, M. N. B., Affal, H. B., and The remaining authors declare that the research was conducted in the absence of
Hussin, N. B. (2019). Affective computing in education: a systematic review any commercial or financial relationships that could be construed as a potential
and future research. Comp. Educ. 142:103649. doi: 10.1016/j.compedu.2019.10 conflict of interest.
3649
Yarkoni, T., and Westfall, J. (2017). Choosing prediction over explanation in Copyright © 2020 Luan, Geczy, Lai, Gobert, Yang, Ogata, Baltes, Guerra, Li and Tsai.
psychology: lessons from machine learning. Perspect. Psychol. Sci. 12, 1100– This is an open-access article distributed under the terms of the Creative Commons
1122. doi: 10.1177/1745691617693393 Attribution License (CC BY). The use, distribution or reproduction in other forums
Zawacki-Richter, O., Marín, V. I., Bond, M., and Gouverneur, F. (2019). Systematic is permitted, provided the original author(s) and the copyright owner(s) are credited
review of research on artificial intelligence applications in higher education– and that the original publication in this journal is cited, in accordance with accepted
where are the educators? Int. J. Educ. Technol. Higher Educ. 16:39. doi: 10.1186/ academic practice. No use, distribution or reproduction is permitted which does not
s41239-019-0171-0 comply with these terms.

Frontiers in Psychology | www.frontiersin.org 11 October 2020 | Volume 11 | Article 580820