8

Chapter 2

FRAMEWORK OF THE STUDY

This chapter includes the related literature and studies done by local and foreign

researchers, authors and contributors. This part also includes the discussions of

theoretical framework, conceptual framework, presentation of null hypothesis and

definition of terms.

Review of Related Literature and Studies

Artificial Intelligence in Education

Technology has become an indispensable part of our life (Rokan & Monang,

2023) and has an impact on numerous sectors of society (Simeone et al., 2022; Anderson

& Horrigan, 2016; Muydinovich et al., 2022), including education (Anderson, 2016;

Andre et al., 2015; Jobirovich, 2022). In recent years, particularly during and after the

Covid-19 pandemic (Urbina et al., 2022; Romero-Martin et al., 2022; Khosravi et al.,

2022), the integration of artificial intelligence (AI) and machine learning (ML) has

become more prevalent across different industries (Lee & Hwang, 2022).

Chatbots, which are designed to help users with information, customer support,

and problem-solving, are among the most important uses of artificial intelligence.

Chatbots are a common tool for individuals as well as organizations since they are made

to communicate with people using natural language and assistance. ChatGPT, an AI

program that has generated a lot of buzz worldwide, is the most recent advancement of

this technology in education.

 9

The integration of AI has had a significant impact on various sectors, including

healthcare, education, and communication. In the field of education (Evans, 2014), AI

solutions such as chatbots have been created to help students study (Dyson et al, 2015;

Greenwood et al, 2016). Chatbots are intended to offer students individualized support,

feedback, and direction. Chatbots can also assist students develop cognitive abilities by

presenting new ideas and information (Farley, 2013).

ChatGPT is an artificial intelligence (AI) chatbot developed by OpenAI that is

designed to understand and generate human-like text based on user input. It is built on the

Generative Pre-trained Transformer (GPT) architecture, which enables it to process

natural language and provide coherent and contextually relevant responses. ChatGPT is

widely used for various purposes, including answering questions, assisting with writing

tasks, coding, and generating creative content (OpenAI, 2025). Additionally, ChatGPT

has been used in the education sector to assist students and teachers in their learning

process.

Although the long-term impacts of this particular application have not yet been

detected, ChatGPT has demonstrated the potential to be a useful tool for teachers to

systematically analyze and provide feedback to students (Kasneci et al., 2023; Meyer et

al., 2024). Einarsson et al. (2024) reported that preliminary findings indicate ChatGPT is

an effective tool for generating classroom materials by reframing problems and creating

engaging content for learning activities. Its capacity to produce instructional materials

rapidly may prove to be a valuable tool for educators, enabling them to tailor activity

content to individual students and reach a higher degree of content personalization to

meet the needs of each student. These activities typically demand a significant amount of
 10

time, which teachers may not always have due to their severe workload (Baidoo-Anu &

Owusu Ansah, 2023).

In addition to that the administrative support skills of AI chatbots can be

beneficial to educators. Teachers can devote more time to instructional preparation and

student engagement by saving time on repetitive processes like scheduling, grading, and

informing students. For instance, ChatGPT can effectively provide a wide range of types

of questions and answer keys across multiple subject areas. To fit their particular

educational situations, educators should, nevertheless, critically assess and adapt their

methods. Since AI cannot yet replace a science teacher, the knowledge, experience, and

understanding of the instructor are important in making well-informed educational

decisions (Cooper, 2023).

AI chatbots can be used by educators to enhance their lessons and offer tailored

support. Teachers can use ChatGPT in a variety of ways to improve their teaching

practices, strategies and evaluation techniques (Herft, 2023). For example, educators can

use ChatGPT's ability to create open-ended question prompts that closely correspond to

the lesson unit's intended outcomes and focused learning objectives. Al Ka'bi (2023) and

Fariani et al. (2023) suggest that this approach enables educators to offer individualized

learning resources and activities tailored to each student's unique needs, interests, and

learning preferences.

Recent studies have explored the role of AI technologies in promoting self-

regulated learning (SRL) in digital learning environments. A study by [Ng et al., 2024]

compared the effectiveness of generative AI-based chatbots (SRLbot) and rule-based AI

chatbots (Nemobot) in facilitating SRL among Secondary 4 students in Hong Kong. Over
 11

a three-week science learning experience, students in the experimental group utilized

SRLbot, while the control group used Nemobot. Findings revealed that SRLbot

significantly enhanced students' science knowledge, behavioral engagement, and

motivation.

Role of AI and ChatGPT in Personalized Learning

Personalized learning, which aims to adapt teaching strategies and content to each

student's unique needs, interests, and preferences, has become a key strategy in today's

education. Personalized learning has shown significant potential in enhancing students'

conceptual understanding and engagement in science classes. Research indicates that

personalized education can significantly boost students' interest in subjects like science

and mathematics. A study by Hogheim and Reber (2015) demonstrated that context

personalization, adapting learning materials to align with students' interests, resulted in

increased triggered situational interest (TSI) and maintained situational interest related to

feelings (MSI-F) among middle school students with initially low interest in the subject

matter (Reber et al., 2018). Similarly, Bernacki and Walkington (2018) found that

personalizing learning contexts led to an increased in situational and individual interest in

mathematics among ninth-grade students.

The integration of artificial intelligence (AI) into personalized learning

environments has opened new avenues for enhancing science education. AI-powered

tools like ChatGPT can provide real-time feedback, answer student queries, and adapt

learning materials to individual needs, thereby fostering a more personalized learning

experience (ElSayary, 2024). While specific studies on ChatGPT's impact are still
 12

emerging, the broader application of AI in education has shown promise in improving

student engagement and understanding.

Kasneci et al. (2023) investigated the use of chatGPT at different educational

levels, emphasizing its beneficial effects on student learning. These included the use of

chatGPT to help elementary school students improve their reading and writing skills,

middle and high school students learn language and writing styles for a variety of

subjects, university students with research and writing assignments, facilitate group

discussions and debates, and combine chatGPT with text-to-speech or speech-to-text to

empower students with disabilities. Students of all skill levels can benefit from

chatGPT's ability to summarize and explain complex materials, provide practice

questions and quizzes, hone their research and problem-solving abilities, and support a

variety of open research issues.

Academic Performance and ChatGPT

Several studies highlight the growing role of AI-powered tools like ChatGPT in

enhancing students' academic performance. In a study conducted by Ahmad et al., 2025

which examines the impact of ChatGPT on students' creativity and academic

performance. Findings suggest that ChatGPT is widely regarded as an effective academic

tool, enhancing creativity through brainstorming and idea generation, though its role in

creative problem-solving is moderate. It was also found out that it improved academic

performance by saving time and enhancing clarity in learning. A strong positive

correlation was found between ChatGPT usage, creativity, and academic performance.

Similarly, the study by Caratiquit and Caratiquit (2023) have shown that ChatGPT serves
 13

as an effective educational tool, with its impact on students' academic performance being

significantly influenced by their learning motivation. The use of ChatGPT has been found

to positively affect academic outcomes by fostering greater engagement and motivation

in learning. The study by Madkaikar (2024) examined the impact of ChatGPT on

academic performance among undergraduate students in Goa. It explored how ChatGPT

influences learning dynamics, study habits, and overall academic growth. The research

found that ChatGPT significantly enhances academic support, skills development, and

performance. However, in the study conducted by Forero and Herrera-Suárez (2023)

where they explored the impact of ChatGPT on student performance in a second-semester

engineering physics course. Findings revealed that students who used ChatGPT exhibited

a significant decline in academic performance, as indicated by lower grades and negative

Hake factors when compared to the control group. Additionally, student surveys showed

mixed perceptions of ChatGPT’s effectiveness in learning physics. While many

acknowledged its usefulness in explaining concepts and providing information, concerns

were raised regarding its potential to diminish critical thinking and independent learning

(Forero & Herrera-Suárez, 2023). These findings highlight the need for a balanced

approach in integrating AI-based tools in education to ensure they enhance, rather than

hinder, student learning outcomes.

Boosting Student Engagement with AI Tools

Engagement is a critical factor in effective learning, and AI tools like ChatGPT

have been shown to positively influence this aspect. Villan and dos Santos (2023)

implemented ChatGPT as a co-advisor in project-based learning within elementary

 14

education. The study reported increased student engagement and participation, attributing

these outcomes to the interactive nature of ChatGPT (Villan & dos Santos, 2023).

Additionally, Lee and Zhai (2024) examined pre-service teachers' integration of ChatGPT

into science lesson planning. Their study found that ChatGPT not only enhanced lesson

interactivity but also supported teachers in creating more engaging learning experiences

for students (Lee & Zhai, 2024). Similarly, Kotsisi (2024) demonstrated that integrating

ChatGPT into primary school physics through AI-driven simulations and interactive

activities effectively transforms abstract concepts into tangible learning experiences,

enhancing student involvement, inquisitiveness, analytical thinking, and positive attitudes

toward physics, while underscoring the need for ongoing assessment of engagement and

academic outcomes.

AI in Science Education Assessment

AI's role extends beyond instruction to the realm of assessment, offering

innovative approaches to evaluate student learning. Zhai (2024) discusses the

transformative potential of AI and Machine Learning (ML) in science assessments,

emphasizing the need for new types of assessments that align with the Framework for K-

12 Science Education. Traditional methods like multiple-choice questions often fail to

capture the complexities of scientific thinking. AI-powered performance-based

assessments can engage students in scientific practices such as modeling, explanation,

and argumentation, providing timely and objective feedback that alleviates the burden on

educators.
 15

In a research conducted by Dijkstra et al. (2024) where they created an end-to-end

multiple-choice test generator. It was found out that many of the quizzes that were

developed showed a high degree of coherence and reasonableness. By using quizzes to

train and assess their knowledge, especially when it comes to exam preparation, the

authors stress that their work positions chatGPT as a useful tool for both teachers and

students, allowing them to improve their learning experience.

ChatGPT can also be used as an automatic grading system in which

students' tasks such as research articles, essays, and a wide range of written

assignments could be checked for strengths and weaknesses (Kasneci, 2023).

AI-Powered Chatbots on Student Learning Outcomes

Recent literature indicates that AI-powered chatbots are increasingly recognized

for their potential to enhance student learning outcomes through personalized, adaptive

instruction and immediate feedback. For instance, Luckin et al. (2016) argue that AI

technologies can transform education by providing customized learning experiences that

cater to individual student needs, thereby improving academic performance. Chatbots like

ChatGPT, specifically, have been shown to offer tailored support by answering questions,

reframing problems, and providing scaffolding that can deepen conceptual understanding

(Holmes, Bialik, & Fadel, 2019).

Several studies have investigated ChatGPT's effectiveness in promoting deeper

understanding of scientific concepts. For instance, Albdrani and Al-Shargabi (2023)

conducted a case study examining ChatGPT's role in data science education. Their

findings suggest that ChatGPT can provide personalized feedback and explanations,
 16

thereby enhancing students' grasp of complex topics (Albdrani & Al-Shargabi, 2023).

Similarly, Vasconcelos and dos Santos (2023) explored the use of ChatGPT as an

"object-to-think-with" in STEM education. Their research indicates that ChatGPT

facilitates reflective and critical thinking, contributing to improved conceptual

understanding among students (Vasconcelos & dos Santos, 2023).

Student Perceptions of AI-Assisted Learning Tools

Virtual assistants and chatbots have the ability to revolutionize support services,

increase accessibility, and create more effective and efficient learning environments

when integrated into educational settings (Chen et al., 2023; Essel et al., 2022). A few

studies that have already been conducted that look at how students feel about using

ChatGPT in the learning process show that students think positively of the app, value its

features, and find it useful for their work and studies (Kasneci et al., 2023; Shoufan,

2023). Students recognize that ChatGPT cannot replace human intelligence and stress the

necessity for strong background knowledge to use it effectively, acknowledging that its

responses are not always accurate (Shoufan, 2023). Scholars have identified the

following as the most significant benefits: Assistance with homework and studies.

Chatbots with AI capabilities can give thorough feedback on student assignments,

pointing out areas that need work and making recommendations for additional education

(Celik et al., 2022).

For instance, ChatGPT can be a useful study aid by offering clarifications and

explanations on various kinds of subjects. It can help students with homework

assignments by providing detailed answers and assisting them in navigating complex

 17

problems (Crawford et al., 2023; Fauzi et al., 2023; Lo, 2023; Qadir, 2023; Shidiq, 2023).

Additionally, ChatGPT allows students to test themselves on a variety of subjects,

strengthening their understanding and getting ready for tests (Choi et al., 2023;

Eysenbach, 2023; Sevgi et al., 2023; Turzo et al., 2023). It can also be used in skills

development. It can help improve writing skills by suggesting syntactic and grammatical

corrections (Kaharuddin, 2021), develop problem-solving skills by offering detailed

solutions (Benvenuti et al., 2023), and support group discussions and debates by offering

discussion structures and real-time feedback (Ruthotto et al., 2020; Wang et al., 2020).

Lastly, AI can help in providing personalized learning. Students can now receive

personalized guidance and feedback from AI-powered chatbots in general, which helps

them get through difficult ideas and enhance their comprehension.

The State of AI Adoption in Philippine Public Schools

The integration of artificial intelligence (AI) in education has gained momentum

globally, and the Philippines has taken significant steps to incorporate AI-driven tools

into its public school system. The Department of Education (DepEd) has been at the

forefront of this transformation, launching the Education Center for AI Research (E-

CAIR) in February 2025, marking the first dedicated AI research hub for education in the

country (Department of Education [DepEd], 2025). The primary objective of E-CAIR is

to revitalize basic education by developing AI-powered tools that enhance teaching

methodologies, student learning experiences, and school administrative processes. This

initiative aligns with the Philippine government's commitment to modernizing education,

 18

ensuring that Filipino students have access to global standards of technology-enhanced

learning (DepEd, 2025).

Higher education institutions in the Philippines have also made significant strides

in AI adoption. The University of the Philippines (UP) has taken a leadership role in AI

education, offering undergraduate and graduate programs in AI, including the country’s

first Ph.D. program in the field (University of the Philippines, 2023). The establishment

of the UP Center for Intelligent Systems further underscores the institution's commitment

to AI research and development. These efforts contribute not only to producing AI-

literate graduates but also to fostering AI-driven innovations that could enhance

pedagogical strategies in secondary education (University of the Philippines, 2023).

Challenges and Opportunities of AI-Powered Learning in the Philippines

Despite the promising advancements in AI adoption, challenges remain,

particularly in terms of infrastructure and accessibility. As of February 2025,

approximately 2,000 public schools in the Philippines still lack electricity, while 40,000

schools require upgrades to their electrical systems, posing significant barriers to AI-

driven education (Philippine News Agency [PNA], 2025). The digital divide remains a

critical issue, as many rural schools lack the necessary technological infrastructure to

support AI-powered learning. Addressing these gaps is essential for ensuring that AI

adoption benefits all students, regardless of their geographic and socioeconomic

backgrounds (PNA, 2025).

 19

Implications and Concerns

One of the primary concerns regarding students' use of ChatGPT is academic

honesty and integrity, an issue highlighted early on by Stokel-Walker (2023). A recent

survey found that approximately 89% of college students in the United States use

ChatGPT to complete their homework assignments, with 53% relying on it to write

essays, 48% utilizing it during exams, and 22% using it to generate paper outlines

(Tangermann, 2023). Similarly, Stanford University has expressed concerns about

students leveraging ChatGPT for various academic tasks, raising ethical issues regarding

its use in coursework (Cu & Hochmann, 2023). Without proper citation and

acknowledgment, AI-assisted work constitutes plagiarism, a challenge that is becoming

increasingly difficult to detect as ChatGPT's responses grow more sophisticated and

human-like (Cotton et al., 2023; Else, 2023; Shiri, 2023). This misuse of AI technology

further complicates the fairness of learning assessments, making it difficult to accurately

evaluate students' knowledge and skills.

AI chatbots have the potential to give inaccurate or biased information (Kasneci et

al., 2023; Sedaghat, 2023). Students may be misled and have their learning progress

impeded if the chatbot gives them inaccurate information or advice. Even though

ChatGPT provided interesting and thought-provoking responses, Sevgi et al. (2023)

contend that it shouldn't be considered a trustworthy source of information. An AI

chatbot may unintentionally replicate biases in its responses if the training data used to

create it contains them. These biases could include distorted viewpoints, stereotypes,

discriminating language, or biased recommendations. In the context of education, this is

especially concerning.
 20

Another difficulties instructors encounter while integrating chatbots into the

classroom is assessing students' work, especially when it comes to written assignments

responses. Even though AI-generated text identification is always becoming more

advanced it is still not perfect and can result in false positives or negatives. This breeds

doubt and may damage the legitimacy of the assessment procedure. The accuracy of

grading and feedback may suffer if teachers are unable to determine whether the

responses were actually created by students or whether an AI gave them. Fair assessment

procedures and academic integrity are called into question by this (AlAfnan et al., 2023;

Kung et al., 2023).

Theoretical Framework

The following theories serve as the foundation of this research study. Theories on

teaching performances and academic achievement of students were included to give light

on where the research study is coming from and revolving into.

Constructivism and Active Learning. Constructivist theory states that learners

actively construct knowledge based on their experiences and prior understandings

(Piaget, 1972). In this view, learning is not a passive absorption of facts but an active

process of making sense of new information by integrating it with what is already known.

By engaging in interactive dialogue with ChatGPT, students are provided with

opportunities to question, reflect, and build upon their existing knowledge. This process

aligns with the constructivist approach, as it encourages active engagement with scientific

concepts and promotes deeper understanding through inquiry-based learning.

 21

Personalized learning theories advocate for instruction that adapts to individual

student needs, learning paces, and interests (Pane, Steiner, Baird, & Hamilton, 2015). AI-

powered systems like ChatGPT analyze student interactions in real time and adjust

instructional strategies accordingly. This adaptive instruction not only addresses

individual learning gaps but also enhances the relevance and accessibility of scientific

content, thereby improving overall conceptual understanding. The dynamic nature of AI

enables continuous personalization, making learning both more engaging and effective.

Conceptual Framework

This study investigates how AI-powered personalized learning implemented

through ChatGPT affects the academic performance and engagement of Grade 11

students in Earth and Life science.

• Performance in Earth and Life Science

• Pre – Test Results
• Student Engagement

• Use of ChatGPT
• Creation of ChatGPT Usage Guide
• Implementation of ChatGPT as an intervention

• Post – Test Results

• Survey Questionnaire

• Mean
• Standard
Deviation
• Frequency
• Percentage

Effect of ChatGPT on the Academic Performance

of Grade 11 Earth and Life Students
 22

Figure 1. Paradigm of the Study

This conceptual framework outlines the research process for examining the impact

of ChatGPT on the academic performance of Grade 11 Earth and Life Science students.

The study begins with a pre-intervention phase, where students' initial academic

performance is assessed through pre-test results, and their level of engagement is evaluated.

Following this, the intervention phase introduces ChatGPT as a learning tool, accompanied

by the creation of a usage guide to ensure proper implementation. ChatGPT is then

integrated into students' learning activities as an intervention to support their academic

development.

After the intervention, a post-test is conducted to measure any improvements in

academic performance, along with a survey to gather students' perceptions and experiences

with ChatGPT. To analyze the collected data, various statistical methods will be employed,

including mean and standard deviation to assess overall performance, frequency and

percentage to summarize survey responses, and a paired t-test to compare pre-test and post-

test scores to determine the significance of any improvements. Ultimately, the study aims

to evaluate the overall effect of ChatGPT on student learning outcomes, providing insights

into its role as an educational tool in Earth and Life Science instruction.