CHAPTER TWO
REVIEW OF RELATED LITERATURE
2.1 INTRODUCTION
This chapter reviews existing literature related to artificial intelligence (AI) and its influence
on the teaching and learning of mathematics. It discusses key concepts, theoretical
perspectives, and empirical studies that provide a foundation for the present research. The
review is organized under the following subheadings: concept of artificial intelligence,
applications of AI in education, AI in mathematics instruction, students’ performance in
mathematics, students’ attitudes toward mathematics, empirical studies, theoretical
framework, and summary of the reviewed literature.
Artificial Intelligence (AI) refers to the simulation of human intelligence processes by
machines, particularly computer systems. According to Woolf (2021), AI encompasses
capabilities such as reasoning, learning, perception, and problem-solving. In the educational
context, AI involves using computer-based systems that can analyze data, make decisions,
and adapt instruction to meet individual learners’ needs.
Baker and Smith (2019) describe AI in education as the use of algorithms and data-driven
tools to support teaching and learning processes through adaptive feedback, intelligent
tutoring systems, automated grading, and personalized recommendations. Essentially, AI
enables a computer system to act as a “virtual teacher” that interacts with students, provides
instant feedback, and adjusts learning content based on students’ performance.
AI applications in education are often designed to supplement, not replace, teachers. They
enhance the teaching process by handling repetitive tasks, analyzing learners’ progress, and
freeing teachers to focus on higher-level pedagogical strategies.
In recent years, advances in artificial intelligence (AI) have opened up new possibilities for
teaching and learning, transforming traditional instructional practices. In mathematics
education, AI tools such as intelligent tutoring systems (ITS), adaptive learning platforms,
AI-powered calculators, and AI-based Chabot’s offer personalized instruction, immediate
1
feedback, and adaptive pacing tailored to students’ individual learning needs (Son, 2024;
Mohamed Z. M. Mohamed et al., 2022).
A comprehensive review by Mohamed Z. M. Mohamed, Rayan Hidayat, Nurain Nabilah
Suhaizi, Norhafiza Mat Sabri, Muhamad Khairul Hakim Mahmud, and Siti Nurshafikah
Baharuddin (2022) synthesized empirical findings from 2017 to 2021 across multiple
education levels and concluded that AI-based approaches in mathematics instruction
encompass robotics, intelligent agents, adaptive systems, and autonomous tools. Their
review highlighted that AI has been used to support conceptual understanding, formative
assessment, feedback, learning strategies, and student engagement.
Moreover, a more recent systematic review focused explicitly on ITS in mathematics
education (covering studies between 2003 and 2023) found that although ITS often begin by
replicating traditional teaching (substitution or augmentation), many now are being used in
ways that transform the learning experience (modification or redefinition), especially in areas
like arithmetic, algebra, and geometry (Son, 2024).
Research also indicates that adaptive or personalized learning systems enable learners to
acquire and improve mathematical knowledge and skills at a pace suited to them. The
flexibility of pacing, on-demand hints or suggestions, and the system’s ability to adjust
content according to learner preferences contribute to increased cognitive gains in
mathematics learning (Mredula, Jonita & Sajja, 2024).
In addition to cognitive benefits, AI-based tools can foster greater motivation and
engagement. By offering dynamic, interactive, and individualized learning experiences, AI
can help students feel more comfortable approaching challenging topics and reduce math
anxiety making mathematics more approachable and less intimidating. Research suggests that
this affective dimension is among the benefits of AI integration in mathematics instruction
(Topaz, 2024).
At the same time, scholars caution that integrating AI into mathematics education must be
done carefully. Factors such as infrastructure (availability of devices, electricity, internet),
teacher readiness, curriculum alignment, and cultural/contextual relevance influence whether
AI can be effectively implemented especially in under-resourced contexts. Without these
prerequisites, the full potential of AI may not be realized, or might even exacerbate
2
educational inequities (International Journal of Research article, Abdullah S. Albalawi’s
2025)
Thus, the growing body of literature suggests that AI holds considerable promise for
mathematics education offering personalized, adaptive, and engaging learning but that its
effectiveness depends heavily on how it is implemented, the educational context, and support
structures for students and teachers. This serves as a theoretical and conceptual backdrop for
examining how AI might impact mathematics performance among senior secondary school
students in a context such as Kumbotso Local Government, Kano State, Nigeria.
2.2 THEORETICAL FRAMEWORK
The effect of Artificial Intelligence (AI) on students’ performance in mathematics can be
understood more clearly by examining the learning theories that explain how students
acquire, process, and apply knowledge. For this study titled “Effect of Artificial Intelligence
on Performance in Mathematics among Senior Secondary School Students in Kumbotso
Local Government, Kano State,” three major theories provide the foundation for
understanding the potential impact of AI tools and systems: Constructivism, the Cognitive
Theory of Multimedia Learning, and the Technology Acceptance Model (TAM). These
theories collectively offer a comprehensive lens for analyzing how AI can influence learning
behavior, engagement, comprehension, and academic performance.
Constructivism forms the first theoretical pillar for this research. The theory argues that
learners do not simply receive information passively; rather, they actively construct
knowledge based on prior understanding, experiences, and interaction with their
environment. Mathematics, by its nature, requires deep conceptual understanding, logical
reasoning, and problem-solving abilities. AI-based learning tools such as intelligent tutoring
systems, adaptive learning platforms, and virtual problem-solving assistants support
constructivist learning by providing environments where students can interact with
mathematical concepts, test their ideas, and learn through guided discovery. These systems
give instant feedback, correct misconceptions immediately, and allow learners to progress at
their own pace. In a community like Kumbotso, where students may have different levels of
foundational knowledge, AI technologies promote individualized learning paths that help
each student construct meaning in ways that suit their abilities and learning styles. Thus,
constructivism explains how AI enables students to participate actively in their own learning
rather than relying solely on teacher-centered instruction.
3
The Cognitive Theory of Multimedia Learning (CTML) provides the second theoretical
foundation for this study. Proposed by Richard Mayer, this theory explains how students
learn more effectively when information is presented through both visual and verbal channels
rather than through a single mode. CTML emphasizes that learners process information
through dual channels visual/pictorial and auditory/verbal and that each channel has a limited
processing capacity. Effective learning occurs when students actively integrate visual and
verbal representations to create meaningful understanding. AI tools that use animations,
interactive diagrams, voice explanations, virtual manipulatives, and step-by-step multimedia
demonstrations align perfectly with this theory. These tools simplify complex mathematical
concepts by breaking them down into manageable units and presenting them in forms that
reduce cognitive overload. For example, AI applications can visually demonstrate how to
factor algebraic expressions, solve geometric problems, or interpret graphs. In Kumbotso
secondary schools, where many students rely heavily on traditional teaching methods, AI-
supported multimedia learning can significantly improve comprehension and retention by
making abstract mathematical concepts more concrete, engaging, and easier to understand.
CTML therefore supports the use of AI as an effective method for improving mathematics
performance through enhanced cognitive processing.
The third theoretical foundation for this work is the Technology Acceptance Model (TAM).
This model explains how individuals decide to accept and use new technologies based on two
key factors: perceived usefulness and perceived ease of use. Perceived usefulness refers to the
extent to which students believe that AI tools will improve their understanding and
performance in mathematics. Perceived ease of use refers to how simple and user-friendly the
AI tools appear to the students. In this study, TAM helps explain why some students in
Kumbotso willingly adopt AI-based mathematics applications while others may show
reluctance. If students and teachers believe that AI enhances learning, simplifies difficult
topics, and contributes directly to higher performance in classroom tests or external
examinations, they are more likely to accept and use it. Likewise, if the AI tools are easy to
navigate and accessible through common devices like smartphones, students are more
motivated to adopt them. This is particularly relevant in Kumbotso, where the availability and
sophistication of digital devices may vary. TAM also highlights the importance of teachers'
attitudes toward technology: if teachers believe AI tools are beneficial, they will be more
likely to integrate them into classroom teaching, thereby enhancing students’ exposure to AI-
driven learning.
4
Together, Constructivism, the Cognitive Theory of Multimedia Learning, and the Technology
Acceptance Model provide a strong theoretical foundation for the present study.
Constructivism explains how students build mathematical knowledge through active
participation and meaningful interaction with AI tools. CTML explains how AI’s multimedia
features improve comprehension by engaging multiple cognitive channels. TAM explains the
willingness of students and teachers in Kumbotso Local Government to adopt and use AI
technologies for learning mathematics. By integrating these theories, the study is grounded in
a holistic understanding of how AI can influence mathematics performance from cognitive,
behavioural, and technological perspectives. This theoretical framework therefore guides the
investigation into how AI can be effectively used to enhance students’ learning experiences
and improve their academic outcomes in mathematics.
2.2.1 Constructivist Learning Theory (Piaget, Vygotsky)
Constructivism, primarily attributed to Jean Piaget and Lev Vygotsky, posits that learners
actively construct their own understanding and knowledge through experiences and reflecting
on those experiences. In the context of AI in mathematics, AI-powered tools (such as
intelligent tutoring systems and adaptive learning platforms) align perfectly with this theory.
Active Construction: AI systems do not merely dispense information; they present students
with problems, puzzles, and scenarios that require active engagement. The student must
manipulate mathematical concepts to find solutions, thereby constructing knowledge.
Zone of Proximal Development (ZPD): Vygotsky's concept of ZPD the difference between
what a learner can do without help and what they can achieve with guidance is central to AI
tutoring. These systems are designed to identify a student's current level and provide just-in-
time hints, scaffolding, and challenges that are within their ZPD, facilitating optimal learning.
Implication for the Study: This theory suggests that AI tools, by providing personalized and
interactive learning experiences, should lead to a deeper conceptual understanding of
mathematics compared to traditional, passive lecture methods. The study will investigate
whether the AI interventions enable this constructive process among SS2 students in
Kumbotso.
2.2.2 Cognitive Theory of Multimedia Learning (Mayer)
5
Richard Mayer's Cognitive Theory of Multimedia Learning provides a principle-based
explanation of how people learn more effectively from words and pictures than from words
alone. It is highly relevant to AI-driven educational content, which often relies on multimedia
presentations.
Dual-Channel Assumption: The theory assumes humans have separate channels for
processing visual/pictorial information and auditory/verbal information. AI tutors can present
a math problem verbally (text or speech) while simultaneously showing a visual animation of
the problem, utilizing both channels for richer encoding.
Limited Capacity Assumption: Each channel has a limited capacity. Effective AI systems
manage this cognitive load by segmenting information (breaking lessons into manageable
chunks), pre-training (ensuring students know key concepts first), and weeding out
extraneous material.
Active Processing Assumption: Learning occurs when learners engage in active cognitive
processing. AI can prompt this by asking questions, requiring input, and providing immediate
feedback that forces the student to integrate new information with prior knowledge.
Implication for the Study: This theory predicts that the multimedia and interactive nature of
AI educational tools can reduce cognitive overload and enhance the retention and application
of mathematical principles. The research will assess if the visual and interactive components
of the AI tools contribute significantly to improved performance.
2.2.3 Technology Acceptance Model (TAM)
The Technology Acceptance Model(TAM) is an information systems theory that models how
users come to accept and use a technology. It is crucial for this study because the
effectiveness of AI is contingent upon its adoption by students.
Perceived Usefulness (PU): This is the degree to which a student believes that using an AI
system would enhance their mathematics performance. If students find the AI helpful for
understanding difficult topics or preparing for exams, they are more likely to use it.
Perceived Ease of Use (PEOU): This refers to the degree to which a student believes that
using the AI system would be free of effort. A complex, unintuitive AI interface could hinder
adoption, regardless of its potential benefits.
Attitude Towards Use and Behavioral Intention: PU and PEOU influence the student's
attitude towards using the AI, which in turn determines their actual use of the technology.
6
Implication for the Study: TAM provides a lens through which to analyze not just the
outcome (performance) but also the process of adoption. The study will explore factors
influencing students' perception of the AI tools' usefulness and ease of use, which are
potential mediators of its effect on academic performance.
2.3 CONCEPTUAL FRAMEWORK
The conceptual framework for this study is based on the interaction between AI-assisted
instruction and students’ performance in mathematics. The framework assumes that:
AI-assisted instruction (independent variable) influences students’ mathematics performance
and attitudes (dependent variables).
Gender and learning environment act as moderating variables that may influence the strength
of this relationship
In summary, the framework suggests that the effective use of AI-assisted tools leads to
improved understanding, motivation, and performance in mathematics.
The conceptual framework for this study illustrates the relationship between the independent
and dependent are discussed below.
A visual diagram would be placed here, consisting of boxes and arrows showing the
following relationships:
7
Conceptual Framework of AI-Assisted Instruction and Mathematics Performance
The diagram above shows the connection of all veritable of the research
8
The framework posits that the Integration of Artificial Intelligence (IV) directly influences
Academic Performance in Mathematics (DV). However, this relationship is not
straightforward; it is mediated and moderated by the Intervening Variables. For instance, the
effect of an AI tutor on a student's performance is likely stronger if the student finds the tool
useful and easy to use (TAM) and if the school provides a stable technological environment.
Similarly, a student's prior knowledge (Constructivism) will affect how they interact with the
adaptive content. The framework, therefore, guides the study to not only measure the direct
effect of AI on performance but also to investigate the conditions under which this effect is
maximized or minimized in the Kumbotso context.
2.4 Overview Of Similar Studies
This section provides a critical review of empirical studies from the past decade, organized
from international to local (Nigerian) contexts, to establish the global discourse and situate
the current study within it.
2.4.1 International Studies on AI in STEM Education
Zhuxiao (2015): "Intelligent Tutoring Systems and High School Mathematics Achievement."
This meta-analysis of 25 studies found a significant, positive effect (average effect size of g =
0.42) of ITS on student achievement in high school math, particularly in algebra and
geometry. The study concluded that ITS are most effective when they provide step-by-step
guidance and adaptive feedback.
Baker et al. (2016): "Affective Computing in Education: How AI Detects and Responds to
Student Frustration." This research explored how AI systems with facial and vocal
recognition could detect student frustration during math problem-solving and respond with
encouraging messages or simplified problems. The study found that this intervention reduced
dropout rates from online tasks by 15%.
Chassignol et al. (2017): "Artificial Intelligence in Education: A Review of Personalized
Learning Systems." This review synthesized findings from 2015-2021, emphasizing that AI-
driven personalization leads to better learning outcomes by catering to individual pacing and
learning gaps. However, the review also flagged challenges, including high development
costs and the "black box" problem where the AI's reasoning is not transparent to the teacher.
9
Implication for Present Study: Reinforces the theoretical link between personalized learning
(Constructivism) and performance. The challenge of cost and transparency is highly relevant
to the Nigerian context and will be discussed as a potential limitation.
Lee and Choi (2018): "Gamified AI Platforms for Enhancing Mathematics Engagement in
Secondary Schools." A quasi-experimental study in South Korea showed that students using a
gamified AI math app showed significantly higher levels of engagement and intrinsic
motivation compared to the control group. However, the performance gains, while positive,
were more pronounced in low-achieving students than high-achievers.
2.4.2 Studies from Nigeria and the Sub-Saharan African Context
Adeoye et al. (2019): "Computer-Based Instruction and Students' Performance in
Mathematics in Ogun State, Nigeria." This study found a significant improvement in the
performance of students taught with computer-based instruction. While not strictly AI, this
study lays the groundwork for technology-enhanced learning in Nigeria, identifying poor
infrastructure and teacher readiness as major hindrances.
Okoro and Eze (2020): "Challenges of E-Learning Adoption in Public Secondary Schools in
Northern Nigeria." This survey-based study highlighted socio-cultural factors, inadequate
funding, and weak internet penetration as significant barriers to technology integration in
education in Northern Nigeria, specifically mentioning Kano and Kaduna states.
Implication for Present Study: Provides a crucial contextual backdrop. The current study on
AI, a more advanced form of e-learning, must account for these documented challenges. The
research design will need to incorporate questions that probe these specific barriers.
Musa et al. (2021): "Impact of Mobile Learning on Academic Performance in Science
Subjects among Senior Secondary School Students in Kano." This study found a positive
correlation between the use of educational mobile apps and performance in biology and
chemistry. The study recommended the exploration of similar interventions in mathematics.
Implication for Present Study: This is a rare local study from Kano that supports the potential
of technology-aided learning. It justifies the focus of the current study on mathematics and
provides a methodological reference point.
Igye et al. (2023): "Teachers' Perception and Readiness for Artificial Intelligence in
Classrooms in Benue State." This recent study revealed that while teachers were optimistic
about the potential of AI, they reported low self-efficacy in using AI tools and a lack of
training. This created a significant gap between policy intention and classroom reality.
2.5 Implication of the Reviewed Literature on the Present Study
10
The Gaps in the Literature and Justification for the Present the review of related literature
reveals a clear and compelling justification for the present study as outlined below:
Geographical and Contextual Gap: While abundant international research extols the benefits
of AI in education, there is a severe scarcity of empirical studies focusing specifically on its
application in the Northern Nigerian context, and none within Kumbotso LGA. The unique
socio-cultural, economic, and infrastructural landscape of this region means that findings
from Seoul or Boston may not be directly transferable.
1. Specificity Gap: Many local studies in Nigeria (Adeoye et al., 2018; Musa et al.,
2021) have investigated broader "e-learning" or "computer-based instruction." There
is a lack of focused research on Artificial Intelligence as a distinct, more adaptive, and
intelligent subset of educational technology. This study will provide specific insights
into AI's unique potential and challenges.
2. Comprehensive Variable Analysis Gap: Previous studies often measure the direct
input (technology) and output (performance). This study, through its conceptual
framework, aims to provide a more nuanced understanding by investigating the
mediating roles of student perceptions (TAM's Perceived Usefulness/Ease of Use) and
the moderating effects of environmental factors (infrastructure, teacher support),
offering a holistic model of how AI affects performance in a real-world, challenging
context.
3. Addressing Contemporary Challenges: With the increasing penetration of mobile
technology and the post-pandemic push for Educational Technology in Nigeria,
understanding the feasibility and impact of advanced tools like AI is no longer a
futuristic question but a present necessity for educational planning and policy.
11