EURASIA Journal of Mathematics, Science and Technology Education, 2024, 20(7), em2478

ISSN:1305-8223 (online)
OPEN ACCESS Review Article https://doi.org/10.29333/ejmste/14762

A systematic review of artificial intelligence in mathematics education:

The emergence of 4IR
Oluwaseyi Aina Gbolade Opesemowo 1* , Habeeb Omoponle Adewuyi 1
1 University of Johannesburg, Johannesburg, SOUTH AFRICA

Received 11 March 2024 ▪ Accepted 29 May 2024

Abstract
The integration of artificial intelligence (AI) in mathematics education, focusing on its implications
in the 4th Industrial Revolution (4IR) era. Through a comprehensive analysis of 10 relevant studies
in Scopus and Google Scholar from 2015 to 2023, this review identifies the research methods,
research instruments, participants, and AI tools used in mathematics education. Some key ideas
include using AI-driven personalized learning and enhanced mathematics instruction, real-time
assessment and feedback, curriculum development, and empowering educators, which were
highlighted. The study aligns with the preferred reporting items for systematic reviews and meta-
analysis. Based on the analysis, most studies reviewed utilized qualitative research methods. The
study indicates that questionnaires were mainly used to gather data from students and teachers
who were the most significant participants in the reviewed papers. Further results revealed that
ChatGPT were the primary AI tool used in mathematics education, among other AI tools, as
identified in this review. Additionally, this review discusses the transformative potential of AI in
addressing educational disparities and preparing learners for the demands of 4IR.
Keywords: artificial intelligence, artificial intelligence tools, mathematics education, 4th Industrial
Revolution, chatbot

enabling them to perform tasks that typically require

INTRODUCTION human intelligence, such as visual perception, speech
Artificial intelligence (AI) in mathematics education recognition, decision-making, and translation between
can revolutionize how students learn and interact with languages (Hoffmann, 2022). Technology’s rapid and
complex concepts. Using AI, teachers can provide steady growth in the 21st century has resulted in
personalized learning experiences tailored to each disruptive transformations in various areas. 4th
student’s needs and abilities. Subsequently, AI can Industrial Revolution (4IR) has emerged because of the
analyze student performance data to identify areas of convergence of digital technology, data analytics, and
weakness and provide additional support and resources AI, altering how we live, work, and learn (Adelana et al.,
to help students improve their understanding of 2023; Moloi & Marwala, 2023; Opesemowo & Ndlovu,
mathematical concepts. With AI technology, the 2023; Rambe & Maime, 2023). Education is one of the
possibilities for engaging and enhancing mathematical critical areas undergoing this shift, particularly in the
education are endless. AI is arguably the driving teaching and learning of mathematics. Since the
technological force of the first half of this century and beginning of human history, mathematics has been an
will transform virtually every industry, if not human essential discipline, contributing to scientific and
endeavors at large (Davenport et al., 2020; Makridakis, technological advancements, cognitive growth, and
2017). Businesses and governments worldwide are logical reasoning. Structures and principles inherent to
spending massive amounts of money on AI (Dwivedi et AI have created a framework for understanding
al., 2021), and education cannot be left out in this regard. complicated events and comprehending our world. In
AI incorporates the computational understanding this era of technological progress and abundant
and imitation of intelligent behavior in machines, information, the ability to read, interpret, and apply
mathematical concepts is more vital than ever.
© 2024 by the authors; licensee Modestum. This article is an open access article distributed under the terms and conditions of
the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/).
oopesemowo@uj.ac.za (*Correspondence) ahabeeb@uj.ac.za
Opesemowo & Adewuyi / A systematic review of artificial intelligence in mathematics education

Contribution to the literature

• This research seeks to unravel the research method most adopted for AI in mathematics education.
• Another contribution of this study is identification of the various research instruments utilized for data
collection.
• Through the systematic review, this study contributes to provide valuable information for educators and
policymakers to infuse AI technology into mathematics.

4IR is distinguished by the convergence of AI can provide real-time assessment and feedback
technology that blurs the distinctions between the (Hooda et al., 2022), allowing students to track their
physical, digital, and biological worlds. AI is at the progress and find areas for growth rapidly. This
center of this revolution, together with other immediate feedback loop develops a more profound
technologies such as the Internet of things and big data knowledge of mathematical ideas and will enable
(Bhat & Huang, 2021; Lee et al., 2022; Misra et al., 2022; students to overcome misconceptions quickly. In
Sharma et al., 2021; Tien, 2017). It has resulted in addition, AI technology can provide recommendations
automation, data-driven decision-making, and the for additional resources and practice materials based on
development of intelligent systems that can learn and each student’s strengths and weaknesses (Alam, 2023).
adapt over time. These improvements have far-reaching This ensures that students can access the right tools and
ramifications for education, necessitating a change from materials to enhance their mathematical skills further.
traditional educational approaches to more adaptable Moreover, AI can also assist teachers in identifying
and individualized ones. Similarly, 4IR’s fusion of struggling students and intervening early to provide
technologies requires an educational paradigm shift extra support and assistance (Srinivasa et al., 2022). By
aligned with its dynamic and interconnected nature. In leveraging AI in mathematics education, students can
this scenario, AI appears as a powerful instrument become more actively engaged in their learning journey
capable of supplementing, refining, and reinventing and achieve better results. However, none of these
mathematics education, encouraging a more significant reviews provides a detailed analysis of how AI has been
grasp and appreciation for the subject. deployed in mathematics education, which this study
The emergence of 4IR has been characterized by the delved into. The study aimed to fill this gap in the
introduction of AI, which involves algorithms literature by examining the research method adopted for
(Oosthuizen, 2022). Recently, the attention of AI has been AI integration into mathematics education. It explored
shifted to teaching mathematics education, which will the various research instruments utilized for data
aid the teaching and learning process of mathematics. If collection, the respondents, and AI tools used to teach
AI is deployed adequately in teaching and learning mathematics education. This study seeks to provide
mathematics education, it will positively boost the deeper and valuable insights for educators and
mathematics performance of the learners. National policymakers to incorporate AI technologies into
Research Council of the USA argued that mathematics is mathematics education and inform future research and
vital for various tasks, including simulation practice in the field by conducting a comprehensive
constructions, statistical data analysis, and expressing analysis.
and applying quantitative relationships. Mathematics
education perfectly matches AI integration because of its Research Questions
structured character and the increasing demand for
This study systematically analyzed and synthesize
mathematics abilities in a data-driven economy (Singh et
empirical published articles focusing on AI in
al., 2022). AI plays vital roles in mathematics education,
mathematics education in the classroom. For relevance,
such as personalized learning, real-time assessment and
this study was restricted to articles dealing with AI in
feedback, curriculum enhancement, and empowering
mathematics education at the elementary, secondary,
educators.
and university school levels. In light of this, being
Personalized learning is essential to mathematics grounded in mathematics starts at the elementary school
education since it allows students to learn at their own level (Hill & Seah, 2023, 2017; Rittle-Johnson, 2017).
pace and concentrate on areas requiring extra assistance Hence, the research questions for this study were:
(Yu et al., 2022; Zhilmagambetova et al., 2023).
1. What is research method adopted for AI in
Algorithms can monitor students’ performance and
mathematics education?
personalize the learning experience to their specific
needs using AI integration (Fernandes et al., 2023; 2. What kind of research instruments are used for
(Opesemowo & Adekomaya, 2024). This improves data collection?
students’ understanding of mathematical ideas and 3. Who are respondents?
increases their confidence and motivation. Furthermore,

2 / 11
 EURASIA J Math Sci Tech Ed, 2024, 20(7), em2478

4. What kind of AI tools were used in identified Quality Assessment

papers?
To ascertain quality assessment, the authors deleted
all articles not published between 2000 and 2015, none-
METHODOLOGY journal articles, articles not focusing on AI, mathematics
education, and 4IR articles not published in English. The
Search Strategy authors transferred the identified papers’ titles,
The study adopted the advanced document search to abstracts, keywords, authors’ names, journal names, and
identify relevant and appropriate literature on AI in publication years to an MS Excel spreadsheet. The
mathematics education. The search strategy was article’s abstract was rigorously examined for purity,
fashioned using keywords such as “AI”, “mathematics rationality, and adherence to the established criteria.
education”, and “4IR” in the Scopus database and However, the four PRISMA review processes were not
Google Scholar. Initially, without setting the data distorted as criteria for the study. In addition, twenty-
parameters, 1,020 articles were found. The search was nine articles were removed based on duplication of
conducted on 11th July 2023 without specifying any records. The titles of articles were scrutinized, resulting
criteria, and a total of 562 papers comprising conference in the removal of twenty-nine articles because they were
papers, book chapters, books, conference reviews, not concentrating on the study’s keywords (i.e., AI,
conference proceedings, book series, editorials, mathematics education, and 4IR).
magazines, reports, lecture notes, errata, and articles in
press were expunged from the first search. Therefore, PRISMA Systematic Review Method
458 papers of peer-reviewed academic journal articles
documented in English and published from 2015 to 2023 Identification
emerged for selection criteria. The articles identified were sourced from the Scopus
database and Google Scholar database. From both
Selection Criteria databases, a total of 110 published papers were
Preferred reporting items for systematic reviews and discovered using the main construct or variables of the
meta-analysis (PRISMA) statement proposed by Moher study; that is, 90 articles were from the Scopus database,
et al. (2009) was adopted for sorting articles. In PRISMA while 20 articles were identified using Google Scholar.
statement, a 27-item evidence-based checklist, including
a four-phase workflow diagram, is provided for Screening
evaluating systematic reviews that have been published. In this method, we adopted several inclusion and
PRISMA aims to ensure consistency and accountability exclusion criteria. In ensuring proper scrutiny of the
that is devoid of bias when documenting the systematic review process, systematic review articles, books, book
literature analysis. The primary objective of the selection chapters, proceedings, or conference papers were not
criteria was to identify existing research and empirical included in the articles selected while focusing solely on
investigations of AI used in mathematics education. The AI, mathematics education, and 4IR published journal
following inclusion criteria were used to select the articles between 2015 to 2023. Nevertheless, countries or
papers reviewed for this study. These are studies on the regions, where the papers were published were not
use of AI in teaching and learning mathematics in considered for exclusion. At the end of the screening
schools, studies that describe AI tools used in process, authors focused on articles with at least one
mathematics education, and studies published in peer- mathematics reference. During the screening process, 23
reviewed journals written in English Language only and articles were spotted as not complying with the study’s
published between 2015 to 2023. However, the criteria requirements, while 29 papers were discovered as
were carefully chosen to ensure that the selected papers duplicates. In addition, there are only 58 articles left.
were relevant to the research topic and met a certain
level of rigor. By including studies on the use of AI in Eligibility
mathematics education, the research team aimed to
comprehensively understand the current landscape and In determining the eligibility of all the 110 articles
the potential impact of AI in the field. In addition, the that fit the study’s selection process, each article’s title,
requirement for peer-reviewed journal publications abstract, methodology, results, and discussion were
helped to ensure the reliability and accessibility of the thoroughly reviewed. At this junction, 100 journal
information gathered. By setting these specific criteria, articles were found to be culpable for not fully
the study strived to maintain a transparent and unbiased addressing AI in mathematics education in 4IR era. To
approach to analyzing the literature on AI in this end, 10 articles were selected in the final stage of the
mathematics education. systematic review process, as presented in Figure 1.

3 / 11
Opesemowo & Adewuyi / A systematic review of artificial intelligence in mathematics education

Figure 1. PRISMA flow diagram (Source: Authors’ own elaboration)

Inclusion & exclusion In addition, the publication stage of the selected

articles was considered. Finally, published articles were
For the inclusion criteria, articles published between
included, and articles in the press were excluded.
the year 2015 to 2023 were selected, while articles below
Therefore, 10 published articles were reviewed for this
2014 were excluded from the review. Also, articles
study.
published in the English language were included in the
review, whereas articles published using languages Table 1 shows list of the 10 reviewed journals in this
other than English were excommunicated. study.

Table 1. List of 10 reviewed journals in this study

Data Research
Reference Title Respondent AI tool
collection design
Büscher (2020) Scaling up qualitative Video N/A Professional Coding & transcribing
mathematics education research recording development
through artificial intelligence course
methods
Chiang (2021) Estimating the artificial Questionnaire Survey University Data envelopment
intelligence learning efficiency for students analysis &
civil engineer education: A case Mahalanobis distance
study in Taiwan approaches
Ferro et al. Gea2: A serious game for Questionnaire Quasi Students & Gae2
(2021) technology-enhanced learning in experimental teachers
STEM design
Flogie and Transdisciplinary approach of Questionnaire Experimental Students & Not specified
Aberšek (2015) science, technology, engineering teachers
and mathematics education
Forbus et al. Sketch worksheets in science, Not specified N/A Instructors Sketch worksheets
(2020) technology, engineering, and
mathematics classrooms: Two
deployments
Lee and Yeo Developing an AI-based chatbot Online form Design- Pre-service Chatbots
(2022) for practicing responsive teaching based schoolteachers
in mathematics research

4 / 11
 EURASIA J Math Sci Tech Ed, 2024, 20(7), em2478

Table 1 (Continued). List of 10 reviewed journals in this study

Data Research
Reference Title Respondent AI tool
collection design
Schindler and Students’ collaborative creative DUET & SRI Case study Graduate Dual eye-tracking
Lilienthal (2022) process and its phases in approach education
mathematics: An explorative students
study using dual eye tracking and
stimulated recall interviews
Schindler et al. Small number enumeration Eye-tricking Experimental 3rd to 5th grade Cluster algorithm
(2022) processes of deaf or hard-of- device & students
hearing students: A study using stimuli
eye tracking and artificial
intelligence
Sun et al. (2023) A theoretical framework for a Questionnaire Survey Primary & ISM
mathematical cognitive model for & interview secondary
adaptive learning systems teachers
Wardat et al. ChatGPT: A revolutionary tool for Triangulation Instrumental Students & Chatbots
(2023) teaching and learning study educators
mathematics

Figure 2. Methodology used by authors (Source: Authors’ Figure 3. Data collection analysis (Source: Authors’ own
own elaboration) elaboration)

RESULTS RQ2. What Kind of Research Instruments Are Used

for Data Collection?
The systematic analysis of the study conducted
between 2015 to 2023 on AI in mathematics education in Research question two presents the method of data
this paper was used to answer the research question. collection. It showed that most of the study sample
employed questionnaires (i.e., n=3: Chiang, 2021; Ferro
RQ1. What Is Research Method Adopted for AI in et al., 2021; Flogie & Aberšek, 2015). In addition, one
Mathematics Education? reviewed study used a questionnaire and interview (i.e.,
Sun et al., 2023) to collect data. Others used various
The first research question focused on the research methods of data collection, such as dual eye-tracking
method adopted for AI in mathematics education. The (DUET) and stimulated recall interviews (SRI) (Schindler
findings showed that the reviewed study used three et al., 2022), online form (Lee & Yeo, 2022) did not specify
research methodologies: quantitative, qualitative, and the method of data collection. In addition, a
mixed methods. The findings also demonstrate that 70% questionnaire was the most commonly used data
(n=7) of the examined study used qualitative research collection method (Figure 3).
method (Büscher, 2020; Ferro et al., 2021; Flogie &
Aberšek, 2015; Lee & Yeo, 2022; Schindler et al., 2022; AI in mathematics education in the new era of 4IR.
Wardat et al., 2023) indicating that majority of the These methods allow researchers to gather both
reviewed study utilized qualitative method. quantitative and qualitative data, providing a
Subsequently, it was uncovered that 20% (n=2) of the comprehensive understanding of the research topic.
examined study were quantitative research methods Questionnaires enable researchers to collect data from
(Chiang, 2021; Flogie & Aberšek, 2015). Lastly, the mixed many participants, while interviews allow in-depth
method was used with 10% (n=1) of the reviewed study exploration of participants’ experiences and
(Sun et al., 2023) (Figure 2). perspectives.

5 / 11
Opesemowo & Adewuyi / A systematic review of artificial intelligence in mathematics education

Figure 4. Sample respondents (Source: Authors’ own elaboration)

Including video recording as a data collection ensures that the perspectives of both groups are well
method suggests a focus on capturing non-verbal cues represented, providing a more well-rounded
and behaviors. Triangulation, on the other hand, understanding of the subject. Also, including graduate
involves using multiple data sources or methods to education students, instructors, and pre-service teachers
enhance the validity and reliability of findings. Using further inflates the scope of the research, allowing for in-
eye-tracking devices and stimuli indicates a focus on depth insights from individuals actively involved in the
studying visual attention and perception. Finally, field of education. Including various grade levels, such
combining DUET and SRI suggests a sophisticated as primary and secondary teachers and third to fifth-
approach to understanding participants’ thought grade students, provides a broader perspective on the
processes and decision-making. Overall, the variety of topic, considering the different experiences and
data collection methods employed in this study reflects challenges faced at various stages of education.
researchers’ commitment to capturing a comprehensive Moreover, the inclusion of university students and
range of data for analysis and interpretation. Also, it educators in the studies acknowledges the importance of
allows researchers to explore various aspects of AI higher education in shaping teaching practices and
implementation and its effects on teaching and learning educational approaches. Overall, these studies’ diverse
practices in mathematics education. range of respondents enhances the findings’ validity and
applicability, contributing to a more comprehensive
RQ3. Who Are Respondents? understanding of AI in mathematics education.
To address this research question, the articles (n =10)
RQ4: What Kind of AI Tools Were Used in Identified
used for this systematic study showed that data were
Papers?
collected from various respondents, as shown in Figure
4. The authors organized their data from different Research question four focuses on the various AI
respondents. The respondents range from graduate tools utilized by the author. AI tools range from specific
education students to university students. However, the methodologies, such as interpretive structural modeling
authors (Ferro et al., 2021; Flogie & Aberšek, 2015) used (ISM) and dual eye-tracing, to broader approaches,
students and teachers as their respondents, representing including ChatGPT and coding/transcribing. The
the highest number of respondents for this review. studies reviewed indicated that more authors used
In addition, Büscher (2020), Chiang (2021), Forbus et ChatGPT as a revolutionary AI tools for teaching and
al. (2020), Lee and Yeo (2022), Schindler et al. (2022); learning mathematics in schools. The study includes Lee
Schindler and Lilienthal (2022), Sun et al. (2023), and and Yeo (2022) and Wardat et al. (2023). Büscher (2020),
Wardat et al. (2023) indicated graduate education Chiang (2021), Ferro et al. (2021), Forbus et al. (2020),
students, instructors, pre-service teachers, primary and Schindler et al. (2022), Schindler and Lilienthal (2022),
secondary teachers, professional development course, and Sun et al. (2023) utilized Sketch worksheets, ISM,
students and educators, third to fifth grade students, and coding and transcribing, cluster algorithm, data
university students, respectively as respondents in their envelopment analysis and Mahalanobis distance
study. Subsequently, these studies included diverse approaches, DUET, and Gae2, respectively. One
respondents, allowing for a comprehensive topic reviewed study (Flogie & Aberšek, 2015) failed to specify
analysis. Including students and teachers as respondents AI tool used. Table 2 shows distribution of AI tools.

6 / 11
 EURASIA J Math Sci Tech Ed, 2024, 20(7), em2478

Table 2. Distribution of AI tools

Reference AI tools Frequency (n) Percentage (%)
Büscher (2020) Coding and transcribing 1 10
Chiang (2021) Data envelopment analysis & Mahalanobis distance 1 10
Ferro et al. (2021) Gae2 1 10
Flogie and Aberšek (2015) Not stated 1 10
Forbus et al. (2020) Sketch worksheets 1 10
Lee and Yeo (2022) ChatGPT 1 10
Schindler and Lilienthal (2022) Dual eye-tracking 1 10
Schindler et al. (2022) Cluster algorithm 1 10
Sun et al. (2023) ISM 1 10
Wardat et al. (2023) ChatGPT 1 10
Total 10 100

DISCUSSION methods approach could be particularly beneficial in

capturing the multifaceted nature of how AI is applied
The study focused on the review of AI in in the field. This would involve collecting quantitative
mathematics education: the emergence of 4IR. The first data on the effectiveness of AI tools and qualitative data
research question reviews the methodology adopted for on the experiences and perceptions of teachers and
AI in mathematics education by the sample studies. This students using these tools. By integrating these different
study reveals the three research methods (i.e., perspectives, researchers can gain a more nuanced
qualitative, quantitative, and mixed method research) understanding of the impact of AI on mathematics
were used in the review study. The study reviewed 10 education and make more informed recommendations
research-published articles. We discovered that the for future practice and research.
qualitative research method was the most prominent
Another main research question answered in this
methodology used by researchers. This result aligns
review was using the data collection method. It was
with Ogegbo and Ramnarain (2022), who also uncovered
uncovered that questionnaires are the foremost form of
in their systematic review of computational thinking in
data collection, among others. Questionnaires allow
science classrooms that qualitative research design was
researchers to gather large amounts of data from a wide
the most common research method used by authors. The
range of participants relatively efficiently. This method
prevalence of qualitative research methods in the
is often preferred for its ability to collect standardized
reviewed studies strongly emphasizes exploring
responses, making it easier to analyze and compare
subjective experiences, perspectives, and meanings
results across different studies. Arundel (2023) reported
related to the research topic. This approach allows for a
that questionnaire testing is essential for ensuring high-
deep understanding of the phenomenon under
quality data collection and achieving content validity,
investigation, providing rich and detailed insights that
inter-rater validity, and reliability. In addition,
may not be captured through quantitative methods
questionnaires can be easily distributed and completed
alone. This result opposes the findings of Mohamed et
by participants at their convenience, making them a
al. (2022) in a similar review study, which stated that
convenient option for researchers conducting studies
researchers often use the use of quantitative method
with limited resources or time constraints. However, it is
approach since it emphasizes the objective measurement
essential to note that questionnaires also have
and analysis of statistical, mathematical, or numerical
limitations. One major drawback is the potential for
data gathering via questionnaires and surveys for the
response bias ( Adewuyi & Oluwole, 2016; Wetzel et al.,
used of AI in mathematics education.
2016), where participants may provide inaccurate or
Subsequently, it is essential to note that qualitative socially desirable responses and can skew the data and
and quantitative methods have strengths and lead to misleading conclusions. Another limitation is the
limitations. While quantitative methods provide lack of depth in responses, as questionnaires typically
numerical data that can be easily analyzed and only remain valuable for researchers to gather large
compared, qualitative methods offer a more in-depth amounts of data efficiently.
understanding of the complexities and nuances of a
Further research questions were raised to identify the
particular phenomenon. They allow researchers to
respondents in the reviewed articles. It was detected that
gather context-specific information on human
various respondents served as participants in the
experiences, behavior, attitudes, and emotions within
reviewed articles. This includes graduate education
specific sociocultural contexts (Dzogovic & Bajrami,
students, instructors, pre-service schoolteachers,
2023; Onyemah & Omoponle, 2022). Therefore,
primary and secondary teachers, professional
combining both approaches can provide a more
development courses, students and educators, students
comprehensive and holistic view of the research topic. In
and teachers, 3rd to 5th grade students, and university
the context of AI in mathematics education, a mixed-

7 / 11
Opesemowo & Adewuyi / A systematic review of artificial intelligence in mathematics education

students. This study indicates that most of the authors of CONCLUSIONS & RECOMMENDATIONS
the reviewed articles used students and teachers in their
research. This finding is in tandem with a similar survey With AI driving 4IR, integrating AI into mathematics
by Kurowski et al. (2022), whose report focused on education represents a shift with immense potential.
primary school teachers, pupils, and parents. The Collaboration between educators, researchers, and
diverse range of respondents in the sample provides a technologists will be critical as we traverse this dynamic
comprehensive perspective on the effectiveness of the terrain to reshape the future in which AI empowers
reviewed articles across various educational levels. The learners and prepares them for the challenges and
emphasis on students and teachers in their studies aligns opportunities of 4IR as an innovation characterized as
with the importance of understanding the impact of the bedrock of any developed society.
educational interventions on these critical stakeholders. However, the path to this future is not without
The survey of Kurowski et al. (2022) further supports difficulties, including issues surrounding ethics,
the notion that involving primary school teachers, accessibility, and the striking balance between AI and
pupils, and parents is crucial in evaluating the success of human abilities. This investigation into AI’s role in
educational initiatives. mathematics education aims to reveal the way forward,
promoting collaboration and dialogue among educators,
The last research question that this study answered
policymakers, and technologists as we navigate 4IR’s
was to identify the various AI tools used in mathematics
undiscovered areas.
education. Emphasis on these AI tools was necessary to
understand how teachers improve the learning of By examining the intersection of AI and mathematics
mathematics education concepts. In this review study, education, we can uncover new ways to enhance
ChatGPT was spotted to have the highest percentage of learning experiences and equip students with the
AI tools, which accounted for 20%. necessary skills for success in a rapidly evolving digital
world. The potential benefits of integrating AI into
In contrast, sketch worksheets, ISM, coding and
mathematics education are enormous, from
transcribing, data envelopment analysis, Mahalanobis
personalized learning experiences to real-time feedback
distance approaches, and dual accounted for 10%,
and assessment. But AI can be used responsively if
respectively. Consequently, one reviewed paper did not
ethical considerations are carefully addressed and
disclose AI tool used. These results further showed how
equitably in educational settings. Also, ensuring
ChatGPT has contributed to AI in mathematics
accessibility for all learners, regardless of their
education. This result supports the study of Wardat et al.
background or abilities, is crucial in harnessing full
(2023), who acknowledged that using ChatGPT in
potential of AI in education. As we strive to balance AI
mathematics education has shown promise in
and human abilities in mathematics education,
improving learning outcomes and increasing student
collaboration and open dialogue will be critical in
engagement. It is also known that ChatGPT can enhance
shaping a future, where AI empowers learners to
mathematics capabilities and increase educational
triumph in 4IR and beyond.
success by providing learners with essential
mathematical knowledge on various topics. Limitations
However, several studies (Adiguzel et al., 2023; Deng
This systematic review study has its limitations, like
& Yu, 2023; Lee & Yeo, 2022; Wu & Yu, 2023) have
any other study. This review was restricted by the
explored using ChatGPT, such as ChatGPT, in teaching
selection criteria used for including relevant literature in
mathematics. Their studies have found that ChatGPT
mathematics education, possibly affecting the
can provide comprehensive instruction and assistance in
generalizability of this study. It is also important to note
geometry, increasing educational success by offering
that the findings of this study may not apply to all
learners basic mathematics knowledge.
educational settings or populations due to the specific
ChatGPT can provide students with personalized criteria used for selecting literature. Additionally, the
support and instant feedback, guiding them through scope of this study may not encompass all aspects of
complex mathematical problems and identifying areas, mathematics education, potentially limiting the depth of
where they may need additional assistance. By offering analysis and conclusions that can be drawn. Also, the
an interactive learning experience, ChatGPT helps study was constrained to AI, which was not written in
students build confidence in their mathematical skills, full. Despite these limitations, this study provides
ultimately leading to tremendous success in the subject. valuable insights and contributes to the existing
Also, using ChatGPT in mathematics education can knowledge in mathematics education. It is constrained
alleviate the burden on teachers, giving them more time in terms of number of reviewed publications, which may
to concentrate on other academic activities and impact the inferences of the results. But findings of this
providing additional support to students outside the study can still be used to inform future research and
classroom. practice in mathematics education.

8 / 11
 EURASIA J Math Sci Tech Ed, 2024, 20(7), em2478
Author contributions: All authors have sufficiently contributed to future of marketing. Journal of the Academy of
the study and agreed with the results and conclusions
Marketing Science, 48(1), 24-42. https://doi.org/
Funding: No funding source is reported for this study.
10.1007/s11747-019-00696-0
Ethical statement: The authors stated that the ethical issues were
adhered to, and secondary data were used in this study. By Deng, X., & Yu, Z. (2023). A meta-analysis and
utilizing secondary data, the study was able to build upon existing systematic review of the effect of chatbot
research and provide valuable insights into the topic at hand. technology use in sustainable education.
Therefore, the study results were more robust and reliable due to
the careful consideration of ethical issues and the use of secondary
Sustainability, 15(4), 2940. https://doi.org/10.3390
data sources. /su15042940
Declaration of interest: No conflict of interest is declared by the Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G.,
authors.
Coombs, C., Crick, T., Duan, Y., Dwivedi, R.,
Data sharing statement: Data supporting the findings and
Edwards, J., Eirug, A., Galanos, V., Ilavarasan, P.
conclusions are available upon request from the corresponding
author. V., Janssen, M., Jones, P., Kar, A. K., Kizgin, H.,
Kronemann, B., Lal, B., Lucini, B., . . ., & Williams,
REFERENCES M. D. (2021). Artificial intelligence (AI):
Multidisciplinary perspectives on emerging
Adelana, O. P., Ayanwale, M. A., Ishola, A. M., Oladejo, challenges, opportunities, and agenda for research,
A. I., & Adewuyi, H. O. (2023). Exploring pre- practice and policy. International Journal of
service teachers’ intention to use virtual reality: A Information Management, 57, 101994.
mixed method approach. Computers & Education: X https://doi.org/10.1016/j.ijinfomgt.2019.08.002
Reality, 3, 100045. https://doi.org/10.1016/j.cexr.
Dzogovic, A., & Bajrami, V. (2023). Qualitative research
2023.100045
methods in science and higher education. Journal
Adewuyi H. O., & Oluwole, D. A. (2016). Low teacher Human Research in Rehabilitation, 13(1), 156-166.
efficacy, poor parental involvement, truancy and https://doi.org/10.21554/hrr.042318
gender as indices of students mathematics
Fernandes, C. W., Rafatirad, S., & Sayadi, H. (2023).
underachievement in Osun State Nigeria. Journal of
Advancing personalized and adaptive learning
Applied Psychology, 15(1), 29-51.
experience in education with artificial intelligence.
Adiguzel, T., Kaya, M. H., & Cansu, F. K. (2023). In Proceedings of the 32nd Annual Conference of the
Revolutionizing education with AI: Exploring the European Association for Education in Electrical and
transformative potential of ChatGPT. Contemporary Information Engineering. https://doi.org/10.23919/
Educational Technology, 15(3), ep429. EAEEIE55804.2023.10181336
https://doi.org/10.30935/cedtech/13152
Ferro, L. S., Sapio, F., Terracina, A., Temperini, M., &
Alam, A. (2023). Harnessing the power of AI to create Mecella, M. (2021). Gea2: A serious game for
intelligent tutoring systems for enhanced technology-enhanced learning in STEM. IEEE
classroom experience and improved learning Transactions on Learning Technologies, 14(6), 723-739.
outcomes. In G. Rajakumar, K.-L. Du, & Á. Rocha https://doi.org/10.1109/TLT.2022.3143519
(Eds.), Intelligent communication technologies and
Flogie, A., & Aberšek, B. (2015). Transdisciplinary
virtual mobile networks. Springer. https://doi.org/
approach of science, technology, engineering and
10.1007/978-981-99-1767-9_42
mathematics education. Journal of Baltic Science
Arundel, A. (2023). How to design, implement, and analyze Education, 14(6), 779-790. https://doi.org/10.33225
a survey. Edward Elgar Publishing. https://doi.org /jbse/15.14.779
/10.4337/9781800376175.00008
Forbus, K., Garnier, B., Tikoff, B., Marko, W., Usher, M.,
Bhat, S. A., & Huang, N. F. (2021). Big data and AI & McLure, M. (2020). Sketch worksheets in science,
revolution in precision agriculture: Survey and technology, engineering, and mathematics
challenges. IEEE Access, 9, 110209-110222. classrooms: Two deployments. AI Magazine, 41(1),
https://doi.org/10.1109/ACCESS.2021.3102227 19-32. https://doi.org/10.1609/aimag.v41i1.5189
Büscher, C. (2020). Scaling up qualitative mathematics Hill, J. L., & Seah, W. T. (2023). Student values and
education research through artificial intelligence wellbeing in mathematics education: Perspectives
methods. For the Learning of Mathematics, 40(2), 2-7. of Chinese primary students. ZDM–Mathematics
https://www.jstor.org/stable/27091146 Education, 55(2), 385-398. https://doi.org/10.1007/
Chiang, T. (2021). Estimating the artificial intelligence s11858-022-01418-7
learning efficiency for civil engineer education: A Hoffmann, C. H. (2022). Is AI intelligent? An assessment
case study in Taiwan. Sustainability, 13(21), 11910. of artificial intelligence, 70 years after Turing.
https://doi.org/10.3390/su132111910 Technology in Society, 68, 101893. https://doi.org/
Davenport, T., Guha, A., Grewal, D., & Bressgott, T. 10.1016/j.techsoc.2022.101893
(2020). How artificial intelligence will change the

9 / 11
Opesemowo & Adewuyi / A systematic review of artificial intelligence in mathematics education

Hooda, M., Rana, C., Dahiya, O., Rizwan, A., & Hossain, Onyemah, T. N., & Omoponle, A. H. (2022). Child abuse
M. S. (2022). Artificial intelligence for assessment and family background as predictors of poor
and feedback to enhance student success in higher academic performance among adolescents in
education. Mathematical Problems in Engineering, special schools of Ibadan-Nigeria. Journal of
2022, 5215722. https://doi.org/10.1155/2022/ Lexicography and Terminology, 6(1), 106-119.
5215722 Oosthuizen, R. M. (2022). The Fourth Industrial
Kurowski, M., Černý, M., & Trapl, F. (2022). A review Revolution–Smart technology, artificial
study of research articles on the barriers to intelligence, robotics and algorithms: Industrial
inclusive education in primary schools. Journal on psychologists in future workplaces. Frontiers in
Efficiency and Responsibility in Education and Science, Artificial Intelligence, 5. https://doi.org/10.3389/
15(2), 116-130. https://doi.org/10.7160/eriesj. frai.2022.913168
2022.150206 Opesemowo, O. A. G., & Adekomaya, V. (2024).
Lee, D., & Yeo, S. (2022). Developing an AI-based chatbot Harnessing artificial intelligence for advancing
for practicing responsive teaching in mathematics. sustainable development goals in South Africa’s
Computers & Education, 191, 104646. https://doi.org higher education system: A qualitative study.
/10.1016/j.compedu.2022.104646 International Journal of Learning, Teaching and
Lee, J., Abid, A., Le Gall, F., & Song, J. (2022). Recent Educational Research, 23(3), 67-86. https://doi.org/
trends on artificial intelligence-enabled internet of 10.26803/ijlter.23.3.4
things platform and standard technologies. In Opesemowo, O. A. G., & Ndlovu, M. (2023). Status and
Proceedings of the 2022 IEEE 8th World Forum on experience of mathematics teachers’ perception of
Internet of Things. https://doi.org/10.1109/WF- integrating computer adaptive testing into unified
IoT54382.2022.10152226 tertiary matriculation examination mathematics.
Makridakis, S. (2017). The forthcoming artificial Multicultural Education, 9(2), 66-78.
intelligence (AI) revolution: Its impact on society Rambe, P., & Maime, R. (2023). The impact of 4IR
and firms. Futures, 90, 46-60. https://doi.org/10. technologies on venture creation and technology
1016/j.futures.2017.03.006 commercialization: Insights and exemplars from an
Misra, N. N., Dixit, Y., Al-Mallahi, A., Bhullar, M. S., emerging economy context. In R. E. Hinson, D. A.
Upadhyay, R., & Martynenko, A. (2022). IoT, big Odame, E. K. Adae, & K. Adom (Eds.), Small
data, and artificial intelligence in agriculture and business and entrepreneurial development in Africa: A
food industry. IEEE Internet of Things Journal, 9(9), route to sustained economic development (pp. 149-178).
6305-6324. https://doi.org/10.1109/JIOT.2020. Springer. https://doi.org/10.1007/978-3-031-
2998584 37675-7_7
Mohamed, M. Z. B., Hidayat, R., Suhaizi, N. N. B., Rittle-Johnson, B. (2017). Developing mathematics
Mahmud, M. K. H. B., & Baharuddin, S. N. B. knowledge. Child Development Perspectives, 11(3),
(2022). Artificial intelligence in mathematics 184-190. https://doi.org/10.1111/cdep.12229
education: A systematic literature review. Schindler, M., & Lilienthal, A. J. (2022). Students’
International Electronic Journal of Mathematics collaborative creative process and its phases in
Education, 17(3), em0694. https://doi.org/10.29333 mathematics: An explorative study using dual eye
/iejme/12132 tracking and stimulated recall interviews. ZDM–
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. Mathematics Education, 54(1), 163-178.
(2009). Preferred reporting items for systematic https://doi.org/10.1007/s11858-022-01327-9
reviews and meta-analyses: The PRISMA Schindler, M., Doderer, J. H., Simon, A. L., Schaffernicht,
statement. Annals of Internal Medicine, 151(4), 264- E., Lilienthal, A. J., & Schäfer, K. (2022). Small
269. https://doi.org/10.7326/0003-4819-151-4- number enumeration processes of deaf or hard-of-
200908180-00135 hearing students: A study using eye tracking and
Moloi, T., & Marwala, T. (2023). The Fourth Industrial artificial intelligence. Frontiers in Psychology, 13.
Revolution. In T. Moloi, & T. Marwala (Eds.), https://doi.org/10.3389/fpsyg.2022.909775
Enterprise risk management in the Fourth Industrial Sharma, D., Singh, A., & Singhal, S. (2021). The
Revolution (pp. 11-20). Springer. https://doi.org/ technological shift: AI in big data and IoT. In V.
10.1007/978-981-99-6307-2_2 Jain, P. Kumar, & V. Ponnusamy (Eds.) The smart
Ogegbo, A. A., & Ramnarain, U. (2022). A systematic cyber ecosystem for sustainable development (pp. 69-
review of computational thinking in science 90). Wiley. https://doi.org/10.1002/978111976165
classrooms. Studies in Science Education, 58(2), 203- 5.ch4
230. Singh, G., Pathak, S., Singh, J., & Tiwari, S. (2022).
https://doi.org/10.1080/03057267.2021.1963580 Implication of mathematics in data science

10 / 11
 EURASIA J Math Sci Tech Ed, 2024, 20(7), em2478

technology disciplines. In Proceedings of the 2022 Wetzel, E., Böhnke, J. R., & Brown, A. (2016). Response
IEEE Conference on Interdisciplinary Approaches in biases. In F. T. L. Leong, D. Bartram, F. Cheung, K.
Technology and Management for Social Innovation. F. Geisinger, & D. Iliescu (Eds.), The ITC
https://doi.org/10.1109/IATMSI56455.2022.10119 international handbook of testing and assessment (pp.
311 349-363). Oxford University Press. https://doi.org
Srinivasa, K. G., Kurni, M., & Saritha, K. (2022). /10.1093/med:psych/9780199356942.003.0024
Harnessing the power of AI to education. In K. G. Wu, R., & Yu, Z. (2023). Do AI chatbots improve students
Srinivasa, M. Kurni, & K. Saritha (Eds.), Learning, learning outcomes? Evidence from a meta-analysis.
teaching, and assessment methods for contemporary British Journal of Educational Technology, 55(1), 10-33.
learners: Pedagogy for the digital generation (pp. 311- https://doi.org/10.1111/bjet.13334
342). Springer. https://doi.org/10.1007/978-981- Yu, X., Xia, J., & Cheng, W. (2022). Prospects and
19-6734-4_13 challenges of equipping mathematics tutoring
Sun, S., Wu, X., & Xu, T. (2023). A theoretical framework systems with personalized learning strategies. In
for a mathematical cognitive model for adaptive Proceedings of the 2022 International Conference on
learning systems. Behavioral Sciences, 13(5), 406. Intelligent Education and Intelligent Research.
https://doi.org/10.3390/bs13050406 https://doi.org/10.1109/IEIR56323.2022.10050082
Tien, J. M. (2017). Internet of things, real-time decision Zhilmagambetova, R., Kopeyev, Z., Mubarakov, A., &
making, and artificial intelligence. Annals of Data Alimagambetova, A. (2023). The role of adaptive
Science, 4(2), 149-178. https://doi.org/10.1007/ personalized technologies in the learning process:
s40745-017-0112-5 Stepik as a tool for teaching mathematics.
Wardat, Y., Tashtoush, M. A., AlAli, R., & Jarrah, A. M. International Journal of Virtual and Personal Learning
(2023). ChatGPT: A revolutionary tool for teaching Environments, 13(1), 1-15. https://doi.org/10.4018/
and learning mathematics. EURASIA Journal of IJVPLE.324079
Mathematics, Science and Technology Education, 19(7),
em2286. https://doi.org/10.29333/ejmste/13272

https://www.ejmste.com

11 / 11