Optimized Math Journey: AI-Driven Approaches to
Personalized Learning in Mathematics
Abstract— This study examined the effects of AI-driven teaching has proved imperfectly flexible compared to the
personalized learning systems on students' mathematics present adaptive environments that create real-time
performance, investigated the approaches employed by personalized support for students to advance at their speed
teachers in integrating AI tools to support personalized [3]. History has produced various platforms, such as Khan
learning, and explored the challenges and opportunities Academy and DreamBox Learning, which can adapt in
surrounding AI adoption in mathematics instruction at terms of the difficulty and speed of lessons and their
Isabela State University. Using a qualitative approach, accompanying actions taken toward feedback for each
data were collected through interviews, observations, student, i.e., intervention[9].
and focus group discussions to gain insights into Despite the various fruit-bearing measures proving the
educators' instructional practices and experiences beneficial aspects of this form of education, such measures
utilizing AI-based tools in their classrooms. The findings have failed to address the critical challenges inherent in
revealed that AI-driven systems positively impacted employing artificial intelligence [5] in education; these
student performance, enhancing engagement, include resources, teacher training, and access to the newest
motivation, and mastery of mathematics through tools. Research has been on K-12 students and
customized learning experiences. Students benefited predominantly one AI tool; hence, there is still a gap in
from adaptive learning paths and real-time feedback, understanding how AI impacts all types of education in the
which allowed them to progress at their own pace and long run and across populations.
improved their understanding of mathematical concepts. Research is needed to determine how AI can redefine
Mathematics teachers employed a variety of AI- mathematics education, scale across contexts, and ultimately
supported approaches, such as adaptive learning overcome the challenges toward adoption. The study will
modules, scaffolded problem-solving exercises, and evaluate the effectiveness, pros, and cons of AI-powered
progress-monitoring tools, to personalize instruction math learning and give an idea of its use in reforming
based on each student's unique mathematics education.
needs. These techniques enabled teachers to address
diverse Research Objectives
learning abilities within their classrooms, making 1. Determine how much mathematics teachers use the
mathematics more accessible and effective for all different AI-based approaches to facilitate personalized
students. Despite these benefits, several challenges learning.
hindered the full integration of AI technologies,
including limited technical support, high maintenance 2. Identify the degree to which AI-driven
requirements, and data privacy concerns. Financial personalized learning system affects the mathematics
constraints and insufficient infrastructure in some performance of the students
schools posed further barriers to seamless AI adoption.
The study recommended increased institutional support 3. Assess the relationship between the extent to which
for technical infrastructure, continuous professional mathematics teachers use the different AI-based
development for educators, and the establishment of approaches to facilitate personalized learning and the
stringent data privacy policies. These measures were degree to which AI-driven personalized learning system
considered essential to optimize AI’s role in enhancing affects the mathematics performance of the students
mathematics education and to ensure a secure, equitable,
and efficient learning environment for students at 4. Explore the challenges and opportunities for
Isabela State University. adopting AI-driven technologies in personalized
mathematics learning.
Keywords: Optimized Math Journeys, AI-Driven
Approaches, Personalized Learning in Mathematics,
Artificial intelligence RESEARCH METHODOLOGY
This study employed a mixed-methods design. The
quantitative part involved a descriptive-correlation approach
based on surveys conducted to explore the use of AI-based
INTRODUCTION methods by mathematics instructors and their effect on the
student's performance. The research analyzed the
Schooling in mathematics, like many other phenomena, has correlations of the above constructs. For the qualitative part,
dramatically evolved due to digital technological thematic analysis was applied to the data generated from
advancement and artificial intelligence[15]. Traditional semi-structured focus group discussions (FGDs).
XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE
Information Classification: General
Data were gathered via a researcher-administered at 4.02, or nearly entirely used. Many techniques are highly
questionnaire and FGDs with 42 purposively selected state effective. Again, virtual AI tutoring tutors such as ChatGPT
university mathematics teachers in the Philippines, whose were rated the most intensive at 4.48 because they are
experience included applying AI tools in teaching. There primarily used for on-demand support. ITSs (4.32),
were two parts to the survey instrument: part one aimed to Personalized Practice Generators (4.22), and AI-driven
assess the level at which teachers apply AI-driven methods recommendations for content (4.25) are also highly
in their teaching processes (8 indicators). In contrast, the implemented in the personalization of learning experiences.
second part probed the degree of impact on students' Adaptive learning systems (4.12) and instant feedback tools
performance by AI-driven personalized learning methods (8 (4.12) are predominantly used for dynamic problem
indicators). Response scores were anchored on a 5-point adaptations and error corrections, whereas tools for peer-
Likert scale. based problem-solving scored 4.13. AI-based diagnostic
The survey instrument was subjected to intensive validation tests scored much lower at 2.53, suggesting limited usage of
and reliability testing. Expert content validation refined the these tools to identify knowledge gaps and provide
items, and a pilot test established the questionnaire's improvement recommendations.
reliability with Cronbach's alpha values of 0.894 and 0.854 The AI tools effectively personalize math education,
for the two sections, which shows strong internal focusing on real-time adjustments and customization. Still,
consistency. The key AI tools assessed included adaptive there are significant opportunities for improvement in the
learning systems, intelligent tutoring systems, diagnostic uptake of diagnostic testing.
assessments, and virtual tutors, with a category mean of Table 2. The degree to which AI-driven personalized
4.02, showing high application levels in general. Once learning system affects the mathematics performance of
permission to float the questionnaires was granted, the students
researcher gathered data through the survey questionnaire
for the respondents. Indicators Mean Descriptive
AI-driven personalized learning system impacts on the Value
1 AI-driven personalized learning systems Very High
performance of the students in mathematics were interpreted improve students' accuracy in solving
through the following: mathematical problems by providing 4.23
targeted practice on weak areas.
Point Mean Range Descriptive Value 2 Students using AI-based platforms High
demonstrate increased retention of
5 4.20 – 5.00 Very high mathematical concepts through adaptive 3.45
4 3.40 – 4.19 High revision tailored to their learning pace.
3 2.60 – 3.39 Moderate 3 AI systems help reduce math anxiety by High
offering step-by-step guidance and instant
2 1.80 – 2.59 Low Skill feedback, boosting students' confidence in 3.57
1 1.00 – 1.79 Very Low problem-solving.
4 AI in personalized learning significantly High
enhances students' understanding of
Pearson's research was conducted regarding the engagement complex mathematical topics by breaking
of mathematics teachers with AI-based methods and their them down into more straightforward, 4.08
customized learning paths.
outcomes on student performance. JAMOVI software was 5 AI-powered assessment tools provide real- High
applied for the statistical test with the 0.05 alpha level. time feedback, allowing students to correct
A schedule of visits was produced to ensure 100% retrieval, mistakes immediately and refine their 4.11
and a schedule of the FGD also ensued. The ethical understanding of mathematical principles.
6 Personalized AI systems lead to improved Very High
standards were also observed in all parts of the study. Data student engagement in math by adapting
gathered mainly were treated confidentially and only for the lessons to their interests and maintaining 4.37
purposes of the study. The data collected in the study were optimal challenge levels.
processed using the weighted mean. In addition, the data for 7 The personalized nature of AI platforms Very High
encourages more independent learning,
the extent to which mathematics teachers use the various allowing students to take control of their 4.58
AI-based approaches in offering personalized learning and mathematical education at their own pace.
how the AI-driven personalized learning system impacts the 8 AI-based tools help improve students’ High
students' mathematics performance. problem-solving strategies by analyzing
their approach to math problems and 3.89
An FGD of 12 mathematics teachers engaged with AI-
suggesting more efficient methods.
driven method-based teaching explored the opportunities Category Mean 4.04 High
and challenges in personalized math learning. Semi-
structured interviews provided detailed insights, and data The table highlights the impact of AI-driven personalized
analysis was done using thematic analysis. learning systems on students' mathematics performance,
with an average category score of 4.04 ("High").
RESULTS AND FINDINGS Independent learning support received the highest rating at
4.58 ("Very High"), emphasizing student control over
Table 1. The extent to which learning pace. Lesson tailoring for better engagement (4.37)
mathematics teachers use AI-based approaches to facilitate and targeted practice in weak areas (4.23) also scored "Very
personalized learning. High," demonstrating their positive effects on problem-
The table assesses AI-based methods for personalized solving and comprehension.
mathematics education, with an average across all categories
Information Classification: General
Other measures rated "High" include tailored learning paths Themes Key Points Sample Quotes
for complex topics (4.08), immediate feedback (4.11), software, hardware, and budget for AI integration.
reduced math anxiety (3.57), and improved problem-solving consistent updates. (P6)
Sophistication - Limited budgets and - AI demands constant
strategies (3.89). However, while rated "High" at 3.45, and Resource infrastructure hinder updates and training. (P2)
adaptive revision for concept retention showed room for Intensity implementation. - Old systems and high
improvement. - Teachers need training to costs make AI adoption
AI-driven systems significantly enhance engagement, use AI systems effectively. difficult. (P4)
independence, and targeted learning, though refining - Teachers need training to
- Teachers need the
use AI effectively.
adaptive revision could further boost their effectiveness. Teacher
- Professional development
confidence to use AI tools
Table 3. Relationship between the extent to which Readiness and effectively. (P1)
workshops are necessary.
Professional - Workshops help teachers
mathematics teachers use the different AI-based approaches Development
- AI tools demand changes in
integrate AI with
pedagogy and teaching
to facilitate personalized learning and the degree to which traditional methods. (P3)
approaches.
AI-driven personalized learning system affects the - Concerns about data
mathematics performance of the students - Is student data secure?
privacy and ethical use.
Effect of AI- Use of different AI-based approaches by Ethical Issues (P3)
- Clear policies are needed to
driven mathematics teachers to facilitate personalized and Data - Privacy and transparency
protect student data.
personalized learning Protection are key for trust in AI
- Risks of algorithmic bias in
learning system Correlation Probability Value Interpretation systems. (P7)
assessment.
on the coefficient - Regular maintenance and - Without tech support,
mathematics 0.57 0.000 Significant updates are crucial. glitches slow down
performance of (Moderate Continuous
- Lack of tech support leads lessons. (P4)
the students correlation) Support and
to interruptions in learning. - Flexibility is essential for
Versatility
- AI tools must be adaptable diverse educational needs.
*Significant at 0.05 level. to different curricula. (P1)
A p-value of 0.000 shows a strong association between
teachers using AI-based approaches and enhanced student Enhanced Differentiation and Interest
performance in mathematics. Increased exposure to AI- Using AI technologies, teachers could develop differentiated
driven learning increases student engagement, concept learning paths that cater to each student's needs, abilities,
retention, and problem-solving skills by filling in knowledge and pace. Adaptive algorithms change lessons according to
gaps and personalizing resources to the individual. This ability levels, giving more opportunities for practice and
shows how AI is effective in providing adaptive, feedback where needed and more challenging work with
personalized instruction. higher achievers. As a result, teachers ensure that students'
The findings stress the importance of teachers in utilizing AI needs are met while keeping them maximally challenged.
for student-centered learning. With AI, teachers can AI promotes learning through interactive, gamified activities
automate routine tasks and focus on mentorship, leading to and instant feedback, promoting accountability and
better outcomes. This calls for integrating AI-based methods motivation. For the teacher, AI simplifies task preparation
into math curricula and teacher training to maximize their for diverse learners, freeing up more time to interact with
impact. and support students. AI promises to enhance engagement
Challenges and Opportunities for Adopting AI-Driven and personalize math teaching and learning [7].
Technologies in Personalized Mathematics Learning Technological Sophistication and Resource Intensity
This section provides key insights based on the interviews The technical demands of AI tools for personalized
conducted with participants, identified through codes such instruction pose a considerable challenge in adoption.
as P1, P2, P3, and so forth, to ensure anonymity. Schools with limited resources often do not have the
Additionally, the quotations have been edited to make them required high-end software, hardware, reliable internet, data
clearer and translated for easier understanding in English. storage, and cybersecurity for such platforms. This outdated
The key themes identified from the interview data highlight infrastructure and high costs make adopting such
their significance in relation to each research objective. technologies particularly challenging for underserved
These themes provide valuable and often unexpected communities.
perspectives on how AI-driven personalized math Many teachers are not trained in the operation and
instruction affects long-term learning outcomes across troubleshooting of AI tools, which complicates
diverse student populations. implementation. Proper infrastructure and professional
AI in Personalized Mathematics Teaching and Learning development are required to integrate AI. Strategic
Themes Key Points Sample Quotes planning, investment, and collaboration with technology
- AI adapts lessons to each providers are necessary to make AI tools accessible to more
- AI adjusts material to
student's skill level for
every student's level, schools and educators.
engagement.
ensuring appropriate Teacher Readiness and Professional Development
- Provides additional practice
challenges. (P5)
Enhanced for struggling learners and
- Students work at their
The impact of AI on personal math instruction will largely
Differentiation advanced material for high be determined by teachers' willingness to use the tools
own pace with content
and Interest achievers.
- Encourages accountability
matching their skills. (P3) proficiently. Most teachers lack the confidence and
- Adaptive algorithms experience with AI systems and, therefore, require training.
and motivation through
adjust lessons to
gamified tasks and instant
performance. (P7)
Training programs, workshops, and technical support
feedback. increase teachers' skills and confidence while using AI [8].
Technological - AI tools require advanced - Our school lacks the
Information Classification: General
Teacher education must adapt to reflect new advancements access to technical assistance, and continued professional
as AI develops. This involves adding personalized learning development programs to instruct educators on how best to
to conventional approaches to pedagogy. Schools should utilize the capability of AI tools for the delivery of
focus more on professional development to allow teachers to mathematics instructional services. Accordingly, institutions
embrace AI, making learning more engaging and effective. must make a stand on issues regarding data privacy and
Ethical Issues and Data Protection security while at the same time attaining trust in the
Personalized AI instruction raises ethical concerns, institution-stakeholder relationship. Student data would
especially on data privacy and security [1]. AI platforms hence be kept safe. These would further the application of
collect large amounts of student data, raising questions AI in sustainable ways for personal mathematics learning in
about protecting sensitive information. There is a need to an equitable and effective learning environment at Isabela
ensure that AI systems comply with privacy regulations and State University.
maintain trust with teachers and parents.
In addition, AI assessments can be biased; therefore, they REFERENCES
may tend to evaluate students according to data patterns [1] Baker, R. S., Smith, L., & D'Mello, S. K. (2019). Educational data
mining and learning analytics: Applications and opportunities. Springer.
rather than objective measures. Schools, educators, and
technology providers must collaborate in implementing data [2] Bangayan- Manera, A. (2020). Writing Without Permission: A Case
policies that are transparent, fair, and safe for all learners. Study on Skinner's Analogy through Vandalism. International Journal of
Continuous Support and Versatility of AI Tools Psychosocial Rehabilitation. 24 (08), 571-578
One of the challenges when using AI in personalized math [3] Bangayan-Manera, A, Vecaldo, R. & Saquing J. (2020). Cultural
teaching is system maintenance and adaptability. Since AI Structuring of Urok Practices: An Intercultural Communication of Bago
tools require regular updates, troubleshooting, and technical Tribe in Kalinga, Philippines. International Journal of Psychosocial
support, it can be challenging to do so if proper resources Rehabilitation.24 (6), 13193-13217
are not established. This calls for dependable technical [4] Bangayan-Manera, A. (2019). Doodle: Towards A Hermeneutical
support to minimize disruption or interference with the Appreciation in Jacques Derrida’s Deconstruction. The Asian EFL Journal.
instruction process. 24 (4.2), 291-204.
AI systems should also be versatile in terms of their learning
[5] Bangayan-Manera, A. (2019). Textual Analysis of School Graffiti. The
models, curricula, and student needs. If a tool is inflexible, it Asian EFL Journal. 21 (2.3), 273-285.
cannot be applied to different classes and grades. Effective
partnerships between schools and technology providers must [6] Domingo- Alejo, J. (2022). Delving the Perceived Administrator’s
be ensured for uniform support to improve AI in Personality in Basic Education Unit. CEMJ. 30 (4), 2052-2060.
personalized learning[13]. [7] Domingo-Alejo, J. (2024). AI Integrated Administration tool design
with ML Technology for Smart Education System. 2024 4th International
Conclusion Conference on Advance Computing and Innovative Technologies in
Engineering (ICACITE). 1423-1428.
The study discovered that AI-driven personalized learning
systems improved students' mathematics performance. [8] Eustaquio, G. L. (2024). A Deep Analysis of Implementation of
These systems enhanced engagement and motivation and Technical Strategy using AI in MBS. 2024 4th International Conference on
allowed for mastering concepts through customized learning Advance Computing and Innovative Technologies in Engineering
(ICACITE), Greater Noida, India, 2024, pp. 1472-1478, doi:
pathways and real-time feedback. At Isabela State 10.1109/ICACITE60783.2024.10617117.
University, teachers employed AI-supported approaches
such as adaptive learning paths, scaffolded exercises, and [9] Kerr, M. (2018). The promise of artificial intelligence in education: A
student progress monitoring to address learners' varying review of AI applications in teaching and learning. Journal of Educational
Technology, 45(3), 52-61.
needs.
Challenges have included the technical complexity of AI [10] Miller, J., & Richards, K. (2022). Adaptive Learning Systems and
tools, the necessity for frequent updates, the lack of Their Impact on Math Performance: A Meta-Analysis. Journal of Learning
resources, and concerns over data privacy. Despite these Sciences, 29(1), 75–95.
challenges, opportunities abound with AI tools to support [11] Pamittan, S.T., Caranguian, C.B, Andres, A.D., Saquing, J.B.,
differentiated instruction and more effectively serve Vecaldo, R.T., Canay, P.C. (2022). Correlates of Employability among the
advanced and struggling students. Bachelor of Technical Teacher Education Graduates of the Philippine
Recommendations Public University—International Journal of Learning, Teaching and
Educational Research, 214 94.
To effectively take optimal advantage of the opportunities
that AI can present in mathematics instructional settings, the [12] Pane, J. F., Steiner, E. D., Baird, M. D., & Hamilton, L. S. (2015).
study recommended enhanced institutional support for Continued progress: Promising evidence on personalized learning. RAND
technology infrastructure, including regular maintenance, Corporation
Information Classification: General
