1 Designing Education Technologies in the age of AI – A learning sciences driven approach

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.
https://doi.org/10.1111/bjet.12861
By improving the understanding of how humans learn, the paper now aims to teach and develop AI for use in
education and training. They use 3 case studies to illustrate how learning sciences research will develop judicial
analysis of multimodal data, such that it will be helpful for scaffold students as well as teachers. Using AI
technology will help us leverage faster, more nuanced and enable individualized scaffolding for learners. This
paper argues about the role of learning sciences (psychology, sociology, computer science, education and
cognitive sciences.) in the design of educational AI, through case studies the author illustrates how learning
sciences research can power AI educational technology. They also provide a framework for inter-stakeholder,
interdisciplinary partnerships that is aimed at helping educators to understand and develop AI in educational
field.
The paper begins with an introduction to learning sciences field followed by 3 empirical case studies. Then they
mention a framework for AI educational technology development, where they focus on data, evidence, and
research.
 Finally, they conclude with the potential of AI in educational field. The author also, talks about inter stakeholder
and interdisciplinary partnerships that enhances AI’s benefits.
2 Artificial Intelligence in Education: A Review

L. Chen, P. Chen and Z. Lin, "Artificial Intelligence in Education: A Review," in IEEE Access, vol. 8, pp. 75264-
75278, 2020, Doi: 10.1109/ACCESS.2020.2988510.
This study talks about how AI is improving learners experience and overall quality of learning. It talks about
how AI is impacting decision making, learning and adaptability capabilities. It explores the use of humanoid
robots and web-based chatbots to automate lame tasks. Using AI, instructors have been able to perform
administrative functions efficiently and effectively, which include but are not limited to grading students’
assignments, intuitive teaching with high quality materials. He also, talks about how ML algorithms have
leveraged adaptability and the curriculum content has been customized and personalized for students. This has
fostered retention and interest of the students and thus enhancing the overall quality of learning.
3 Accelerated move for AI Education in China

Yang, X. (2019). Accelerated Move for AI Education in China. ECNU Review of Education, 2(3), 347–352.
https://doi.org/10.1177/2096531119878590
This paper mentions recent developments in applications of AI in Chinese education industry. It reviews
government policies and suggests a course for future progress in this domain. The paper discusses how Xi
Jinping’s government is working towards the elevation of AI in the field of highly strategic national import of
Artificial Intelligence technologies. The paper tells 260 experts were hired on for a 4-year job who revised
standard high school curriculum in order to incorporate AI as per country’s standards. The Ministry of
 Education has ensured that the revised IT curriculum is no longer focused on computers and the Internet, instead
it is focusing on algorithms, information systems, and the information society. The 2 credit AI course is
equivalent to 36 classroom hours, which enables students to learn about history and development of AI
algorithms. Student develop their own simple intelligent applications, gain hands-on experience with the
process of designing and creating intelligent systems to serve human development. According to the paper,
China had 32 colleges of Ai until 2018 with the aim of training professionals specializing in AI.
The paper also discusses how Ministry of Science and Technology is collaborating with industries to popularize
AI. National platforms in partnerships with companies like BAIDU for Autonomous Driving, Tencent for
Medical Imaging, Ali Baba for City Brain, iFlytek for Voice Recognition and many more.
Education enterprises are actively utilizing AI and other technologies for building smart campus which use Big
Data intelligence. They have focused on raising general IT attainment rather than focusing only on students and
professionals. A popular example is of school credit bank system that uses innovative blockchain technology
for making education sector flexible and customized. AI is actively used in voice recognition, gesture
recognition, facial expression recognition, and brain wave recognition. Overall, the paper concludes with the
positive sentiment that AI is bringing in the education sector of China.
4 Deep Learning goes to school- towards a relation understanding of AI in education
Citations: Deep learning goes to school: toward a relational understanding of AI in education. (2020).
Retrieved September 25, 2021, from Learning, Media and Technology website:
https://www.tandfonline.com/doi/citedby/10.1080/17439884.2020.1686017?scroll=top&needAccess=true
In this paper, the author actively talks about neural networks and advance AI techniques that are used to
automate human tasks. The experiment undertaken by the researcher deals with adding AI elements to Online
learning websites such as Khan Academy, Google, Fun toot, and more. The paper evaluates scope of AI in
Science and Technology Studies, by taking emphasizing on Weak AI- learning from data and its aspects. It
explores the dependence of Weak AI algorithms on data sets and exact algorithm implementation in order to
produce acceptable results. They also explore RNN and the disagreements and confusions involved in
 educational data science. The paper discusses the effect of applying RNN based algorithms on datasets obtained
from Khan Academy, Fun toot, ASSIST, and other sources. Finally, they conclude with potential of AI and its
complications, that are hindering in the adoption of this capable technology.
5 AI in medical Education
Masters K. (2019). Artificial intelligence in medical education. Medical teacher, 41(9), 976–980.
https://doi.org/10.1080/0142159X.2019.1595557
This guide introduces broad concepts and applications of AI on medicine and how its implications impact
educators. According to the paper, AI will not be able to replace the doctor, but it can definitely replace the
doctor’s role and develop new roles. The paper talks about AI replacing teachers and AI assisting doctors for
roles like minor surgeries, predicting patient’s health, predicting the outcome of various medications depending
on the patient and thus delivering a personalized prescription. AI can cater to salient issues like appalling
medical experiments, decline empathy levels among students, drug price gauging, predicting hospital, wards,
beds schedules, and much more. The paper also talks about advance AI methods, which involves utilizing AI
along with Robotic technology, to create a fully functional surgical doctor, that can be allotted to every patient.
It can efficiently monitor and predict patients’ health. This data can be further analyzed by students in order to
facilitate researches and discover new trends in medical field. In addition, the paper also talks about E-Patients,
being proactive in AI system design and drawbacks associated with the same.
6 A Holistic Approach to the Design of Artificial Intelligence (AI) Education for K-12 Schools
Chiu, T.K.F. A Holistic Approach to the Design of Artificial Intelligence (AI) Education for K-12 Schools.
TechTrends 65, 796–807 (2021). https://doi.org/10.1007/s11528-021-00637-1
This study analyses data collected from 12 K-12 schools and develops a model for designing curriculum which
includes content, produce, process, praxis. The data includes individual interviews, teaching documents, and
much more. Based on the experiments, the author has proposed a model that aims at holistic development of
AI curriculum for K-12 Schools. He focusses on Student relevance in terms of authenticity and local
understanding, teacher student communication in terms of consistent terminology and visual representation,
 flexibility in terms of student requirements and school environment, knowledge in AI in terms of definition and
development, process in AI in terms of perception and technical skills, and impact of AI in terms of social and
ethical impact.
7 An effective analysis of online education model based on AI
This paper is an effort in analyzing the education model that works on AI principles. The paper proposes a
learner model known as the Neighbor Mean Variation Multi-Objective Particle Swarm Optimization-Generic
Algorithm (NMVMOPSO-GA). This model is a combination of 3 models which include:
• Learner’s interest sub model
• Learner’s cognitive ability sub-model
• Learner’s knowledge sub-model

This algorithm helps in improving the performance and stability of the online course resource generation
method.
 Figure 1 Algorithm flow
8 An investigation of barriers to Hong Kong K-12 schools incorporating AI in education

9 Analysis of Worldwide Research trends on the impact of AI in Education

Paek S, Kim N. Analysis of Worldwide Research Trends on the Impact of Artificial Intelligence in Education.
Sustainability. 2021; 13(14):7941. https://doi.org/10.3390/su13147941
This study is aimed at exploring the future direction of education and the impact of AI along with research
trends and collaboration status country wise, through network analysis, global research trends and topic
 modelling. This study used Latent Dirichlet Allocation algorithm. The paper explains the ubiquitous nature of
AI in today’s world, and how it is transforming the society. Be it online education, medical services,
automobiles, manufacturing, finance, media, agriculture, communication, energy, or distribution. The paper
explains AI and Education by citing major use cases such as ‘AI for All’. The author has divided the use of AI
in Education in 3 parts:
o Use of Intelligent tutoring system (ITS): Step-by-step learning path for structured knowledge, includes-
domain models, learner models, pedagogical models.
o Dialogue-based Tutoring System: Engages students in dialog of learning by using NLP and NLP
Generation.
o Exploratory learning environment: provides automatic feedback to modify erroneous learning outcomes.

10 AI in education Big Data: Black Boxes and Technological Solutionism

Giró Gràcia, Sancho-Gil, X. J. (2020). Artificial Intelligence in Education: Big Data, Black Boxes, and
Technological Solutionism. Artificial Intelligence in Education: Big Data, Black Boxes, and Technological
Solutionism, Vol 17(No 2), ISSN: 1504–4831. https://doi.org/10.7577/seminar.4281
This article introduces the concepts of AI, Big Data, Machine Learning, Algorithms, and how they are deployed
as black boxes for generating a positive impact on the education. The paper toured the way of freeing teachers
 and students from lectures, textbooks, and recitation. The paper aims at making readers aware of how big data
is fueling the use of Artificial Intelligence algorithms. The paper also introduces with Black Box algorithms,
using case studies like Google Maps, Netflix suggestions, Apple’s Siri, YouTube’s Content ID system.
The paper then, mentions the scope of AI in education and learning. Education is more about learning to know,
learning to do, learning to live together, learning to live with others, discovering others, working towards
common goals, and learning through life; and AI can provide an immersive learning experience to students,
developing their cognitive abilities.
11 AI in education – the 3 paradigms
Fan Ouyang, Pengcheng Jiao, Artificial intelligence in education: The three paradigms, Computers and
Education: Artificial Intelligence, Volume 2, 2021, 100020, ISSN 2666-920X,
https://doi.org/10.1016/j.caeai.2021.100020.
This paper explores the contribution of AI in 3 paradigms: AI- Directed learner as recipient, AI-Supported
learner as collaborator, AI empowered learner as leader. The paper explains how AI has opened new
opportunities, potentials, and challenges in educational domain. The paper explores the possibilities of AI, it
has provided personalized learning, challenged instructor’s role, and developed complex educational systems.
Coupled with NLP, ANN, ML, DL, and genetic algorithms AI has been able to create intelligent learning
environments that can detect behavior, predict model building, and provide learning recommendation. The
paper’s main focus is on different roles of AI in education and to what extent it can influence learning and
instruction.
12 AI in education- Using heart rate variability (HRV) as a biomarker to assess emotions objectively
Joanne Wai Yee Chung, Henry Chi Fuk So, Marcy Ming Tak Choi, Vincent Chun Man Yan, Thomas Kwok
Shing Wong, Artificial Intelligence in education: Using heart rate variability (HRV) as a biomarker to assess
emotions objectively, Computers and Education: Artificial Intelligence, Volume 2, 2021, 100011, ISSN 2666-
920X, https://doi.org/10.1016/j.caeai.2021.100011.
In this paper, the author tries to assess the emotions of joy and sorrow among the students, in order to develop
an emotion sensor based on AI. Random videos were projected to mentally stable and fit adults with no history
of health problems and mood disorders. Based on these random videos, their perspective, emotion ratings and
HRV were monitored using a Polar watch. The model had a sensitivity of 70.7% for identifying happiness and
58.4% sensitivity for identifying sadness. Statistical analysis was made using 2-way ANOVA and factors such
as BPM, HF, LF, SDNN, SDSD RMSSD, pNN50, were analyzed. Based on this, many classification algorithms
were used such as SVM, NN, LDA and more. Based on all these models, sensitivity and specificity was
calculated. Finally, it was concluded that PLS-DA algorithm was suitable for emotion classification. It gave an
overall sensitivity of 64.54% compared to other algorithms. Using this algorithm, the author was able to classify
student’s emotions as happy or sad using student’s actions, HRV and behavior.
13 A ML Based Recommender system for improving students learning experiences
 Yanes, N., Mostafa, A.M., Ezz, M., & Almuayqil, S. (2020). A Machine Learning-Based Recommender System
for Improving Students Learning Experiences. IEEE Access, 8, 201218-201235.
In this paper, a recommendation engine is used with different machine learning algorithms. It is proposed for
predicting suitable actions based on the factors such as course specifications, academic records, course learning
outcomes and assessments. They use a multi-label multi-class binary dataset and then after translating the
dataset into different problems, such as label powerset, classifier chain, binary relevance, they used a ML-KNN
classifier to derive the output. This hybrid recommendation approach uses student’s average score in specific
modules of the course and based on the features, the best suitable action for each taught course is recommended
based on overall student academic records in the concerned course. Some of the parameters include, number of
students in each section, course credit hours, number of topics in each course, grade distribution, course learning
outcomes, and total number of questions. Data is filtered, missing values are handled, and actions are mapped
to classes. Different classifiers like the OVA classifier, BR classifier, LP Classifier, CC classifier, and ML-
KNN adaptive classifier are tested on the dataset. Analysis is done and based on that important feature are
determined. Based upon the analysis it is determined on which topics more work need to be executed.
14 Challenges and future directions of Big Data and AI in Education
Luan, H., Geczy, P., Lai, H., Gobert, J., Yang, S. J. H., Ogata, H., … Tsai, C.-C. (2020). Challenges and Future
Directions of Big Data and Artificial Intelligence in Education. Frontiers in Psychology, 11.
https://doi.org/10.3389/fpsyg.2020.580820
The purpose of this paper is to reflect present status, opportunities, and challenges that Big Data and AI are
facing in education industry. It covers three types of major challenges which include Research, Policy making
and Industry based. For each type of challenge, they provide some possible solutions.
The paper mentions about the enormous potential of AI and Big Data, and how they can be utilized to provide
a highly effective learning and teaching experience. The paper explains, how AI and Big Data can help in policy
making such as making micro-credit and micro-degrees that address specific subject and not the domain as a
whole. Big Data can help in commercializing intelligent educational tools and systems, which can provide
educators with tools to design curricula and assessments more effectively and efficiently. The author talks about
freely available tools like the MOODLE, LMS, whose use peaked during the COVID period. Similarly, it
proposes solutions for integrating cutting-edge research findings, innovative theory-driven curricula, greater
practical experience, a suitable legislative framework to protect personal data from unscrupulous collection,
unauthorized disclosure, and commercial exploitation. Some interesting developments that big data and AI can
 bring in education can include use of AR/VR technology which helps in fast-development of concepts,
algorithms and are suitable for personalized learnings.
15 Out of the laboratory and into the classroom – the future of AI in education
Schiff, D. Out of the laboratory and into the classroom: the future of artificial intelligence in education. AI &
Soc 36, 331–348 (2021). https://doi.org/10.1007/s00146-020-01033-8
This paper focuses on the status of AIEd with respect to intelligent tutoring systems as well as
anthropomorphized artificial educational agents. The paper talks about AI’s ability to simulate teachers, provide
robust student differentiation, foster socio-emotional engagement, AI enabled digital distance education, and
more. The paper then also talks about the downsides of AI in education industry. It explores the possible reason
for which AI has not been able to reproduce teacher-student interaction, why massive open online education
has not been able to provide high quality universal education.
The main focus of the paper is intelligent tutoring systems and how AIEd can penetrate into educational
systems. The paper explores the sensational history of education technology, the technological efforts made in
order to provide interactive learning to the students, it also talks about immensely popular Massive Open Online
Courses (MOOCs), and about the shift from educational infrastructure towards skill based online education,
and further credential-based education in physical spaces. The paper talks how MOOCs failed to explore new
markets, and why employers and students were not able abandon traditional degrees. The paper also discusses
about implementation barriers, technological linearity, teacher student interaction, and differentiation in AIEd.
Main focus of the paper lies in futures of AIEd which include Pedagogy and bias, Curriculum emphasis on
STEM subjects, Teachers’ role and displacement, international development and premature automation,
flexibility and ownership of educational systems, nudging and manipulation and the responsibilities of an AIED
researcher. The paper finally concludes with uncertainty and side effects involved. At the same time, he
encourages, role of AIEd in automating teachers’ tasks, automating curriculum planning.
31 Online Student Authentication and Proctoring System Based on Multimodal Biometrics Technology
M. Labayen, R. Vea, J. Flórez, N. Aginako and B. Sierra, "Online Student Authentication and Proctoring System
Based on Multimodal Biometrics Technology," in IEEE Access, vol. 9, pp. 72398-72411, 2021, doi:
10.1109/ACCESS.2021.3079375.
With an increase in online education platforms, even though institutions have managed to deliver information
through digital means effectively, still the industries have not been able to develop a method for accrediting
and certifying students. The process of verification of students is too complex through digital means, and this
paper focusses on devising a methodology, that uses different biometric technologies along with automated
proctoring system in order to get a credible certification platform. The methodology is highly scalable,
autonomous, affordable, reliable, passive for students as well as least hardware and software dependent. It is
based on constant biometric recognition system as well as automatic continuous proctoring using automatic
image and audio processing. The proposed solution uses face, voice, typing pattern as a biometric verification
method, while device monitoring, lock-down and inappropriate behavior detection as physical proctoring
means. The solution is cloud hosted, that makes it scalable and free from any geographical boundations.
The paper explains, even though online proctoring idea dates back to 2006, the solutions offered are still not
scalable, reliable, authentic, and biometric.
The processing core of the proposed system is composed of an image processing and audio processing
mechanism. Both of these help in authentication and proctoring process. In addition to this, a Human
verification module will also be integrated as initial check of the student.
The system was tested on more than 57 activities in 5 different e-learning institutions. The focus was laid on
technical capabilities of a system and a user experience satisfaction matrix. Results demonstrate the system has
high precision with 100% accuracy in final result report. All the false positive and false negatives were driven
to human cross verification in order to enhance the accuracy. The system achieved a 6.01/7 rating on Likert
scale for the UX test.
 None of the technical shortcomings exist in the proposed solution, however the functionality of the system can
be enhanced further by incorporating more robust biometric models that avoid undesirable deviations due to
face poses, light, noise conditions, etc.
32 Academic Performance Prediction Based on Multisource, Multifeatured Behavioral Data
L. Zhao et al., "Academic Performance Prediction Based on Multisource, Multifeature Behavioral Data," in
IEEE Access, vol. 9, pp. 5453-5465, 2021, doi: 10.1109/ACCESS.2020.3002791.
The paper proposes a model named as Augmented Education (“AugmentED”). It aims to solve challenges such
as combining student’s digital data from disparate sources in order to obtain holistic view of a student. At the
same time, utilize the data to accurately predict the academic performance and thereby using these predictions
to promote positive student-university relationship.

Academic performance prediction is considered a classification problem in this paper, and thus by using
classification-based algorithms, this system is divided into 3 parts i.e.
 the data module: Collects behavioral raw data from various sources, processes it to a feature extraction sub
module, which delivers feature using algorithms like BC-Linear, BC-Nonlinear, and BC-LSTM. Further it is
10-fold cross-validated by dividing the dataset into 7:3 ratio.
the prediction and feedback module: AugmentED is executed on major algorithms like RF, GBRT, KNN, SVM,
and XG Boost. Indexes like accuracy, precision, recall, f1 score, and AUC are used to evaluate the performance
of the models. Based on the results it is concluded that the minimum precision of AugmentedED is no less than
86.3%. It is also concluded that the difference between accuracy obtained by executing different algorithms is
negligible, hence AugmentedED appears to be independent of ML algorithms.
The paper concludes by summarizing the need for a prediction model in universities. The model imports data
from a variety of sources and not just academic activities, such as campus life behaviors inside and outside the
classrooms. The feature evaluation is done by linear, LSTM, nonlinear methods that provide a systematic
prediction.
The drawbacks of this study include that it is based on limited dataset that deals with passive data point capture
system.
33 Creating a Recommender System to Support Higher Education Students in the Subject Enrollment
Decision
A. J. Fernández-García, R. Rodríguez-Echeverría, J. C. Preciado, J. M. C. Manzano and F. Sánchez-Figueroa,
"Creating a Recommender System to Support Higher Education Students in the Subject Enrollment Decision,"
in IEEE Access, vol. 8, pp. 189069-189088, 2020, doi: 10.1109/ACCESS.2020.3031572.
This study introduces an approach that predicts the dropout risk or students’ performances as well as helps both
students and educational institutions by assisting students in selecting their subjects. The experiment is done by
taking data from a public Spanish university. The paper initially talks about the importance of maximizing
graduation rates, and why a Recommendation system is necessary in course recommendation to students, so as
to limit the dropout rates. Next the paper also talks about difficulty in gathering quality data as universities
don’t collect clean data and also sharing personal data is illegal in some situations. The main innovations
discussed in the paper are about utilizing small datasets to deliver precise output, using recommendation system
rather than predictive models, and using Data Mining Techniques. The algorithms used to create a
recommendation system are Decision Tree, Random Forest, Gradient Boosting Classifier, Logistic Regression,
SVM, K Nearest neighbors, and multi-layer perceptron. These algorithms along with encoded, feature
engineered, scaled and resampled data is fed to the training model, based on which appropriate model is selected
from a list of models.
 This selected model is then utilized to generate recommendations. The models have been evaluated based on
F1-Scores. The recommendation model was able to achieve 78.5% of accuracy for the first-year students, while
in some cases, the accuracy enhanced to 83.5%.
34 Design and Evaluation of a Deep Learning Recommendation Based Augmented Reality System for
Teaching Programming and Computational Thinking

P. Lin and S. Chen, "Design and Evaluation of a Deep Learning Recommendation Based Augmented Reality
System for Teaching Programming and Computational Thinking," in IEEE Access, vol. 8, pp. 45689-45699,
2020, doi: 10.1109/ACCESS.2020.2977679.

The paper talks about developing a Deep Learning based recommendation system, that uses AR technology
along with learning theory to help student learn and apply programming concepts. Most teachers teach
programming unilaterally, which makes students think it to be a difficult area to understand, thereby making it
more difficult to implement the concepts. The paper proposes AR technology to dynamically overlap digital
information with real time objects that provides a context-aware learning environment for writing and
compiling program codes in a problem-solving manner. In the experimental study, it was found the students
belonging to experimental group performed better in the dimensions of computational thinking, creativity,
logical computing, critical thinking and problem-solving skills compared to those in the control group. The
students in experimental group had a much more immersive experience in the context-aware learning
environment.
The paper claims in their survey that Deep learning analytic engine like the Google TensorFlow can predict the
development of students after study according to their academic performance by processing their subject
performance in subjects like History, Chinese, English, and their non-study performance such as learning
behavior, sports, arts, learning services, and much more. Deep learning is one of the illustrious technologies for
analysis, classification, and predictions. Deep learning can be used for a variety of tasks such as high energy
physics, data processing, image and speech recognition, automatic detection of retinopathy, computational
chemistry, analyzing skin disorders, translating handwritten text, MOOCS video analysis, and much more.
Deep Learning mechanisms can also be used to predict individual student who would discontinue their studies,
by analyzing his video watching pattern.
The paper explains four main dimensions of Computational thinking ability i.e., Decomposition, Pattern
recognition, Pattern Generalization and Abstraction, and Algorithm Design.
The experiment performed by the author on 97 non-IT students is named as “Program Logic Thinking
Education” that focuses on exploring the differences between students’ learning outcomes and computation
thinking. It provides a basic programming experience to students, so as to develop their computational skills in
the process of program compilation.
For preparing the platform, Unity 3D was used as a cross platform AR game engine, while the recommendation
function was built on Unity Machine Learning Agents. A recurrent neural network was adopted to model
dynamic and sequence data that can achieve supervised learning path recommendations. Firebase was used for
backend development on Cloud.
The Deep Learning recommendation-based AR system is composed of 2 parts, Teacher (Account, Materials,
Missions, Learning Portfolio) and Student (Know-How, We’re Pro, Common term, Exam). For the student
operated part, the activity flow was from Log In-> Reading -> Missions -> Exam -> Log Out.
 Based on the experiment, T test was done on the outcomes, to obtain a result. Based on the analysis, it was
concluded that, AR system is easy to operate, the interface is clean and easy to use, the system can be used
intuitively, and they felt satisfied with their ability to solve the problem. The paper concludes with a positive
sentiment among students for the system. A major research gap in this experiment was that the students were
taught only programming concepts. The results of this experiment can vary, if some other subject is opted. Also,
Arts students are more comfortable in analyzing shapes and real-world entities, thus for non-Art students, results
are not very positive.

Uma Gaur et al. [16] in their review paper they examine the challenges and opportunities faced by medical schools
in implementing remote learning for basic science teaching in response to the COVID-19 crisis. They searched
relevant literature in PubMed, Scopus, and Google Scholar using specific keywords, e.g., “COVID-19 pandemic,”
“preclinical medical education,” “online learning,” “remote learning,” “challenges,” and “opportunities.” The
pandemic has posed several challenges to premedical education (e.g., suspension of face-to-face teaching, lack of
cadaveric dissections, and practical/laboratory sessions) but has provided many opportunities as well, such as the
incorporation of online learning in the curriculum and upskilling and reskilling in new technologies. To date,
many medical schools have successfully transitioned their educational environment to emergency remote teaching
and assessments. During COVID-19 crisis, the preclinical phase of medical curricula has successfully introduced
the novel culture of “online home learning” using technology-oriented innovations, which may extend to post-
COVID era to maintain teaching and learning in medical education. However, the lack of hands-on training in
the preclinical years may have serious implications on the training of the current cohort of students, and they may
struggle later in the clinical years. The use of emergent technology (e.g., artificial intelligence for adaptive
learning, virtual simulation, and telehealth) for education is most likely to be indispensable components of the
transformative change and post-COVID medical education.
Xuesong Zha et al. [17] in this paper they reviewed AI in education research from 2010 to 2020. It is found that
the research to date could be classified into three dimensions: the dimension of development including
classification, matching, recommendation, and deep learning; the dimension of an extraction involving feedback,
reasoning, and adaptive learning; and the dimension of application including affection computing, role-playing,
immersive learning, and gamification. Moreover, based on the research questions and the related AI techniques,
four research trends were identified. They are the Internet of things, swarm intelligence, deep learning, and
neuroscience, as well as an assessment of the effect of AI in education. The challenges of AI in education were
also conspicuously seen in terms of technique perspective, teachers’ and students’ roles, and social ethical issues.
These findings could be valuable references for educational researchers, students, and AI developers who plan to
contribute to the relevant studies. Furthermore, it seems clear that educators need to work with AI engine
Chong Guan et al. [18] In this study they sought to examine the past research themes and trends in the field of AI
and DL applications in education via a comprehensive analysis of the literature published in reputable journals
over the last two decades. They utilize an automated method (web-scrapping) to retrieve and compile journal
articles on major publishers’ databases, including Association for Computing Machinery (ACM), EBSCO,
Emerald, IEEE, JSTOR, ScienceDirect, Taylor & Francis, and Wiley.
The following criteria were used to define the studies to be included in the review:
1. A focus on AI and/or DL application.
2. Deployment of AI/DL in education context.
3. Being published in an influential academic journal.
This research surveyed over 400 journal articles in the domain of AI and DL in education published in reputable
journals from 2000 to 2019. With the aim to uncover the history of research evolution in the field of AIEd, four
main research paradigm shifts have been identified in accordance with the changing role of AI technology in
education. With the prevalence of distant education, 2000e2009 saw an increase of research outputs in
implementation and design of online education in schools. VR emerged as a sub theme under this strand of
research as a technology to bring vivid experiences to classrooms. Then when big data came onto the scene,
student profiling models and learning analytics that supported personalized learning design became a dominant
focus in 2010e2019. With such temporal bibliometric analysis based on multiple-journals of AIEd research in the
past two decades, our research contributes to the AIEd literature by exploring major research paradigms in the
history, and the lifecycles of different research paradigm shifts. It helps to position on-going AIEd research in the
literature. The study also serves as a spur to initiate dialogue regarding the up-and-coming AIEd research themes
in the next decade. Despite our effort to make the coverage of this research as comprehensive as possible, their
research is limited in the following aspects. They use the density of publications to construct the concepts map
and major research paradigms. In this sense, their method is constrained to what have merged or happened in the
field. This can be an issue especially in predicting future research trends. For themes that have just emerged and
have yet diffused in the field, considering the lag of 1e2 year from research idea generation to publication, they
might not be spotted as a possible emerging area. Secondly, we have focused on AIEd literature to study the AI
technology in education. However, it is important to realize that EdTech is not only about technology, but also
about the pedagogical, cultural, social, economic, ethical and psychological dimensions of education. In this sense,
AIEd research is cross-disciplinary in nature and any major development in the above disciplines may induce a
paradigm shift in AIEd research. Hence, for future studies, reviews on AIEd research should include publications
in these disciplines as well.

Stefan A. D. Popenici and Sharon Kerr [19] in their paper explores the phenomena of the emergence of the use
of artificial intelligence in teaching and learning in higher education. It investigates educational implications of
emerging technologies on the way students learn and how institutions teach and evolve. Recent technological
advancements and the increasing speed of adopting new technologies in higher education are explored in order
to predict the future nature of higher education in a world where artificial intelligence is part of the fabric of our
universities. They pinpoint some challenges for institutions of higher education and student learning in the
adoption of these technologies for teaching, learning, student support, and administration and explore further
directions for research.
Dr. Share Aiyed M Aldosari [20] in his paper discussed the potential effects of artificial intelligence on higher
education at Prince Sattam Bin Abdulaziz University. To achieve this goal, a qualitative research methodology
was used by asking an open question on a sample of academics. The results of the analysis showed that there is a
decrease in the level of awareness of the mechanisms of applying artificial intelligence, and that there is a need
to further spread awareness in The Saudi environment on the possibilities of using artificial intelligence
applications in education. The researcher recommends universities' leaders to do these following steps:
-Increasing awareness among specialists of the requirements of applying artificial intelligence in education
-Attention to the challenges arising from the applications of artificial intelligence to the jobs and roles of
academics in universities
- Organize academic conferences on all AI issues.
- Create community cooperation with specialized companies in AI field. Finally, the researcher encourages
academic researchers to conduct more studies about the future of all levels of education in the light of AI
transformations.

Steven A. Wartman et al. [21] in this paper they showed us collaborating with and managing artificial intelligence
(AI) applications that aggregate big data, generate diagnostic and treatment recommendations, and assign
confidence ratings to those recommendations. Thus, an overhaul of medical school curricula is due and should
focus on knowledge management (rather than information acquisition), effective use of AI, improved
communication, and empathy cultivation. Accordingly, they advocate new curricula that respond to the challenges
of AI while being less detrimental to learners’ mental health. These curricula should emphasize 4 major features:
1. Knowledge capture, not knowledge retention;
2. Collaboration with and management of AI applications;
3. A better understanding of probabilities and how to apply them meaningfully in clinical decision making with
patients and families; and
4. The cultivation of empathy and compassion.
Hui Luan et al. [22] In this paper they have discussed the new challenges and directions facing the use of big data
and artificial intelligence (AI) in education research, policy-making, and industry. The purpose of this position
paper is to present current status, opportunities, and challenges of big data and AI in education. The work has
originated from the opinions and panel discussion minutes of an international conference on big data and AI in
education (The International Learning Sciences Forum, 2019), where prominent researchers and experts from
different disciplines such as education, psychology, data science, AI, and cognitive neuroscience, etc., exchanged
their knowledge and ideas. This article is organized as follows: they start with an overview of recent progress of
big data and AI in education. Then we present the major challenges and emerging trends. Finally, based on their
discussions of big data and AI in education, conclusion and future scope are suggested.

Yuying Gu et al. [23] their paper used the context of artificial intelligence, has used the convolutional neural
network structure in deep learning to establish an online teaching gesture recognition model, which can be used
to identify five kinds of gestures that teachers will use in online teaching situations. The recognition efficiency
and accuracy of the model are very high. The model can be extended to analyze the effect and quality of gestures
in teaching and the improvement of the performance of educational robots in human-computer interaction, which
has high practicability. However, at present, the model cannot be used to identify symbolic gestures and
metaphorical gestures that need to be classified according to teachers’ language or actual teaching situation.
Therefore, how to increase the recognition function of these two kinds of gestures on the basis of the existing
model is the focus of the following work.
Jagadeesh Kengam [24] in this paper have described how Artificial Intelligence can be used and is being used
in Educational sector. According to the 21st International Conference on Artificial Intelligence in Education held
in 2020, AIED is one of the currently emerging fields in Educational technologies. The use of AI is still
unclear for the educators how to make pedagogical advantage of it on a broader scale and how AI can
impact on teaching and learning in higher education. The impact of AI in education and its pros and cons are
presented here. It also describes a specific way to develop AI enabled platform for education.
Wojciech Samek et al. [25] In this paper they approached the problem of explain ability in artificial intelligence.
It was discussed why black box models are not acceptable for certain applications, e.g., in the medical domain
where wrong decisions of the system can be very harmful. Furthermore, explain ability was presented as
prerequisite for solving legal questions which are arising with the increased usage of AI systems, e.g., how to
assign responsibility in case of system failure. Since the “right to explanation” has become part of the European
law, it can be expected that it will also greatly foster explain ability in AI systems. Besides being a gateway
between AI and society, explain ability is also a powerful tool for detecting flaws in the model and biases in the
data, for verifying predictions, for improving models, and finally for gaining new insights into the problem at
hand.
Francisco-Javier Hinojo-Lucena et al. [26] in this paper, they used a methodology for bibliometric studies
following the parameters set by the PRISMA declaration (Preferred Reporting Items for Systematic reviews and
Meta-Analyses) . For this, the focus of attention was placed on the metadata of the scientific production in the
last 10 years (2007–2017), as collected in the databases with the greatest impact in social science: Web of Science
(WOS) and Scopus . Thus, first of all, the search equation “Artificial Intelligence” AND “Higher Education” was
established, which was introduced in the WOS (Journal Citation Reports impact index—JCR) and Scopus
(SCImago Journal & Country Rank impact index—SJR) databases. Afterwards, the results were narrowed down
by scientific discipline (education, educational research) and years of production (2007–2017). It should be noted
that the keywords “artificial intelligence” and “higher education” are included in the ERIC Thesaurus as approved
descriptors and used in Education Sciences. In the same way, the union with the Boolean operator “AND” gives
rigor to the search carried out
Lihua Peng [28] in his paper studied the development of online education and the environment and learning
variables of distance education. Based on this, the learning process of online education is modeled and the
functional structure is designed. The overall structure of the learner model NMVMOPSO-GA in the online
education system is introduced. This article only studies the characteristics of learning resources from their
original attributes, but the degree of intelligence of automatic semantic annotation of learning resources will also
have an important impact on learning resource recommendation services. Therefore, using information
technology to carry out data-driven personalized learning resource recommendation service is also a problem
worthy of in-depth study.
Ying Pei et al. [29] in their paper analyzes the data mining technology in cloud environment. Firstly, the time
series analysis model of big data information flow in hybrid cloud environment is constructed to analyze the data
structure, and then the high-dimensional phase space of data information flow in hybrid cloud environment is
reconstructed. In the reconstructed phase space, the association rules are extracted, and the extracted features are
used as pheromones to guide data location mining, so as to improve the data mining algorithm. In order to improve
the data orientation and mining ability in the mixed cloud environment, a data flow mining algorithm based on
phase space reconstruction of data information flow and association rule feature extraction in the mixed cloud
environment was proposed. The time series analysis model of big data information flow in the mixed cloud
environment was built, and association rule features were extracted in the reconstructed phase space. The
extracted features were used as pheromones to guide data location mining, and the improvement of data mining
algorithm was realized. This paper constructs a set of intelligent education platform and explores an online
teaching model based on intelligent question-answering robot. Based on the intelligent question- answering
system, students can ask questions, discuss and study with each other, thus stimulating the ability of students to
study independently. By participating in the Q&A, teachers can timely understand the learning situation of
students, grasp the depth of students' understanding of the problems, find out the difficult problems, analyze the
deficiencies in the teaching process, so as to adjust the teaching content and teaching methods in time. With the
help of this intelligent question- answering system, it is helpful to form a new teacher-student cooperative teaching
mode, and realize the transformation from teacher teaching to teacher guidance, students learn from each other
and cooperate in learning.

Yuxuan Sun et al. [30] In this paper they have presented a recurrent neural network to quickly build a text
readability evaluation model. The model consists of bi-directional gated recurrent unit (bi-GRU) and attention
layer. It can predict the readability level of the sentences or paragraphs. The weight of the attention layer can
finely locate the distribution of difficulty in the comprehension of the sentences. They train the model on rough
leveled reading materials without tremendous reading comprehension tests. Their experiments are performed on
various text materials. Compared with the popular readability formulas, the results show. They basically create
an attention-based bi-GRU method for readability assessment. Instead of consuming a lot of man- power and
material resources to build a hand-crafted formula, Their method can quickly build a text readability measurement
model from rough leveled reading materials. The proposed model can evaluate all kinds of texts. Visualization of
attention weights can be used to locate where the difficult part is in a sentence. It is helpful to key highlight, article
modification and text readability improvement in many language applications.