Artificial Intelligence in the Personalization of Teaching of Higher Mathematics Students in Kazakhstan

Abstract

Higher mathematics is a core component of STEM and engineering education; however, many students encounter difficulties in developing conceptual understanding and problem-solving skills. This study examines students’ use of artificial intelligence-based chatbots to support learning in higher mathematics within a personalized learning framework. The empirical basis of the study consists of a cross-sectional anonymous online survey conducted at Saken Seifullin Kazakh Agrotechnical University (Kazakhstan) (n = 154). The results of the descriptive analysis indicate that 72% of respondents reported an understanding of chatbot operating principles, 49% used chatbots “as needed,” and 7% reported very frequent use. The most common areas of chatbot application were the computation of derivatives and integrals (44%) and graph plotting (28%). Among the most frequently perceived benefits, respondents highlighted explanations of complex topics (45%) and rapid access to problem solutions (28%). At the same time, the main perceived limitations included insufficient accuracy of responses (42%) and difficulties in entering mathematical expressions (35%). Overall, the findings suggest that students primarily perceive chatbots as on-demand support tools for computational tasks and conceptual explanations, while concerns related to accuracy and usability remain significant barriers to their broader adoption in the educational process.

1. Introduction

In contemporary higher education, increasing attention is being devoted to personalized learning, particularly in the teaching of higher mathematics, where students frequently encounter conceptual challenges and require sustained and systematic practice. Personalized learning seeks to accommodate individual learner needs and to support motivation and academic progress by adapting instructional content, delivery modes, learning pace, and task complexity. In mathematically intensive courses, personalization is widely regarded as a promising approach for reducing learning barriers and enhancing learning outcomes through targeted instructional support and feedback (Bhutoria, 2022; Msomi & Mthethwa, 2024).

Recent advances in artificial intelligence (AI), especially in the area of generative AI, have enabled the development of interactive tools capable of providing on-demand explanations, step-by-step examples, and feedback-like support during problem-solving processes. In this study, artificial intelligence is understood as a field of computer science concerned with the design of systems and algorithms capable of performing tasks that typically require human intelligence, including natural language processing, decision-making, and learning from data (Russell & Norvig, 2010). Generative AI, as a subclass of AI technologies, employs large language models to generate new content (e.g., text) based on patterns learned from large-scale data sets (Vaswani et al., 2017; Brown et al., 2020). In educational contexts, generative AI is increasingly discussed as a means of supporting personalized learning resources and enhancing student interaction through the use of chatbots (Sajja et al., 2024; Schei et al., 2024; Wang et al., 2023; Ilieva et al., 2023; Merino-Campos, 2025).

The President of the Republic of Kazakhstan, Kassym-Jomart Tokayev, has emphasized the growing importance of artificial intelligence in modern society, highlighting its substantial potential to transform multiple domains, including education (Bhutoria, 2022).

In Kazakhstan, national initiatives aimed at the development and regulation of artificial intelligence have intensified in recent years. The President announced the establishment of a Ministry of Artificial Intelligence and Digital Development, signaling a strategic institutional focus on AI-driven digital transformation (President of the Republic of Kazakhstan, 2025). Furthermore, the Law of the Republic of Kazakhstan “On Artificial Intelligence” was signed on 17 November 2025 and entered into force on 18 January 2026, thereby establishing a legal framework and core principles for the development and use of AI systems (Republic of Kazakhstan, 2025). Alongside regulatory measures, Kazakhstan is also expanding its computing infrastructure to support AI and digital services, including the national supercomputer cluster Alem.Cloud (Ministry of Artificial Intelligence and Digital Development of the Republic of Kazakhstan, n.d.). Against this backdrop, it is both timely and relevant to examine how university students already employ AI-based chatbots to support learning in higher mathematics.

To ensure terminological consistency and avoid ambiguity, this study adopts standardized definitions of key concepts used throughout the manuscript. Personalized learning is defined as an educational approach in which learning content, pace, and instructional support are adapted to learners’ individual needs and characteristics (Bhutoria, 2022; Merino-Campos, 2025). In educational contexts, a chatbot is understood as a software agent with artificial intelligence capabilities that supports interactive dialogue and learning assistance, typically through natural language processing and generative language models (Brown et al., 2020; Russell & Norvig, 2010; Vaswani et al., 2017; Kuhail et al., 2023). Artificial intelligence in education (AIEd) refers to a broad class of digital technologies designed to support teaching and learning processes, including adaptive learning platforms, intelligent tutoring systems, assessment tools, and content-generation technologies (Chiu et al., 2023; Holmes et al., 2022; Zawacki-Richter et al., 2019). Adaptive learning systems are considered a subset of AIEd solutions that dynamically adjust learning tasks and trajectories based on data about learners and their performance (Gligorea et al., 2023).

Although intelligent educational systems are actively developing, the role of AI chatbots in supporting personalized learning in higher mathematics remains insufficiently explored, particularly in empirical studies examining students’ actual usage patterns and perceptions (Msomi & Mthethwa, 2024; Schei et al., 2024; Kuhail et al., 2023). Previous research has identified both potential opportunities, such as increased accessibility and interactive learning support, and persistent challenges, including issues related to accuracy, usability, and integration into formal instructional designs (Sajja et al., 2024; Merino-Campos, 2025). In the context of Kazakhstan, the limited number of context-specific empirical studies further underscores the relevance of the present research (Yilmaz et al., 2023; Omirali et al., 2025). Accordingly, this study investigates students’ self-reported experiences of using generative AI chatbots (e.g., ChatGPT) as support tools in the study of higher mathematics at Saken Seifullin Kazakh Agrotechnical University.

Aim of the Study

The aim of this study is to examine how students use AI chatbots to support learning in higher mathematics from a personalized learning perspective, with particular attention to usage frequency, typical learning tasks, and perceived benefits and limitations.

Research Questions

To achieve the stated aim, the following research questions (RQs) were formulated:

• RQ1. How frequently do students use AI chatbots when studying higher mathematics?
• RQ2. For which topics and learning tasks in higher mathematics do students most frequently use chatbots?
• RQ3. What benefits and limitations of AI chatbots do students perceive in the context of learning higher mathematics?

2. Literature Review

2.1. Artificial Intelligence in Education

In this study, artificial intelligence (AI) is conceptualized as a set of technologies and algorithms that enable systems to perform tasks typically associated with human cognitive activity, such as learning, analysis, and decision-making. In educational contexts, AI is primarily applied to support the personalization, adaptation, and automation of learning processes.

Recent research emphasizes that AI has the potential to enhance both the accessibility and quality of learning while creating opportunities for the design of more individualized educational pathways (Holmes et al., 2022; Chiu et al., 2023; Kamalov et al., 2023). In particular, AI is increasingly associated with adaptive and sustainable learning systems that dynamically adjust content, feedback, and instructional support in response to learners’ needs (Bhutoria, 2022; Msomi & Mthethwa, 2024; Kamalov et al., 2023). From this perspective, AI should not be viewed as a single tool but rather as an ecosystem of interconnected approaches—including learning analytics, adaptive learning platforms, and generative models—that operate at multiple levels, ranging from content recommendation to the provision of interactive explanations and formative feedback.

From a pedagogical standpoint, AI-supported personalization aligns with the principles of self-regulated learning, such as goal setting, self-monitoring, and feedback, which enable students to manage their learning progress and pace more effectively (Chang et al., 2023; Pintrich, 2000; Zimmerman, 2002). Bibliometric analyses and systematic review studies further indicate a growing scholarly interest in generative AI and its integration into educational technologies for learning support and analytics, underscoring the timeliness and relevance of this research direction (Chiu et al., 2023; Ivanova et al., 2024; Zhu et al., 2022).

At the same time, the literature highlights persistent challenges that constrain the effective implementation of AI in educational practice, including the need for instructor preparedness, adequate institutional infrastructure, and well-designed integration strategies (Wardat et al., 2024; Sajja et al., 2023). While AI-based tools can reduce routine instructional workload—such as automated assessment and personalized recommendations—thereby allowing educators to focus on more complex instructional design and individualized learner support (Heffernan & Heffernan, 2014; Montero-Mesa et al., 2023; Sangwin & Grove, 2006), these benefits are not realized automatically. Rather, they depend on the ways in which AI technologies are embedded within learning environments and on how both students and instructors engage with these tools (Wardat et al., 2024; Sajja et al., 2023).

Overall, existing research suggests that AI holds considerable potential for advancing personalization in education; however, actual educational outcomes are contingent upon contextual factors, the quality of implementation, and users’ digital and pedagogical competencies (Wardat et al., 2024; Sajja et al., 2023). This underscores the need for focused empirical investigations of AI tools that are already readily accessible to students, particularly chatbot-based systems used to support learning in higher mathematics.

2.2. Intelligent Educational Systems

Continuing the discussion of artificial intelligence applications in education, this section focuses on Intelligent Educational Systems (IES). IES are AI-based software systems designed to support personalized learning by adapting learning activities to learners’ individual needs and providing interactive assistance throughout the learning process (Koedinger et al., 1997; Koedinger et al., 2012; Hwang et al., 2020). In practice, such systems can monitor learner progress, adjust the difficulty level and sequencing of learning tasks, and deliver feedback and guidance, including, where appropriate, through natural language interaction (Sangwin & Grove, 2006; Heffernan & Heffernan, 2014; Koedinger et al., 2012).

From a functional perspective, IES typically rely on machine learning techniques and learner modeling to infer students’ current knowledge states and misconceptions, and to generate individualized learning trajectories. This form of adaptive regulation can help reduce cognitive load and promote sustained learner engagement, particularly in cognitively demanding disciplines such as higher mathematics (Msomi & Mthethwa, 2024). Empirical and analytical studies indicate that AI-supported educational systems can enhance learning support by enabling personalized assignments, timely feedback, and interactive learning content, thereby addressing the limitations of the traditional “one-size-fits-all” instructional approach (Sangwin & Grove, 2006; Heffernan & Heffernan, 2014; Koedinger et al., 2012).

Recent research also underscores the expanding role of generative artificial intelligence in educational technologies, including large language models (LLMs), which extend the scope of interactive learning support and facilitate on-demand explanations and dialogue-based assistance (Ali et al., 2024; Chen et al., 2023; Kasneci et al., 2023). Nevertheless, not all AI-in-education research explicitly addresses mathematics learning contexts. Building on these developments, the present study examines students’ use of AI chatbots as tools for problem-solving and learning support in higher mathematics.

Given the diversity of Intelligent Educational Systems (IES), the following section provides an overview of key IES types commonly discussed in the literature and clarifies their relevance for supporting personalized learning in higher mathematics.

2.3. Types of IES

This section summarizes the main types of Intelligent Educational Systems (IES) that are commonly discussed in the literature in relation to personalized learning, including applications relevant to the teaching and learning of higher mathematics:

Intelligent Tutoring Systems (ITS): systems that provide individualized instruction and feedback by modeling learners’ knowledge states and adapting tasks, explanations, and practice activities to their progress (e.g., adaptive problem sets and diagnostic feedback) (VanLehn, 2011).

• Adaptive Hypermedia Systems (AHS): systems that personalize the navigation structure and presentation of learning resources—such as texts, multimedia content, and hyperlinks—based on learners’ characteristics, preferences, and prior knowledge (Brusilovsky, 2001).
• Dynamic Task Generation Systems (DTGS): systems that automatically generate, select, or sequence learning tasks and provide corresponding solutions and feedback using pedagogical and domain models, thereby supporting task variability and individualized practice (Circi et al., 2023).

Conversational Agents/Educational Chatbots: dialogue-based systems that facilitate learning through natural language interaction by answering learners’ questions and providing guidance during problem-solving processes (Kuhail et al., 2023).

• Generative AI and LLM-based Tools: systems that leverage large language models to generate explanations and learning-support content on demand; these tools are increasingly discussed in the literature as flexible and scalable solutions for personalized learning support (Kasneci et al., 2023).

Overall, these types of Intelligent Educational Systems vary in terms of their structural complexity and the degree of instructional guidance they provide. In the context of the present study, conversational agents and LLM-based chatbots (e.g., ChatGPT) are of particular relevance, as they are readily accessible to students and can be used for on-demand support in the study of higher mathematics.

2.4. AI in Mathematics Education

Research indicates that the integration of artificial intelligence (AI) into educational practice enables the adaptation of learning materials and instructional support to students’ current levels of proficiency and individual learning needs (Msomi & Mthethwa, 2024). In mathematics education, AI-based tools—including chatbots—are increasingly employed to support cognitively demanding activities such as solving calculus problems (e.g., derivatives and integrals), visualizing mathematical functions, and interpreting computational results. These applications are particularly relevant in higher mathematics, where students often require timely feedback and step-by-step explanations to address conceptual difficulties and develop problem-solving competence (Wardat et al., 2023; Zhuang, 2025; Sun et al., 2023).

Previous studies also highlight broader instructional use cases of AI in mathematics education, including support for function graphing and visualization through computational tools, facilitation of real-world problem modeling, and enhancement of student engagement via interactive and gamified learning activities (Msomi & Mthethwa, 2024; Xu et al., 2024). For instance, Xu et al. reported that a gamified chatbot intervention in an information technology course was associated with increased student interest, improved academic performance, and higher levels of cognitive activity, suggesting the potential of chatbot-based learning support to foster sustained engagement (Msomi & Mthethwa, 2024; Xu et al., 2024). Although this evidence does not focus exclusively on higher mathematics, it illustrates key mechanisms—such as feedback, interaction, and motivation—that are directly relevant to mathematics instruction.

Overall, these findings suggest that AI-enabled tools can be integrated into a wide range of educational platforms and learning environments, including learning management systems (LMSs) and massive open online courses (MOOCs), to support personalization at multiple levels (e.g., course, module, topic, or lesson) (Alotaibi, 2024). Within the context of the present study, the most pertinent direction is the use of generative AI chatbots to provide on-demand learning support in higher mathematics.

Design Higher Mathematics Courses in Moodle Cloud

Moodle Cloud is a cloud-based deployment of Moodle that enables educational institutions to create, deliver, and manage online courses without the need for local server installation or ongoing technical maintenance. In the present manuscript, Moodle Cloud is referenced solely as an illustrative example of a learning management system (LMS) environment in which personalization mechanisms may be implemented through instructional design, activity configuration, and the use of learning analytics. This description is conceptual in nature and does not imply that a dedicated intervention involving Moodle Cloud was conducted as part of this study. From a functional perspective, an LMS environment such as Moodle Cloud can support personalization in higher mathematics through several widely used mechanisms. First, instructors may organize course content into structured modules and learning pathways, offering differentiated learning resources tailored to students’ prior knowledge and individual learning needs. Second, assignments and assessments can be configured to accommodate individual learning paces and provide formative feedback, which is particularly important in mathematically intensive subjects that require repeated practice and progressive skill development. Third, integrated communication tools, including discussion forums and messaging systems, can facilitate instructor–student interaction and peer collaboration, enabling learners to address conceptual difficulties in a timely manner.

At the same time, cloud-based LMS deployments may face practical constraints related to resource limitations, technical support availability, and the degree of customization permitted, especially under free or low-cost service plans. Nevertheless, opportunities for personalization may be enhanced through system integrations and add-on components, where available, including learning analytics tools and automated assessment workflows.

The potential contribution of learning analytics is supported by Perez-Suay et al., who demonstrate that behavioral data collected within Moodle-based environments can be used to predict student performance in higher mathematics courses, underscoring the relevance of digital learning traces for adaptive support and personalization (Pérez-Suay et al., 2023). More broadly, Mohamed et al. discuss the implementation of adaptive educational platforms and intelligent tutoring systems, emphasizing that AI-supported technologies can promote individualized learning by aligning tasks and instructional materials with students’ current competence levels and learning pace—an especially critical consideration in higher mathematics education (Mohamed et al., 2022). In parallel, research on generative artificial intelligence highlights both opportunities and limitations. Cope and Kalantzis point to the potential of large language models to provide automated explanations, solution verification, and feedback, while stressing the importance of maintaining a balance between technological support and the teacher’s pedagogical role (Cope & Kalantzis, 2024). Similarly, Wardat et al. report positive associations between AI use and academic achievement in mathematics education, while also identifying challenges such as insufficient teacher training and infrastructural constraints, and outlining key conditions for successful implementation (Wardat et al., 2024). Taken together, these studies suggest that LMS environments may serve as a viable foundation for future AI-enabled personalization initiatives, provided that issues of methodological transparency, instructor support, and infrastructural readiness are adequately addressed.

2.5. Chatbots in Education

The integration of artificial intelligence (AI) into education has generated new opportunities to enhance teaching and learning processes. In this context, educators may employ digital and AI-enabled tools to support assessment practices, learning analytics, and instructional decision-making, while students may benefit from interactive tutoring support as well as increased flexibility through asynchronous and self-paced learning opportunities.

One approach that has received increasing attention is the use of chatbots in educational settings, including higher mathematics. Research on educational chatbots has expanded in recent years, reflecting growing interest among both students and instructors. However, empirical evidence regarding the specific contribution of chatbots to personalized learning in higher mathematics—and concerning how students perceive, adopt, and utilize such tools in this domain—remains limited and unevenly reported across educational contexts (Kostikova et al., 2025; Aleshkovski et al., 2024). Accordingly, the aim of the present study was to examine students’ self-reported perceptions and patterns of chatbot usage as a potential means of supporting personalized learning in higher mathematics. The analysis is based on an empirical survey conducted among students at Saken Seifullin Kazakh Agrotechnical University in Kazakhstan.

3. Materials and Methods

3.1. Research Design and Setting

The present study adopted a cross-sectional survey design to investigate students’ use of AI-based chatbots as a means of supporting personalized learning in higher mathematics at Saken Seifullin Kazakh Agrotechnical University (Kazakhstan).

3.2. Participants

Participation in the study was voluntary. The total sample comprised n = 154 students, including 126 first-year students, 18 second-year students, and 10 third-year students. Participants were between 18 and 22 years of age. The gender distribution of the sample was 38% female and 62% male. A summary of the participant’s demographic characteristics is presented in

Table 1. Participant characteristics (n = 154).

Characteristic	Value
Sample size	n = 154
Year of study	Year 1: n = 126; Year 2: n = 18; Year 3: n = 10
Age	18–22 years
Gender	38% female; 62% male

.

3.3. Instrument (Questionnaire)

Data were collected using an online questionnaire administered through Google Forms. The instrument consisted of 12 items, including demographic questions (year of study, faculty, and gender) as well as items examining:

• students’ understanding of chatbots;
• the frequency and purposes of chatbot use in higher mathematics;
• preferred types of chatbots;
• perceived benefits and limitations of chatbot use;
• desired chatbot functionalities.

The questionnaire included a combination of single-choice categorical items, multiple-response (checkbox) items, and an optional open-ended “Other” response field to allow participants to provide additional input. The full survey questionnaire used in the study is provided in Appendix A.

3.4. Data Collection Procedure

The survey link was distributed to students in Years 1–3 through WhatsApp group chats associated with their teaching groups. The questionnaire was completed online, and responses were automatically recorded and stored using Google Forms.

3.5. Data Analysis

Survey responses were analyzed using descriptive statistical methods, including frequencies and percentages, and were summarized visually using bar charts. Chatbot usage frequency was treated as a categorical grouping variable with four response options (“very often,” “when there is a need,” “very rarely,” and “never”). As the dataset consisted primarily of categorical survey items and did not include individual-level numeric outcome variables, the analysis was limited to descriptive summaries; no inferential group comparisons were conducted. Future research will incorporate standardized quantitative outcome measures to enable inferential statistical analyses and the estimation of effect sizes.

3.6. Ethics and Limitations of the Study

Participation in the study was voluntary. The survey was administered anonymously, and no personally identifiable information was collected. Informed consent was obtained from all participants prior to their participation. The study protocol was reviewed and approved by the institutional ethics committee (Protocol No. 13, 17 April 2024).

The study was conducted at a single university, which may limit the generalizability of the findings. Furthermore, the reliance on self-reported data may introduce response bias. Due to the cross-sectional design of the study, causal relationships cannot be inferred. Future research should consider multi-institutional samples and the inclusion of additional data sources to validate and extend the present findings.

4. Results

A conceptual comparison was conducted between traditional teaching methods and AI-supported approaches to the personalization of higher mathematics education. This comparison highlights the potential pedagogical advantages associated with AI-supported learning environments.

Table 2. Conceptual comparison of traditional teaching methods and artificial intelligence-supported approaches in higher mathematics education.

Parameter	Traditional Methods	AI-Support Approaches
Individualization	Limited	High (adaptive tasks)
Feedback rate	Low	Instant
Motivation of students	Average	High (use of game-based elements)
Availability of materials	Limited	Extensive (online access)

presents a literature-informed conceptual synthesis intended to contextualize the survey findings; it is not directly derived from the empirical questionnaire data.

Analysis and interpretation of the survey data (n = 154) yielded the following findings.

4.1. Respondents’ Understanding of Chatbots

A substantial proportion of respondents (72%, n = 111) demonstrated an adequate understanding of the nature and basic operating principles of chatbots. In general, chatbots were described as software programs or systems capable of interacting with users via text-based communication. Their primary functions include processing user queries, providing information, executing Commands, and supporting problem-solving activities.

In contrast, 14% of respondents indicated that they did not possess a sufficient understanding of chatbots but expressed an interest in learning more, suggesting a latent demand for further exposure to and training in the use of such technologies. An additional 14% of participants reported limited interest in chatbots. This lack of interest may be attributed to insufficient prior experience with such tools or limited awareness of their potential applications in educational contexts.

The distribution of respondents’ levels of understanding of chatbots is presented in

Table 3. Respondents’ Understanding of Chatbot Operating Principles (n = 154).

Category of Respondents	Percentage of Respondents
Have the right understanding	72%
Would like to learn more	14%
Do not have a particular interest	14%

.

Overall, the analysis of respondents’ general understanding of chatbot principles indicates that the majority possess a basic level of familiarity with such technologies, suggesting a potential readiness to engage with chatbot-based tools in educational contexts. The subgroup of respondents who expressed interest in acquiring additional information (14%) may represent a target audience for future instructional or awareness-raising initiatives.

4.2. Frequency of Chatbot Use

Analysis of the frequency with which respondents reported using chatbots to obtain assistance with tasks in higher mathematics yielded the following distribution:

(a)

Very often (7%), which may indicate intensive or regular reliance on chatbots for learning support;

(b)

When there is a need (49%), including situations in which problem-solving proves challenging or when rapid access to information is required;

(c)

Very rarely (26%), potentially reflecting confidence in one’s own knowledge or a preference for alternative information sources;

(d)

Never (18%), which may indicate a preference for traditional learning approaches or limited trust in automated solutions.

The distribution of chatbot usage frequency among respondents in the context of higher mathematics is illustrated in Figure 1.

The majority of respondents (49%) reported using chatbots occasionally, suggesting that their engagement with these tools is situational and dependent on the specific learning tasks. The relatively low proportion of regular users (79%) may reflect limited trust in chatbot technologies or a lack of habitual use.

Given the categorical nature of the survey items, only descriptive results are presented, without inferential statistical comparisons. Interpretation of these descriptive patterns in relation to the existing literature is provided in the Section 5.

4.3. Types of Tasks Solved with the Assistance of Chatbots

Respondents reported using chatbots to address a variety of tasks in higher mathematics:

(a)

Solving equations (22%), including finding roots, solving systems of equations, and analyzing algebraic expressions;

(b)

Calculating derivatives and integrals (44%), with chatbots supporting the computation of function derivatives and the evaluation of integrals;

(c)

Plotting function graphs (28%), as chatbots facilitate the construction of graphs and visualization of functional relationships;

(d)

Searching for information related to lecture topics (6%), enabling access to supplementary explanations or learning materials.

The distribution of higher mathematics tasks for which chatbots were employed is illustrated in Figure 2.

The distribution indicates that calculating derivatives and integrals is the most frequently addressed task. The predominance of chatbot use for derivatives and integrals (44%) highlights their practical utility for computational support. Twenty-eight percent of respondents reported using chatbots for graph construction, while the relatively low proportion for information searches (6%) may reflect a preference for traditional learning resources.

4.4. Advantages of Using Chatbots

Respondents identified several key advantages of employing chatbots to support problem-solving in higher mathematics:

• Quick access to solutions (28%): chatbots provide immediate responses, enabling rapid problem resolution;
• 24/7 consulting support (28%): chatbots are accessible at all times;
• Convenience and ease of use (27%): the intuitive interface facilitates user-friendly interaction;
• Support for understanding complex topics (45%): chatbots can clarify mathematical concepts and enhance comprehension.

The distribution of reported advantages is presented in Figure 3.

The ability to explain complex topics (45%) emerged as the most prominent advantage, highlighting the educational value of chatbots beyond merely providing solutions.

4.5. Limitations and Disadvantages

Respondents also identified several limitations and disadvantages associated with using chatbots to solve higher mathematics problems:

• Limited functionality (15%);
• Insufficient accuracy (42%);
• Difficulty in inputting mathematical expressions (35%);
• Difficulty in obtaining answers (8%).

The distribution of reported limitations and disadvantages is presented in

Table 4. Limitations and disadvantages of using chatbots (n = 154).

Limitations and Disadvantages of Using Chatbots	The Response Rate (%)
Limited functionality	15
Insufficient accuracy	42
Difficulty in inputting mathematical expressions	35
Difficult to answer	8

.

Insufficient accuracy (42%) and the complexity of inputting mathematical expressions (35%) were identified as the primary barriers to effective chatbot use. These concerns align with findings in the broader literature, which emphasize challenges related to accuracy, usability, and integration of AI tools in educational contexts (Maghsoudi et al., 2021; Yakov & Partners, 2023).

4.6. Respondents’ Interest in Using Chatbots

Regarding interest in using chatbots for solving mathematical problems, 60% of respondents reported a positive disposition, indicating that the majority of participants (n = 92/154) are receptive to employing chatbots for mathematical tasks. Conversely, 40% of respondents (n = 62/154) expressed a negative disposition, frequently citing concerns about the need to verify results and the reliability of chatbot outputs. The distribution of respondents’ interest in using chatbots is illustrated in Figure 4.

The majority of respondents (60%) reported a positive attitude toward using chatbots, whereas 40% reported doubts or negative attitudes.

5. Discussion

5.1. Comparison of Traditional Teaching Methods and AI Approaches

5.1.1. Interpretation of Results

The results indicate a high level of awareness of chatbots among respondents (72%). However, the fact that 14% are not interested in chatbots may suggest a lack of exposure to AI-based tools in their academic environment. To increase engagement, universities may introduce additional training sessions or workshops on AI-powered chatbots for educational purposes.

The conceptual comparison presented in Table 1 provides a literature-based summary of commonly discussed differences between traditional instructional methods and AI-supported approaches in higher mathematics education. In this literature-informed synthesis, AI-supported approaches are generally characterized by the provision of more immediate feedback, potentially stronger motivational support, and broader access to learning resources. A more detailed interpretation of these themes in relation to prior studies is presented below (Zimmerman, 2002; Holmes et al., 2022; Sajja et al., 2023).

Moreover, our findings are consistent with those reported by Bhutoria and Sun et al. (Bhutoria, 2022; Sun et al., 2023), which indicate that AI can facilitate the adaptation of learning materials to the individual needs of students. However, in comparison with data from countries where AI tools are more extensively adopted (e.g., China and the United States), this sample demonstrates a lower frequency of engagement with such technologies among respondents in Kazakhstan. This pattern may reflect differences in curricular integration and institutional adoption; however, the current study does not allow for causal inferences.

5.1.2. Practical Implications for AI Integration in Higher Mathematics Education

Integrating AI into educational platforms: The high percentage of respondents who correctly understand chatbot functionality suggests that universities can expand AI-based learning resources and integrate them into Learning Management Systems (LMS) such as Moodle Cloud.

Bridging the gap for uninformed respondents: 14% of respondents interested in learning more could benefit from structured AI training programs, including hands-on exercises with chatbots like ChatGPT and Wolfram Alpha.

Consideration of cultural and technological barriers: Since the level of AI implementation varies across educational institutions and countries, future strategies should consider the level of digital literacy, institutional policies, and teacher training. The limitations of the study and directions for further research are presented in Section 6.3.

5.2. Frequency of Chatbot Usage

5.2.1. Interpretation of Chatbot Usage Patterns

The results indicate that while a majority of respondents (49%) use chatbots occasionally, very few (7%) rely on them consistently. This low adoption rate may be attributed to:

• A lack of awareness about the full potential of AI-powered chatbots.
• Limited integration of chatbots into structured learning environments.
• Concerns about accuracy and reliability, leading students to cross-check information using other sources.

Given the categorical nature of the survey items, we report descriptive results (frequencies and percentages) and do not perform inferential group comparisons.

These findings align with previous studies by Holmes et al. and Sajja et al., which highlight the positive impact of AI integration on student learning outcomes (Holmes et al., 2022; Sajja et al., 2023).

5.2.2. Practical Implications for Regular Chatbot Use

Encouraging Regular Chatbot Use: Universities should integrate chatbots into structured learning activities to increase student familiarity and trust.

Addressing Accuracy Concerns: Enhancing chatbot reliability and offering guidance on cross-verification can increase confidence in AI-assisted learning.

Further Research on Learning Outcomes: Future studies should explore whether prolonged chatbot use leads to measurable academic improvements.

Thus, while chatbots offer clear advantages in higher mathematics education, their full potential remains underestimated due to a lack of awareness and trust, highlighting the need for the strategic integration of these technologies into the educational process.

5.3. Interpretation of Results on the Use of Chatbots for Solving Tasks

The histogram indicates that the most common use of chatbots is for calculating derivatives and integrals (44%). This confirms their value in complex mathematical calculations, where automation and accuracy are crucial.

The second most popular use case is plotting function graphs (28%), highlighting their usefulness in visualizing mathematical relationships.

A comparatively smaller percentage of respondents (22%) use chatbots for solving equations, possibly because equation solving often requires additional analysis rather than just computation.

The lowest percentage (6%) of respondents use chatbots to search for information on lecture topics. This could be explained by a preference for traditional sources of information, such as textbooks or instructor-led lectures.

Thus, the findings emphasize that chatbots are most in demand for tasks requiring automated calculations, such as integration, differentiation, and graph plotting, while they are less commonly used for more theoretical and analytical tasks.

5.4. Interpretation of Results on Chatbot Advantages

Based on respondents’ self-reported data, the capacity of chatbots to provide explanations of complex topics emerged as the most commonly utilized advantage, as indicated by 45% of the respondents. This highlights their potential as educational tools, as they help students understand mathematical concepts rather than just providing answers.

Both quick access to solutions (28%) and 24/7 availability (28%) are equally appreciated, reflecting respondents’ need for instant assistance and flexible learning schedules. These features are particularly useful in self-directed learning environments, where students may not have immediate access to instructors.

According to respondents’ self-reports, usability and ease of use (27%) also play a significant role in the adoption of chatbots, suggesting that user-friendly design contributes to their effectiveness as learning tools.

Overall, the results indicate that chatbots are most beneficial for explaining complex mathematical concepts and providing immediate assistance, making them valuable tools for enhancing student engagement and learning efficiency in higher mathematics.

5.5. Interpretation of Results on Chatbot Limitations

The analysis results show that the main drawbacks of chatbots are calculation errors (42%) and the complexity of working with mathematical expressions (35%). This may suggest, from respondents’ perspective, that despite the widespread adoption of such technologies, their algorithms still require improvement, particularly in handling complex mathematical operations.

The limited functionality (15%) suggests a lack of specialized tools in existing chatbots, making them less effective for solving complex mathematical problems.

Additionally, 8% of respondents found it difficult to define chatbot limitations, which may indicate either insufficient experience with these technologies or the need for deeper integration of AI tools into the educational process. This highlights the need to improve functionality and build trust in chatbots.

Thus, to enhance the effectiveness of chatbots in higher mathematics education, the following improvements are potentially relevant:

• Improving algorithm accuracy by integrating more advanced mathematical models and expanding knowledge bases.
• Optimizing the user interface to simplify the input of mathematical expressions.
• Developing specialized chatbots that can solve complex mathematical problems and adapt to students’ needs.

These findings are consistent with a range of recent studies emphasizing that, to enhance the effectiveness of AI-driven solutions in education, several conditions should be satisfied:

• Clear structuring of learning objectives and feedback mechanisms (Chang et al., 2023);
• Consideration of students’ behavioral patterns and cognitive styles (Xu et al., 2024);
• Development of personalized learning strategies based on AI-driven analytics (Wang et al., 2023);
• Use of AI to support self-regulated learning and the development of higher-order thinking skills (Ivanova et al., 2024);
• Adaptation of AI platforms to the specific characteristics of the subject matter, particularly in higher mathematics, where precision and structural rigor are essential (Pérez-Suay et al., 2023).

Accordingly, rather than adopting a one-size-fits-all approach to chatbot design, there is a need to develop specialized intelligent agents capable of:

• accurately interpreting mathematical contexts;
• supporting dialogue aligned with instructional objectives;
• flexibly adapting to individual students’ learning styles and paces.

Taken together, these limitations—calculation errors (42%), difficulties in working with mathematical expressions (35%), and limited functionality (15%)—may help explain why a substantial proportion of respondents (40%) expressed doubts about the use of chatbots. From the respondents’ perspective, concerns regarding reliability, as well as the effort required to input and interpret mathematical notation, may undermine trust and reduce the willingness to rely on chatbots for complex problem-solving tasks. Consequently, addressing these issues through targeted improvements in chatbot accuracy, mathematical expression processing, and user guidance is likely to be essential for strengthening trust and supporting wider adoption of chatbots in higher mathematics education.

5.6. Interest in Using Chatbots

The high level of interest in chatbots (60%), as reported by respondents, reflects a growing demand for automated educational tools, especially in higher mathematics, where students seek instant solutions and explanations.

However, the fact that 40% of respondents expressed skepticism highlights the ongoing concerns about the accuracy and reliability of chatbots.

These observations align with the conclusions of Holmes et al. and Sajja et al., who emphasize that AI can enhance engagement through personalized support and 24/7 availability (Holmes et al., 2022; Sajja et al., 2023). However, challenges such as insufficient accuracy and difficulties in integrating AI tools into educational platforms remain important issues for further research (Maghsoudi et al., 2021; Yakov & Partners, 2023).

To summarize, according to respondents’ self-reports, the survey indicates several perceived contributions of chatbots to mathematics learning:

• learning support and immediate answers to questions;
• personalized guidance;
• timely feedback; and
• availability and flexibility.

In addition, the questionnaire items on preferred problem types and desired chatbot functionalities help clarify practical implementation priorities (i.e., which content areas to support and what form of assistance is perceived as most useful). These insights can inform instructional guidance and training modules, as well as the design and evaluation of mathematics-oriented chatbots aligned with learners’ expressed needs.

Prospects and Challenges:

• Prospects

AI can improve the accessibility and quality of education, particularly in technical disciplines. Research by Wardat et al. demonstrates that interactive AI-powered assignments increase students’ motivation and improve their academic performance (Kasneci et al., 2023).

• Challenges

It is necessary to address issues related to insufficient algorithm accuracy and the complexity of integrating technologies into the educational process. Additionally, it is crucial to provide training for educators to ensure the effective use of AI.

6. Conclusions

This study indicates that AI chatbots have the potential to support personalized learning in higher mathematics, as reflected in respondents’ self-reported perceptions and usage patterns. Such tools can offer on-demand assistance, including adaptive explanations, timely feedback, and round-the-clock availability, which respondents perceive as valuable for learning support. Given the categorical nature of the survey items, only descriptive analyses (frequencies and percentages) were conducted, and no inferential group comparisons were performed. The findings underscore respondents’ usage patterns and perceived benefits of chatbots in higher mathematics, highlighting their potential role as instruments for personalized educational support. These results are broadly consistent with previous research (Holmes et al., 2022; Sajja et al., 2023).

6.1. Key Findings

• According to respondents’ self-reports, AI chatbots may enhance the perceived accessibility and effectiveness of learning support, particularly in relation to complex mathematical topics.
• Respondents most frequently used chatbots for calculating derivatives and integrals (44%), plotting function graphs (28%), and explaining complex mathematical concepts (45%).
• The primary perceived limitations reported by respondents include computational inaccuracies (42%), difficulties in inputting mathematical expressions (35%), and potential risks associated with over-reliance on AI tools.

6.2. Practical Recommendations

• Integration of AI tools into educational platforms such as Moodle Cloud, with attention to the specific needs of mathematics education.
• Professional development for instructors focused on interpreting AI-generated recommendations and recognizing common student errors.
• Development of learners’ critical thinking skills when interacting with AI, including the ability to verify, analyze, and consciously apply chatbot-generated outputs.

6.3. Limitations and Future Research

Limitations of the study:

Despite the promising results of AI guidance in mathematics education, several limitations need to be considered:

• Accuracy of AI-generated solutions: Chatbots sometimes provide incorrect answers due to limitations of the algorithms that require human verification.
• Limited problem-solving capabilities: Current AI models are good at routine calculations, but have difficulty providing non-routine mathematical problems that require deep conceptual understanding.
• Over-reliance on AI tools: Excessive reliance on chatbots may hinder the development of independent problem-solving skills and reduce opportunities for critical thinking and verification practices.
• Technological barriers: AI-based platforms require a stable internet connection and support from educational institutions, which may be a challenge in some educational environments.

Research perspectives:

• Longitudinal comparative study: The prospective study will compare the academic performance of learners using AI tutors with those taught using conservative methods in a broad context.
• AI model enhancement: Develop advanced machine learning technologies such as deep learning and reinforcement learning to enhance the learning capabilities of chatbots.
• Teacher-AI collaboration: Explore how AI tools can complement rather than replace teaching to deliver optimal pedagogical outcomes.
• Ethical considerations: exploring the ethical implications of AI in education, including issues of data privacy and algorithmic bias.
• Fostering critical thinking: Researching strategies to encourage students to independently analyze and critically evaluate AI-generated information to prevent blind trust in technology.
• Training and responsible use: Introducing courses or workshops on the effective and responsible use of AI in the classroom.

These areas of research have the potential to advance personalized learning methods in AI and expand their application in higher mathematics education.

6.4. Applicability of Findings

The study was conducted at one of Kazakhstan’s leading agrotechnical research universities, characterized by an advanced digital infrastructure and active participation in international projects. This context enables the extrapolation of the findings to similar higher education institutions and supports the development of national initiatives aimed at the digitalization of higher education. Expanding the research to other academic institutions represents a timely and necessary step toward the development of universal strategies for the integration of AI into teaching and learning processes.