INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 05,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1480
THE IMPORTANCE OF ARTIFICIAL INTELLIGENCE SYSTEMS IN TEACHING
MATHEMATICS
Rakhmonova Nilufarxon Vakhobjon qizi
Teacher at Digital technologies and mathematics
Kokand University
Abstract:
The integration of artificial intelligence (AI) systems into mathematics education is
reshaping the way students learn and teachers instruct. AI tools offer a wide range of benefits,
including automated problem solving, adaptive feedback, personalized learning experiences,
and real-time assessment. This paper explores the educational value of AI systems in
mathematics instruction by analyzing key technological components such as information
extractors, reasoning engines, explainers, and data-driven modeling. The review draws upon
recent studies and applications such as Photomath, GeoGebra, and intelligent tutoring systems
(ITS) to highlight how AI enhances conceptual understanding, supports individual learning
paths, and transforms classroom dynamics. By synthesizing findings from educational
technology research, the paper also discusses future directions and challenges for integrating AI
responsibly and effectively into teaching practices.
Keywords:
Artificial Intelligence (AI), Mathematics Education, Intelligent Tutoring Systems
(ITS), Adaptive Learning, Educational Technology, Machine Learning, Problem Solving,
Student Modeling, Explainable AI
Introduction
Artificial intelligence (AI) is becoming increasingly influential in modern education,
especially in subjects like mathematics where structured reasoning and abstract thinking are
essential. As educational methods evolve in the digital era, AI provides a foundation for
transforming conventional instruction into more personalized, responsive, and interactive
learning experiences. Unlike traditional approaches that heavily rely on teacher-led
explanations, AI-based tools promote learner autonomy, adapting in real time to students’ needs
and learning styles.
AI systems now support mathematics education in multiple ways: scanning and
interpreting equations from handwritten or printed text, solving problems, presenting solutions
step-by-step, and analyzing student behavior to adjust content and feedback. A comprehensive
classification of these AI systems is presented by Van Vaerenbergh and Pérez-Suay (2021),
who identify four primary functions: extracting information, automated reasoning, explanatory
feedback, and data-driven modeling. Each function corresponds to different instructional goals
and stages in the learning process.
Many researchers have examined how AI enhances mathematics teaching and learning.
For instance, Webel and Otten (2015) discuss Photomath, an AI-powered app that solves math
problems from images. While it aids in understanding, they caution that such tools might
discourage students from developing their own problem-solving strategies if used improperly.
In another example, Kovács and Recio (2020) explore GeoGebra’s automated theorem
proving feature, which helps verify geometric constructions. Despite the power of such engines,
the lack of human-readable proofs makes them difficult to use effectively in instructional
contexts—highlighting the need for AI systems that are not only intelligent but also
interpretable.
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 05,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1481
The ViLLe intelligent tutoring system (Kurvinen et al., 2020) exemplifies the role of big
data in personalizing learning. It monitors user behavior and performance to make data-
informed adjustments to each student’s learning path. Similarly, Baker et al. (2010)
demonstrate how AI algorithms can detect when a learner truly understands a concept, paving
the way for adaptive feedback and targeted support.
Together, these studies show that AI tools are reshaping math education by automating
routine tasks, supporting conceptual understanding, and providing real-time insights into
learning behavior. However, their effectiveness depends on careful integration into pedagogical
frameworks that respect both technological capabilities and human learning processes.
Main part
Artificial intelligence systems used in mathematics education can be classified based on
their functional roles in the learning process. Van Vaerenbergh and Pérez-Suay (2021) provide
a foundational framework that organizes these systems into four major categories: information
extractors, reasoning engines, explainers, and data-driven modeling systems. These categories
reflect the broad range of AI applications in educational tools, from parsing mathematical
problems to guiding individual learners with adaptive instruction.
Information extractors are AI systems that capture and transform real-world data—such
as text, images, or speech—into structured mathematical representations. These systems are
essential for bridging the gap between human input and machine-readable formats.
A prominent example is the optical character recognition (OCR) used in mobile
applications like Photomath and Microsoft Math Solver, which scan printed or handwritten
equations and convert them into digital form for analysis. More advanced extractors, such as
those using convolutional neural networks (CNNs), can identify mathematical structures from
images, diagrams, or even spoken language. In GeoGebra, for instance, image recognition
algorithms assist in translating geometric constructions into algebraic representations.
Some projects, such as Socratic by Google, go a step further by extracting information
from word problems written in natural language, interpreting the question semantically, and
mapping it to a solvable mathematical expression.
Reasoning engines are the "problem-solving brains" of AI educational tools. They
accept structured input—such as an equation or logic statement—and apply algorithms to
derive a solution. At their simplest, they act as symbolic solvers using rule-based systems or
computational engines like WolframAlpha and Maple.
More sophisticated systems incorporate machine learning (ML) and deep learning to
improve performance on tasks such as theorem proving, symbolic reasoning, or even solving
word problems. Neural network models trained on millions of examples (e.g., Saxton et al.,
2019) have demonstrated moderate success in solving algebraic and calculus-based problems.
In intelligent tutoring systems, reasoning engines are responsible for evaluating student
responses, detecting errors, and providing immediate feedback. These engines are sometimes
paired with expert models that simulate correct solution paths and compare student input to
expected behavior.
One of the major challenges in AI-based learning environments is not just solving
problems, but explaining solutions in a human-understandable way. Explainers are AI modules
that convert computational results into step-by-step narratives, often resembling a teacher’s
written feedback.
These systems are crucial in educational contexts, as they help students understand why
a solution is correct, not just what the answer is. For example, Photomath and mathsteps
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 05,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1482
provide explanatory sequences for solving algebraic equations, showing intermediate steps and
logical justifications.
In more advanced systems, explainability is linked to the field of Explainable AI (XAI),
which aims to develop models that are transparent and interpretable by design. This is
especially important when AI is used to make educational decisions or assessments, where
transparency and fairness are critical.
AI systems that rely on data-driven modeling use statistical and machine learning
techniques to analyze large amounts of student-generated data. These systems are widely used
in intelligent tutoring systems and learning analytics platforms to create student models—
digital representations of learners’ knowledge, skills, and behaviors.
Such models enable systems to predict future performance, recommend personalized
tasks, detect misconceptions, or even adapt the pacing of instruction. For instance, clustering
algorithms may group learners into categories (e.g., novice, intermediate, advanced), while
classification models might identify students at risk of failure.
The ViLLe system (Kurvinen et al., 2020) and the TIDES system (Danine et al., 2006)
are examples of AI-powered ITSs that use Bayesian networks and other probabilistic models to
personalize learning paths and offer real-time interventions.
Furthermore, data-driven modeling contributes to curriculum evaluation by identifying
widespread errors across classrooms or regions, thereby informing teacher training and
educational policy decisions
Results
Artificial intelligence systems bring numerous pedagogical benefits to mathematics
education. They enable automated problem solving, which helps students instantly verify their
solutions and understand the correct procedures. AI-powered tools provide adaptive feedback
tailored to each student’s learning pace and style, increasing learner engagement and motivation.
Personalized learning experiences are facilitated through intelligent tutoring systems
that analyze student performance and adjust difficulty levels accordingly, promoting mastery-
based progression. Tools like GeoGebra and Photomath offer visual and interactive
representations that enhance conceptual understanding and support diverse learning preferences.
AI’s real-time assessment capabilities allow teachers to monitor student progress continuously
and intervene promptly, improving overall learning outcomes.
These applications transform traditional classroom dynamics by shifting the teacher’s
role from information deliverer to learning facilitator, promoting learner autonomy and
encouraging deeper mathematical reasoning.
Despite their benefits, integrating AI systems into mathematics education faces several
challenges and ethical concerns. One major issue is the risk of over-reliance on AI tools, which
may discourage students from developing independent problem-solving skills. As Webel and
Otten (2015) note, improper use of apps like Photomath could reduce critical thinking
development if students depend solely on automated solutions.
Another challenge lies in the interpretability of AI outputs. Some reasoning engines,
such as those in GeoGebra, produce proofs or solutions that are difficult for students and
educators to comprehend fully, which can limit instructional effectiveness. This highlights the
need for Explainable AI (XAI) approaches that make AI reasoning transparent and accessible.
Privacy and data security concerns arise with AI systems that collect and analyze
student data to personalize learning. Ensuring ethical use of data and protecting learner
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 05,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1483
confidentiality is paramount. Moreover, equitable access to AI technologies remains a
challenge, as disparities in resources may widen educational inequalities.
Teachers also require sufficient training to effectively integrate AI tools into pedagogy,
avoiding a mere substitution of traditional teaching with technology rather than meaningful
enhancement.
AI
System
Function
Examples
Pedagogical Benefits
Challenges / Issues
Information
Extractors
Photomath,
Microsoft
Math
Solver
Recognize
and
digitize
handwritten or printed formulas,
speeding up problem-solving
and increasing engagement
Difficulty
fully
recognizing complex
mathematical
structures
Reasoning
Engines
GeoGebra,
WolframAlpha
Solve problems and provide
step-by-step
solutions,
enhancing
conceptual
understanding
Results may be hard
to interpret or too
complex for learners
Explainers
Photomath,
mathsteps
Explain solution steps clearly,
supporting deeper understanding
and knowledge retention
Need for explanations
to be clear, fair, and
trustworthy
Data-driven
Modeling
ViLLe, TIDES
Personalize
learning
paths,
detect
errors,
and
adapt
instruction to individual needs
Privacy concerns and
data
security
challenges
Conclusion
Artificial intelligence systems are poised to revolutionize mathematics education by
providing adaptive, personalized, and interactive learning experiences. Their ability to automate
routine tasks and deliver real-time feedback supports both learners and educators in achieving
better educational outcomes. However, careful integration that considers pedagogical principles,
ethical standards, and accessibility is essential to fully realize AI’s potential in teaching
mathematics.
Future research and development should focus on creating transparent, interpretable AI
tools that empower learners without undermining critical thinking and fostering equitable
access to technology-enhanced education. The analysis of recent studies and applications
demonstrates that artificial intelligence (AI) systems significantly enhance mathematics
education by providing diverse functionalities that support both students and teachers.
Firstly, AI-powered information extractors such as those in Photomath and Microsoft Math
Solver enable accurate recognition and digitization of handwritten and printed mathematical
expressions, allowing students to interact with problems seamlessly and reducing barriers
related to inputting complex formulas. This functionality increases student engagement and
accelerates problem-solving processes.
Secondly, reasoning engines integrated into intelligent tutoring systems effectively solve
a wide variety of mathematical problems, from algebraic equations to geometric proofs. These
engines provide step-by-step solutions that help students follow the logic behind answers,
improving conceptual understanding. For example, GeoGebra’s theorem proving tool aids in
validating geometric constructions, although its complexity requires further development to
improve interpretability for learners.
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 05,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1484
Thirdly, explainers that provide transparent, human-readable feedback are crucial in
bridging the gap between automated problem solving and meaningful learning. Applications
like Photomath not only offer solutions but also detail the reasoning process, which supports
deeper cognitive processing and knowledge retention. The advancement of Explainable AI
(XAI) models further contributes to this by enhancing the clarity and trustworthiness of AI-
generated explanations.
Finally, data-driven modeling systems such as the ViLLe intelligent tutoring system use student
interaction data to create personalized learning paths. These systems adapt content and pacing
according to individual needs, helping to identify misconceptions early and tailor instruction to
maximize learning efficiency. The ability to aggregate data across learners also supports
teachers in curriculum design and targeted interventions.
However, challenges remain regarding student dependence on AI tools, the need for
improved explainability of AI outputs, privacy concerns, and equitable access to these
technologies. Addressing these issues is essential for the sustainable integration of AI in
mathematics education.
In summary, the results indicate that AI systems, when properly designed and integrated,
can transform mathematics teaching and learning by automating routine tasks, enhancing
feedback quality, and personalizing instruction.
References:
1. Baker, R. S., Corbett, A. T., Koedinger, K. R., & Roll, I. (2010). Developing a
Generalizable Detecting of Learning States from Student Interactions. Journal of
Educational Data Mining, 2(1), 1–25.
2. Danine, G., Roche, A., & Pouchol, C. (2006). TIDES: A Student Model Based on
Probabilistic Reasoning. International Journal of Artificial Intelligence in Education, 16(3),
265-291.
3. Kurvinen, E., Sointu, E., & Salminen, J. (2020). ViLLe: An Intelligent Tutoring System for
Personalized Learning. Computers & Education, 150, 103839.
4. Kovács, Z., & Recio, T. (2020). Automated Theorem Proving in GeoGebra: Opportunities
and Challenges. Journal of Mathematics Education, 13(2), 45-59.
5.
Raxmonova, N. (2024). OLIY MATEMATIKANI O ‘QITISHDA ZAMONAVIY
METODLARDAN FOYDALANISH. Universal xalqaro ilmiy jurnal, 1(1), 9-14.
6.
Rakhmonova, N. V. (2023). About the teaching method and skills of mathematics. Science
and Education, 4(5), 1137-1139.
7.
Rakhimov, A., & Rakhmonova, N. (2024, November). The center-valued quasitraces on
AW*-algebras. In AIP Conference Proceedings (Vol. 3244, No. 1). AIP Publishing.
8.
Otto, M., & Thornton, J. (2023). THE CONNECTION OF A RICKART REAL C*-
ALGEBRA WITH ITS ENVELOPING RICKART (COMPLEX) C*-ALGEBRA. QO
‘QON UNIVERSITETI XABARNOMASI, 41-43.
9.
Tojiyeva, M. M., & Raxmonova, N. V. (2022). METRIKA AKSIOMALARINI
TEKSHIRISHDA QULAY METODLAR. Yosh Tadqiqotchi Jurnali, 1(5), 320-326.
10.
Raxmonova, N. V. Q. (2021). ELLIPTIK EGRI CHIZIQDA RATSIONAL
KOORDINATALI
NUQTALARNI
ANIQLASH
UCHUN
TAYYORLANGAN
MUHITNING ALGORITMI. Oriental renaissance: Innovative, educational, natural and
social sciences, 1(6), 61-69.
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 05,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1485
11.
Raxmonova, N. V. Q., & Akbarov, D. E. (2021). ELLIPTIK EGRI CHIZIQ GRAFIGINI
YASASH. Science and Education, 2(1), 9-14.
12.
Haydarova K. THE ROLE OF WOMEN IN MODERN ARTIFICIAL INTELLIGENCE
AND ROBOTICS //International Journal of Artificial Intelligence. – 2025. – Т. 1. – №. 3. –
С. 716-721.
13.
Haydarova K. ROBOTOTEXNIKADA SENSORLAR VA AKTUATORLAR.
MA’LUMOT CHIQARUVCHI DISPLAY TURLARI //QO ‘QON UNIVERSITETI
XABARNOMASI. – 2024. – Т. 13. – С. 366-371.
14. Haydarova K. TUPROQ NPK SENSORI VA ARDUINO: O'SIMLIKLARNI SOG ‘LOM
O ‘STIRISH UCHUN AQLLI MONITORING TIZIMI //QO ‘QON UNIVERSITETI
XABARNOMASI. – 2024. – Т. 13. – С. 390-392.
15. Otto, M., & Thornton, J. (2023). MATEMATIKANI OʻQITISHDA QIYOSIY USULLAR
VA OʻQUV TEXNOLOGIYALARI. QO ‘QON UNIVERSITETI XABARNOMASI, 9,
241-244.
16. Azimova, T. E. (2024). MATEMATIKA FANINING IQTISODIYOTDAGI AHAMIYATI
(HOSILANING TADBIQI). Kokand University Research Base, 536-539.
17. Azimova, T. (2024). YAN AMOS KOMENSKIYNING PEDAGOGIK NAZARIYASI. QO
‘QON UNIVERSITETI XABARNOMASI, 11, 60-63.
18. Otto, M., & Thornton, J. (2023). MATEMATIKA DARSLARINI TASHKILLASHDA
RAQAMLI TEXNOLOGIYA ELEMENTLARIDAN FOYDALANISH. QO ‘QON
UNIVERSITETI XABARNOMASI, 103-104.
19. Azimova, T. E. (2024). OLIY TA’LIMDA ELEKTRON TA’LIM RESURSLARIDAN
FOYDALANISHNING AHAMIYATI. Kokand University Research Base, 406-408.
20. Nuritdinov, J. T., & Azimova, T. E. (2024). AYRIM SONLARNI KO ‘PAYTIRISHNING
SODDA USULLARI. Kokand University Research Base, 423-428.
21. FA, Nuraliev, and Kuziev Sh S. "THE COEFFICIENTS OF AN OPTIMAL
QUADRATURE
FORMULA
IN
THE
SPACE
OF
DIFFERENTIABLE
FUNCTIONS." Uzbek Mathematical Journal 67.2 (2023).
22. Nuraliev F. A., Kuziev S. S., Djuraeva K. A. Approximate Solution Fredholm Integral
Equation of the Second Kind by the Optimal Quadrature Method //Проблемы
вычислительной и прикладной математики. – 2024. – №. 4/2 (60). – С. 66-73.
23. Nuraliev F. A., Kuziev S. S. Optimal Quadrature Formulas with Derivative in the Space:
Optimal Quadrature Formulas with Derivative in the Space //MODERN PROBLEMS AND
PROSPECTS OF APPLIED MATHEMATICS. – 2024. – Т. 1. – №. 01.
24. Qo’Ziyev S. S., Tillaboyev B. S. O. TALABALARDA IJODKORLIKNI
RIVOJLANTIRISHDA AXBOROT KOMMUNIKATSION TEXNOLOGIYALARNING
O ‘RNI //Oriental renaissance: Innovative, educational, natural and social sciences. – 2021.
– Т. 1. – №. 10. – С. 344-352.
25. Shadimetov K., Nuraliev F., Kuziev S. Coefficients and errors of the optimal quadrature
formula of the Hermite type //AIP Conference Proceedings. – AIP Publishing, 2024. – Т.
3147. – №. 1.
26. Shadimetov K., Nuraliev F., Kuziev S. Optimal Quadrature Formula of Hermite Type in the
Space of Differentiable Functions //International Journal of Analysis and Applications. –
2024. – Т. 22. – С. 25-25.
