GGGA6093 Educational Research and Publication 1 Assignment 2026 | UKM

Request Answer of this Assignment

GGGA6093 ASSIGNMENT | RESEARCH PROPOSAL 

TITLE:  EXPLORING ESL SECONDARY SCHOOL TEACHERS’ ATTITUDES TOWARDS THE USE OF AI GENERATIVE TOOLS IN ENGLISH LANGUAGE TEACHING

CHAPTER I

1.1 INTRODUCTION 

The chapter presents the research on the attitudes of ESL teachers in Malaysian secondary schools regarding the use of generative artificial intelligence (GenAI) tools in teaching and learning English and how these attitudes affect the application of these tools in the classroom by the teacher. The applications of GenAI (e.g., text-generation and interactive tutoring tools) have recently permeated education environments at a very high rate, generating both opportunities (e.g., better instructional support and efficiency in lesson design) and issues (e.g., reliability of outputs and aspects of academic dishonesty). In the Malaysian secondary school setting, where English is established as a significant language to support learning and subsequent employability, educators are once more central to the success or failure of GenAI in becoming significant to class practice.

In this respect, the research is informed by the following two objectives: first, to examine attitudes of ESL secondary school teachers towards GenAI tools in teaching and learning; and second, to investigate the way teachers incorporate GenAI tools into the teaching process. These objectives are denoted in the research questions that will study the perceptions of teachers and their classroom integration reporting. In order to situate the study, the chapter provides the theoretical background of how the attitudes and practices of the teachers are explored with reference to the Technology Acceptance Model (TAM) to explain the concept of perceived usefulness and ease of use together with a perspective of teacher cognition to explain how beliefs intermingle with contextual realities, to influence practice.

The chapter is structured in the following way. The background and contextual foundation of the study are presented in section 1.2, where the author discusses ESL teaching landscape and the applicability of GenAI to the modern classroom issues. Section 1.3 defines the problem statement and explains the gap in research that drove the study. Section 1.4 and 1.5 describe the research questions and objectives and Section 1.6 describes the theoretical and conceptual framework that directs the inquiry. The other sections cover the importance of the study, limitations, operational definitions of the key terms in the study, and a short conclusion that leads to the literature review in Chapter II.

1.2 BACKGROUND OF STUDY

In the contemporary educational landscape, the mastery of digital technologies has transcended being a mere advantage to becoming a fundamental necessity for global participation. This imperative is enshrined in the United Nations Sustainable Development Goal 4 (SDG 4), particularly Target 4.4, which mandates a substantial increase in the number of youth and adults who possess “relevant skills, including technical and vocational skills, for employment, decent jobs and entrepreneurship” by 2030 (United Nations 2016). Within the Malaysian context, this global call to action is deeply integrated into the nation’s core educational ethos, the National Education Philosophy (NEP). The NEP envisions the holistic development of individuals who are intellectually, spiritually, emotionally, and physically balanced (JERI) to contribute to national harmony and prosperity (Malaysian Ministry of Education 2023b). In the 21st century, the intellectual component of this philosophy is increasingly defined by digital fluency and the ability to leverage advanced technologies to solve complex problems.

To operationalize these philosophical and global aspirations, the Ministry of Education has implemented a robust strategic framework, the Malaysia Education Blueprint (MEB) 2026-2035, which explicitly identifies “Digital and Artificial Intelligence (AI)” as a critical new discipline that must be mastered (Malaysian Ministry of Education 2026). This strategic direction is further codified in the Digital Education Policy (DEP), specifically through Core Thrust 2, which demands the development of “Digitally Competent Educators” (Malaysian Ministry of Education 2023a). This policy mandates that teachers must move beyond basic digital literacy to become innovators capable of effectively integrating high-stakes technologies like Generative AI into teaching and learning.

To translate this policy into practice, the Ministry of Education has launched key enabling initiatives. The Digital Educational Learning Initiative Malaysia (DELIMa) platform provides a centralized hub for digital resources and, increasingly, AI-driven educational tools, offering the foundational infrastructure for integration. Complementing this, the AI for Good (Educator) Conference focuses on equipping teachers with the practical knowledge and skills needed to use AI responsibly and effectively in their classrooms (Malaysian Ministry of Digital, 2025a). These initiatives represent the institutional support structure designed to facilitate the transition towards digitally competent educators, as envisioned by the DEP. Thus, the adoption of AI tools in the classroom is not merely a pedagogical option but a direct compliance requirement aligned with Malaysia’s long-term strategic vision for a digitally empowered society.

Within this strategic national framework, English serves as a vital second language (ESL) indispensable for STEM education and global career mobility. Although the MEB 2026-2035 continues to prioritize English proficiency, achieving this standard remains a formidable challenge, particularly at the secondary school level. The impact of this high-pressure environment is reflected in recent student performance data. According to the Sijil Pelajaran Malaysia (SPM) results analysis reports (Malaysian Ministry of Education 2025), while the percentage of students achieving A+ (Cemerlang) increased from 16.0% in 2019 to 30.2% in 2024, 9.8% of students still failed the exam in 2024, and 29.4% only achieved passing grades (A-E). The National Grade Point Average (GPMP) improved (declining from 5.85 in 2020 to 4.75 in 2024), yet this overall progress contrasts sharply with the nearly 10% failure rate, highlighting a polarized achievement gap.

This gap is sustained by significant challenges faced by ESL teachers. Operating under the Kurikulum Standard Sekolah Menengah (KSSM) and the Common European Framework of Reference for Languages (CEFR), secondary school ESL teachers are tasked with a demanding curriculum that prioritizes high-level communicative competence over rote memorization (Nawawi et al. 2021). However, the high-stakes Sijil Pelajaran Malaysia (SPM) examination, where English is a compulsory pass subject, creates immense pressure to balance 21st-century pedagogical goals with exam preparation (Ag-Ahmad et al. 2025). This exam-oriented culture can narrow classroom priorities toward test-oriented tasks, reducing time for sustained interaction and formative language development. Furthermore, teachers grapple with excessive administrative workloads, which limit time for engaging teaching and timely feedback (Chandran et al. 2022). Challenges in teaching speaking skills, managing overcrowded and mixed-ability classrooms, and insufficient facilities further hinder the implementation of modern teaching methods and the provision of individualized support (Ag-Ahmad et al. 2025). These limitations collectively risk widening achievement gaps, especially when access to supportive resources varies between schools.

The incorporation of generative artificial intelligence (AI) tools in ESL teaching and learning presents a feasible avenue to address these multifaceted problems (Liao et al. 2023). According to Law (2024), generative AI is a type of deep learning model capable of producing human-like text, audio, images, and videos by analyzing massive datasets. Pedagogically, its use aligns with Social Constructivist theories of learning, particularly Vygotsky’s (1978) concept of scaffolding. Generative AI can act as a More Knowledgeable Other, providing individualized feedback and linguistic models that scaffold learning within a student’s Zone of Proximal Development (Cai et al. 2025). This theoretical perspective supports the communicative competence goals of the KSSM curriculum (Nawawi et al., 2021). Furthermore, GenAI revitalizes Krashen’s Input Hypothesis by dynamically adjusting linguistic complexity to provide comprehensible input (i+1) tailored to each learner’s level (Javahery & Alizadeh, 2025). This alignment with pedagogical theory is matched by policy alignment, as the MEB 2026-2035 identifies “Building Initial AI Capabilities” as a priority. Empirical studies support this potential, showing that generative AI tools can enhance students’ language skills and learning motivation (Vera 2023; Wei 2023; Qiuyang 2025; Dikaprio & Diem 2024).

For teachers, AI offers significant benefits, such as alleviating administrative load by automating tasks like grading and feedback, thereby freeing time for interactive teaching and addressing individual student needs (Sivanganam et al. 2025). It also provides a means to address classroom diversity by generating specific learning materials for varying proficiency levels. However, the successful adoption of these tools is often viewed through the lens of the Technology Acceptance Model (TAM), which posits that acceptance is driven by Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) (Davis 1989; Granić 2022). These perceptions are influenced by institutional support, training access, workload realities, and perceived pedagogical or ethical risks, explaining why similar technologies are adopted in some schools and not others (Granić 2022).

Consequently, the successful integration of AI into ESL classrooms hinges critically on teachers’ attitudes. Positive attitudes, fostered by perceived benefits and ease of use, enhance willingness to adopt and integrate these tools (Granić 2022). Conversely, negative attitudes or concerns about ethics, training, or technological displacement can lead to reluctance. Understanding these attitudes is therefore essential for identifying barriers and developing strategies to support effective AI integration. This study aims to fill this gap by exploring secondary school ESL teachers’ attitudes toward the use of generative AI tools in teaching and learning, as well as how they integrate these tools into their teaching practices. By examining teachers’ perceptions, this research will provide valuable insights into how AI tools can be effectively used to support English language learning, improve teaching practices, and enhance student outcomes.

1.3 PROBLEM STATEMENT

The primary impetus for this study arises from the persistent and inequitable challenge of English language proficiency among Malaysian secondary school students, a problem that persists despite significant national policy investments. This enduring issue is starkly reflected in the Sijil Pelajaran Malaysia (SPM) results, where data from 2020 to 2024 reveals a troubling divergence in outcomes: while the proportion of top achievers (A+) has risen to 30.2%, nearly 10% of students continue to fail the English paper annually, and the declining Grade Point Mean suggests stagnation in average performance (Malaysian Ministry of Education 2025). This bifurcation indicates that conventional pedagogical approaches have not successfully addressed the foundational proficiency gaps for a substantial segment of the student population. In this context, Generative Artificial Intelligence (GenAI) emerges as a theoretically potent intervention. Robust meta-analytic evidence confirms that AI-enhanced tools can significantly improve key language competencies, particularly writing accuracy and speaking fluency (Ekizer 2025; Sun & Zhou 2024). However, the translation of this technological potential into tangible classroom impact is not automatic; it is fundamentally mediated by the educators responsible for its implementation. Consequently, the core problem this study addresses is the critical disjunction between the strategic promise of AI and the human, contextual factors that determine its effective classroom adoption, a disjunction that can be understood through four interconnected research gaps.

A significant gap persists in understanding ESL teachers’ attitudes towards Generative AI (GenAI), which serve as the primary psychological gatekeepers to its adoption (Xuan & Yunus 2023; Sivanganam et al. 2025; Zulkarnain & Yunus 2023). Current research indicates these attitudes are often characterized by apprehension rather than acceptance, largely stemming from a fundamental lack of knowledge. For instance, a study conducted by Pokrivcakova (2023) found that out of 137 preservice English language teachers, a majority of them reported having no knowledge of how to implement AI pedagogically (61.31%) or of its limitations (63.50%). This cognitive deficit fuels significant affective resistance, with many teachers viewing AI as a threat to humankind (43.8%), fearing the erosion of students’ social skills (53.1%), or believing it could replace them entirely (34.4%). This anxiety has a demonstrable, mechanistic impact, as shown by Li and Thien (2025), whose study of 307 EFL pre-service teachers confirmed a direct mechanistic link between anxiety and rejection. Their structural equation modeling revealed that AI Learning Anxiety significantly and negatively impacts the Intention to Use (β = -0.235). Furthermore, they found that this anxiety degrades Perceived Usefulness (β = -0.230), suggesting that when teachers feel overwhelmed by the difficulty of mastering new tools, they become blind to the technology’s pedagogical benefits. It is therefore critical to investigate whether similar anxieties exist among Malaysian educators, as Pokrivcakova (2023) asserts that knowing these attitudes is a key factor in the success or failure of applying AI in education. Teachers harboring skepticism are more likely to implement restrictive policies, banning AI or limiting it to administrative tasks rather than leveraging it for pedagogical scaffolding. As Eusebio et al. (2025) argue, teachers serve as the primary mediators of technology use. Thus, if they act as resistant gatekeepers, they inadvertently deny students, particularly weaker ones access to the personalized and low-anxiety practice environments that GenAI can provide. Consequently, students remain dependent on the teacher’s limited time and attention, missing crucial opportunities for autonomous and self-regulated learning.

Second, beyond general attitudes, there is a distinct lack of knowledge regarding how ESL teachers practically integrate GenAI tools into their daily instructional cycles (Tripathi et al. 2025). Research by Celik (2023) addresses this gap by arguing that mere technological proficiency is insufficient; rather, understanding the specific pedagogical affordances of AI is essential to move beyond theoretical capability and grasp the technology’s real-world educational function. For instance, does AI serve merely as a time-saving resource generator, or is it leveraged as a cognitive partner for students during writing workshops? This gap is vital to address because the quality and depth of integration directly determine the educational value derived from the technology. Superficial use may yield efficiency gains but fail to foster the higher-order thinking and communicative competence that the KSSM/CEFR curriculum demands.

Third, existing research provides insufficient insight into how unique contextual realities mediate teachers’ decisions to adopt and integrate GenAI (Eusebio et al. 2025; Tripathi 2025). Teacher Cognition framework, as proposed by Borg (2003) posits that teacher practice is invariably filtered through contextual factors such as infrastructure and institutional culture. A recent systematic review by Eusebio et al. (2025) confirms that adoption barriers are not merely individual but systemic, highlighting that environmental barriers and technological infrastructure significantly constrain educator acceptance. In the Malaysian setting, these factors are particularly pronounced, characterized by a high-stakes examination culture and variable digital access. However, the specific mechanisms through which these systemic pressures interact with teachers’ beliefs to either enable or constrain AI use remain underexplored. This gap is crucial because it explains the often-observed “attitude-behavior gap,” where positive intentions fail to materialize in practice due to contextual roadblocks (Eusebio et al. 2025; Sivanganam et al. 2025).

Finally, these substantive gaps are compounded by a prevailing methodological gap in the research literature. The field of AI in education has been predominantly explored through quantitative surveys, which are ill-suited to unpacking the complex, subjective dimensions of teacher cognition. This imbalance is substantiated by Xue et al. (2025), who conducted a systematic literature review of AI adoption among teachers. Their analysis of 26 empirical studies revealed a stark methodological disparity: 65.38% (17 studies) relied exclusively on quantitative methods, while only 7.69% (2 studies) utilized qualitative approaches. As Tripathi et al. (2025) argue, the multifaceted phenomenon of AI integration demands qualitative inquiry to capture the rich, narrative data needed to understand teachers’ meaning-making processes. Therefore, this methodological gap is not merely an academic preference but a fundamental limitation that has left the “how” and “why” of teacher engagement with AI largely unanswered.

Therefore, to address the persistent inequity in English proficiency, this study is grounded in the critical need to bridge the disconnect between AI’s potential and classroom reality. It systematically investigates the intertwined gaps in understanding Malaysian ESL teachers’ attitudes, practical integration, and contextual mediators, employing the qualitative methodological lens required to capture these complex dimensions. The findings aim to generate empirically grounded insights for developing sustainable, context-sensitive professional strategies, thereby directly contributing to the national goal of equitable educational improvement.

1.4 RESEARCH OBJECTIVES

1. To explore the attitudes of ESL secondary school teachers towards the use of generative artificial intelligence tools in English language teaching.
2. To examine how ESL secondary school teachers integrate generative artificial intelligence tools into their teaching practices.
3. To explore the opinions of ESL secondary school teachers regarding the contextual feasibility of integrating Generative AI in their schools.

1.5 RESEARCH QUESTIONS

The research aims to address the following questions:

1. What are the attitudes of ESL secondary school teachers towards the use of generative artificial intelligence tools in teaching and learning?
2. How have ESL secondary school teachers integrated generative artificial intelligence tools in their teaching practices?
3. What are the opinions of ESL secondary school teachers regarding the contextual feasibility of integrating Generative AI in their schools?

1.6 THEORETICAL AND CONCEPTUAL FRAMEWORK

The present study is grounded in two key theoretical foundations, which are the Technology Acceptance Model (TAM) and the Teacher Cognition Framework. These frameworks complement each other in examining ESL teachers’ attitudes toward generative artificial intelligence (AI) tools and how such attitudes shape their teaching practices. While TAM explains the behavioral, motivational, and perceptual aspects influencing technology adoption, the Teacher Cognition Framework focuses on the internal mental processes that inform teachers’ pedagogical decisions and classroom actions. Together, these theories provide a holistic lens through which to understand how ESL teachers perceive, interpret, and integrate generative AI into their teaching contexts. The dual-theory methodology is consistent with the two aims of the study: TAM is better suited to answer the first (Research Question 1) question of exploring the attitudes of teachers, whereas Teacher Cognition offers an explanatory framework of how these attitudes can be translated into classroom integration practices (Research Question 2) and within the context of constraints that may or may not allow their enactment.

1.6.1 Theoretical Discussions

Technology Acceptance Model (TAM)

Figure 1.1 Technology Acceptance Model (TAM) (Davis 1989)

This study is grounded in the Technology Acceptance Model (TAM), originally proposed by Davis (1989), which serves as the primary theoretical lens for examining teacher attitudes. TAM posits that an individual’s intention to adopt a system is determined primarily by two cognitive beliefs: Perceived Usefulness (PU), defined as the degree to which a user believes the system will enhance their job performance, and Perceived Ease of Use (PEOU), the degree to which they believe using it will be free of effort. These constructs are the core predictors of behavioral intention; users are significantly more likely to adopt a technology if they view it as both beneficial to their productivity and easy to master.

The selection of TAM over other prominent frameworks, specifically the Technological Pedagogical Content Knowledge (TPACK) model (Mishra & Koehler 2006) and the Unified Theory of Acceptance and Use of Technology (UTAUT) (Venkatesh et al. 2003) is deliberate and grounded in the study’s focus on psychological acceptance rather than pedagogical competence or organizational compliance. While TPACK is a model of integrative knowledge, it measures the ability to use technology, not the willingness to do so (Huang et al. 2025). As Eusebio et al. (2025) note, psychological barriers often persist even among digitally competent teachers. Since the first research question specifically targets teacher attitudes, TAM is more appropriate as it isolates the psychological drivers (Usefulness and Ease) that precede and inform competence.

Furthermore, TAM is preferred over the more comprehensive UTAUT model due to its scientific parsimony and explicit focus on attitude. UTAUT is often criticized for its complexity and for frequently omitting the “Attitude” construct to focus solely on usage behavior (Dwivedi et al. 2019). In contrast, scholars argue that TAM’s parsimony is its strength, offering a valid and powerful approach that isolates the critical beliefs driving acceptance without diluting them with excessive variables (Granić 2023; Li & Thien 2025). Granić (2023) identifies TAM as the leading paradigm for investigating educational technology acceptance, explicitly defining it as an attitude toward technology. This aligns perfectly with the study’s objective to deconstruct the internal belief systems such as anxiety and ethical concerns that underpin teacher sentiment (Li & Thien 2025).

Therefore, TAM provides the most focused theoretical architecture to understand why ESL teachers hold specific attitudes toward Generative AI. While TAM excels at modeling this formative psychological stage, the study’s subsequent investigation into integration practices (RQ2) and contextual mediators (RQ3) will engage complementary perspectives, such as Teacher Cognition, to explain how these attitudes translate or fail to translate into classroom practice.

Teacher Cognition Model

Figure 1.2 Teacher Cognition Model (Borg 2003)

Complementing the attitudinal focus of Technology Acceptance Model (TAM), the study utilizes the Teacher Cognition Framework to examine the actual implementation of generative AI in the classroom (Borg 2003). Borg (2003) posits that a teacher’s internal mental life, including their beliefs, knowledge, and thoughts is inextricably linked to their observable classroom behaviors. Specifically, this research focuses on two key components of the framework: Classroom Practice and Contextual Factors. By analyzing integration across three distinct phases, which are “Pre-active” (Jackson 1968), “Interactive” (Jackson 1968), and “Post-active” (Clark & Peterson 1986) and considering the external constraints that shape these actions, this framework allows for a holistic examination of how ESL teachers translate their digital acceptance into pedagogical reality.

The decision to utilize the Teacher Cognition Framework rather than the Technological Pedagogical Content Knowledge (TPACK) framework or extending UTAUT into the implementation phase is driven by the specific need to analyze pedagogical reasoning rather than static competence or binary adoption. First, while TPACK is the dominant framework for assessing teacher knowledge, it is fundamentally a model of competence, not execution (Mishra & Koehler 2006). A recent critique by Eyal (2025) argues that TPACK frameworks often overlook distinctive AI characteristics, failing to capture the complex, iterative decision-making required to use AI effectively. TPACK can determine if a teacher possesses the skills to use GenAI, but it cannot explain why a skilled teacher might choose to exclude AI from a specific lesson or use it superficially due to external pressures. As noted by Tran et al. (2025a), successful GenAI integration is less about technical knowledge and more about an ‘AI Mindset’, a pedagogical vision where teachers must constantly evaluate the ethical and instructional value of AI outputs. The Teacher Cognition Framework is superior for this study because it explicitly accounts for the Cognition-Practice Gap, a phenomenon recently modeled by Zhou and Liu (2026) to illustrate how contextual constraints force teachers to deviate from their pedagogical beliefs, whereas TPACK implies that knowledge automatically leads to effective practice.

Second, while UTAUT is robust for predicting the intention to use AI (RQ1), it is insufficient for examining the nature of that use (RQ2). Bautista et al. (2024) emphasizes that GenAI implementation depends heavily on educators understanding and preparedness, which are cognitive states, not just behavioral frequencies. Bayaga and du Plessis (2024) argue that focusing on behavioral intention captures only one aspect of the ‘implementation continuum,’ effectively ignoring the actual pedagogical processes and subsequent usage behaviors required for effective integration. It lacks the theoretical architecture to distinguish between administrative use and pedagogical partnership, a distinction that Alfarwan (2025) identify as critical, noting that while administrative efficiency is well-documented, there remains a significant gap in understanding GenAI’s value in real-time teaching and learning. Since the integration of generative AI involves constant and high-stakes decision-making regarding accuracy and hallucinations (Tran et al. 2025b), it requires a framework that views teaching as a thinking process (Clark & Peterson 1986) rather than just a technical adoption event.

Therefore, to answer RQ2 and RQ3, which probe the how and why of integration within specific contexts, this study requires a framework that views teaching as a dynamic thinking process (Clark & Peterson, 1986). The Teacher Cognition Framework provides precisely this lens, moving beyond technical adoption to analyze the pedagogical reasoning that shapes GenAI use in the classroom.

Integration of Frameworks

The theoretical integration of the Technology Acceptance Model (TAM) and the Teacher Cognition Framework is necessitated by the complex reality that a teacher’s psychological acceptance of technology does not automatically translate into effective pedagogical practice. While TAM provides the definitive lens for deconstructing the attitudinal drivers of acceptance, including Perceived Usefulness and Perceived Ease of Use. It is not designed to explain the nature of classroom implementation or the contextual forces that shape it. To address this limitation, scholars like Granić (2023) argue that TAM’s explanatory power is enhanced when integrated with models accounting for post-adoption behavior and contextual constraints. This study responds to that imperative by synthesizing TAM with the Teacher Cognition Framework (Borg 2003), creating a dual-lens approach that maps the complete journey from internal belief to situated action.

This integration is both logical and sequential, directly aligning with the research questions. TAM serves as the foundational attitudinal lens for RQ1, isolating the cognitive beliefs (Usefulness, Ease of Use) and affective evaluations that form teachers’ initial disposition toward GenAI. However, to progress to RQ2 (integration practices) and RQ3 (contextual mediation), the theoretical focus must shift from acceptance to enactment. Here, the Teacher Cognition Framework provides the necessary analytical lens. It moves beyond the competence metrics of TPACK and the behavioral prediction of UTAUT to treat teaching as a dynamic thinking process. It explicitly accounts for the “Cognition-Practice Gap” (Zhou & Liu 2026), where contextual factors such as exam pressure, infrastructure, or ethical concerns filter and often constrain how a teacher’s attitudes and knowledge are translated into classroom practice. This synthesis ensures the study captures not only why teachers might intend to use GenAI but, critically, how and why they do so in the complex ecosystem of their professional reality.

This combined framework does not merely serve as a conceptual backdrop; it directly operationalizes the research design and analysis. The inquiry is structured to mirror the theoretical progression: instrument design will first elicit TAM-informed appraisals of utility and anxiety before probing the situated decision-making captured by Teacher Cognition. Analysis will specifically examine the phases of “Pre-active,” “Interactive,” and “Post-active” practice to understand integration behaviors, while simultaneously identifying the “Contextual Factors” that Borg (2003) establishes as the primary filters for instructional choices. Consequently, this integrated approach provides the a priori categories to systematically investigate the potential disjunction between positive attitude and limited use, constructing an explanatory model of how acceptance is ultimately realized, adapted, or abandoned in the Malaysian ESL classroom.

1.6.3 Conceptual Framework

Figure 1.3 Conceptual Framework

The conceptual framework for this study synthesizes the Technology Acceptance Model (TAM) and the Teacher Cognition Framework into a sequential trajectory that traces the adoption of Generative AI from an initial technological stimulus to its actualized pedagogical implementation. Unlike single-theory models that may isolate user satisfaction from professional behavior, this integrated framework posits that AI integration is a multi-stage process governed by both psychological drivers and environmental filters. As illustrated in the framework, the process is initiated by the introduction of Generative AI tools (e.g., ChatGPT) as the external input. The immediate response to this input is cognitive: consistent with the established parameters of TAM, the framework identifies Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) as the independent variables that determine the teacher’s attitude. Recent meta-analyses in AI education, such as Ekizer (2025), confirm that this pathway, from belief to attitude remains the strongest predictor of a teacher’s intent to use technology. Consequently, the framework positions “Teacher Attitudes” as the mediating variable, representing the necessary psychological readiness that must precede any attempt at classroom application.

However, a critical rationale for this specific framework design is the recognition that a positive attitude does not automatically guarantee effective classroom usage, particularly in high-stakes environments. To account for this “attitude-behavior gap”, a phenomenon recently documented among Malaysian ESL teachers by Sivanganam et al. (2025), the framework incorporates “Contextual Factors” from the Teacher Cognition Framework as a critical moderating variable. This component explains why teachers with high perceived usefulness might still resist implementation; their professional agency is constrained or enabled by systemic realities such as infrastructure, SPM examination pressures, and digital policy mandates. This flow is supported by Zaimoğlu and Dağtaş (2025), who argue that in the era of AI, a teacher’s cognitive decision-making is inextricably “situated” within their institutional context. Therefore, the arrow in the framework moving from “Attitude” to “Classroom Practice” is not direct but is filtered through these contextual realities, illustrating that integration is a negotiated act rather than a simple technical adoption.

Crucially, this conceptual structure directly operationalizes the research methodology and instrument development. The sequential logic of the framework dictates the architecture of the semi-structured interview protocol, which is segmented to assess each stage of the flow: first probing the independent variables of utility and ease (TAM), then shifting to the moderating influence of school constraints (Context) and finally examining the dependent variable of “Classroom Practice.” This final output is not treated as a monolithic concept but is analytically divided into the pre-active, interactive, and post-active phases of teaching. This specific categorization allows the research to isolate exactly where AI is being used or abandoned within the teaching cycle. By structuring the inquiry in this manner, the framework provides the analytical scaffold for deductive coding, enabling the study to move beyond descriptive findings to construct a causal explanation of how environmental factors intervene to shape, modify, or halt the integration of Generative AI in Malaysian secondary schools.

1.7 SIGNIFICANCE OF STUDY

This study holds significant practical and strategic value, contributing directly to pivotal educational objectives at the global, national, and institutional levels. Its findings are designed to inform policy refinement, teacher development, and classroom practice, ensuring the integration of Generative Artificial Intelligence (GenAI) aligns with broader goals of equitable quality education and national advancement.

First, the research provides direct support for the United Nations Sustainable Development Goal 4 (SDG 4) on inclusive and equitable quality education (United Nations 2016). It addresses Target 4.c by investigating the specific cognitive and infrastructural barriers that must be overcome to increase the supply of qualified teachers capable of leveraging modern tools. Concurrently, it supports Target 4.4 by positioning AI literacy as a crucial skill for future employability. By pinpointing the factors that hinder teacher acceptance and effective use, this study generates the necessary evidence to design targeted professional development. This ensures GenAI acts as a tool for empowerment that enhances learning for all students, rather than a disruptive force that exacerbates existing inequalities.

Second, the inquiry is critically aligned with Malaysia’s National Education Philosophy (NEP), which mandates the holistic development of individuals intellectually, spiritually, emotionally, and physically (Malaysian Ministry of Education 2023b). The study scrutinizes AI integration through this philosophical lens, investigating whether teachers perceive it as a mechanism that supports holistic growth, such as by fostering critical thinking and creativity, or as a potential threat to ethical and emotional development (Chan 2023). By centering the teacher’s mediating role, the findings will help ensure that technological implementation upholds the NEP’s core mission to produce knowledgeable, competent citizens with high moral standards who contribute to societal harmony.

Third, the timing of this research is pivotal for the strategic formulation of the Malaysian Education Blueprint 2026–2035, which aims to cultivate individuals of holistic well-being and Global Citizens (Malaysian Ministry of Education 2026). As the Ministry of Education pivots toward this vision, this study serves as a crucial baseline assessment of teacher readiness. By documenting current pedagogical alignments and gaps in ESL classrooms, the evidence will empower policymakers to allocate resources effectively and craft supportive strategies. This ensures the aspiration for globally competitive citizens is built upon a foundation of capable and confident educators.

Finally, the study offers operational significance for the success of national digital initiatives, namely the Digital Education Policy (DEP) (Malaysian Ministry of Education 2023a) and the National AI Action Plan 2030 (AI for Nation) (Malaysian Ministry of Digital 2025b). While the DEP aims to produce digitally fluent educators and the AI Action Plan envisions pervasive technological adoption, both depend entirely on successful classroom execution. This research acts as a critical reality check, identifying the gaps between high-level policy intent and on-the-ground practical usage. This systemic disconnect is highlighted in studies of educational innovation (Xiaofan & Annamalai 2025). The resulting insights will be instrumental in refining professional development, shifting it from theoretical exposure to practical application, thereby securing Malaysia’s trajectory toward becoming a digitally competitive and future-ready nation.

1.8 LIMITATIONS OF THE STUDY

This study provides meaningful insights into Malaysian ESL secondary school teachers’ attitudes toward and integration of Generative AI tools, yet several methodological and conceptual limitations must be acknowledged.

First, the study is constrained by its qualitative design and relatively small participant sample. As qualitative inquiry prioritizes depth over breadth, the findings cannot be statistically generalized to all ESL teachers in Malaysia. However, as Staller (2021) notes, the purpose of such research is to generate rich and context-specific understanding rather than broad representation. The insights offered are therefore valuable for constructing a detailed and narrative based perspective on teacher experiences within the studied contexts, serving as an important exploratory foundation for further large-scale research.

Second, this research captures teachers’ perceptions and reported practices at a single point in time. Given the rapid evolution of AI technologies, attitudes and integration strategies are likely to change (Seo et al. 2021). While this limits the longevity of the findings, establishing a clear baseline at this pivotal moment is critical. It provides a necessary reference point for tracking how teacher acceptance and pedagogical approaches evolve alongside the technology, thereby informing the timing and focus of future longitudinal studies and policy updates.

Third, the study focuses solely on teachers’ perspectives and does not directly measure student learning outcomes or classroom efficacy resulting from AI integration. As Nguyen et al. (2025) demonstrate, students often engage with AI tools in ways that differ from teachers’ assumptions, suggesting that teachers’ perceptions alone cannot fully represent students’ actual learning behaviours. Although this limits claims about the tangible educational impact of GenAI, understanding the teacher’s role remains fundamentally important. Teachers act as the primary gatekeepers and mediators of technology in the classroom (Sivanganam et al. 2025). Their perceptions, competencies, and decisions directly shape whether and how students gain access to AI tools. Therefore, investigating this mediating perspective is a vital first step in mapping the ecosystem of AI integration, preceding and informing subsequent research into student outcomes.

1.9 OPERATIONAL DEFINITION OF TERMS

The terms used in this study are defined based on how they are applied and measured within the context of the research. Where relevant, these operational definitions are aligned to the study’s theoretical framing (Technology Acceptance Model and Teacher Cognition) and specify how each construct is interpreted within Malaysian secondary ESL classrooms (Davis 1989; Borg 2003).

1.9.1 English as a Second Language (ESL)

English as a Second Language (ESL) refers to the teaching and learning of English by non-native speakers. In this context, ESL encompasses the development of language skills in reading, writing, speaking, and listening, aiming to enable learners to communicate effectively in English (Zhang & Kang 2022). It is typically taught to individuals whose primary language is not English, with an emphasis on helping them achieve proficiency in English for academic, social, and professional purposes. In Malaysia, ESL education focuses on preparing students to meet the demands of English for academic success, particularly in subjects like science, technology, and mathematics, and for future career opportunities in a globalized workforce. Within Malaysian secondary schools, ESL learning is implemented through the KSSM English curriculum aligned to CEFR proficiency targets (Malaysian Ministry of Education, 2019).

1.9.2 English as a Second Language (ESL) teachers

ESL teachers are educators who specialize in teaching English to students whose first language is not English (Babinski et al. 2024). These teachers focus on developing students’ language skills in reading, writing, speaking, and listening and helping them achieve proficiency in English for academic, social, and professional purposes. ESL teachers may use various instructional strategies and methods tailored to the students’ language proficiency levels, cultural backgrounds, and learning needs. In Malaysia, ESL teachers often work with non-native English-speaking students in formal educational settings. In this study, ESL teachers refer specifically to Malaysian public secondary school English teachers who plan, deliver, and assess English lessons under KSSM/CEFR requirements and within the SPM assessment context, and who therefore act as key decision-makers in whether and how generative AI is used in instruction.

1.9.3 Generative Artificial Intelligence (AI)

According to Moorhouse and Kohnke (2024), generative AI tools refer to advanced machine learning technologies designed to create human-like content through the analysis of large datasets. These tools generate content such as text, audio, images, and videos. In the context of ESL teaching, these tools include AI-driven language models like GPT, automated essay graders, virtual tutors, and AI-based educational applications used to support content creation, feedback delivery, and personalized learning experiences for ESL students. To make the point clear, in this research, the concepts of generative AI tools would encompass both publicly available conversational agents (e.g., ChatGPT-like systems) and educational AI applications that produce language output, feedback, or lesson materials, whether delivered through a web-based platform or integrated learning system (Kasneci et al. 2023).

1.9.4 English as a Second Language (ESL) teachers’ attitudes

In this study, attitudes refer to ESL teachers’ psychological dispositions, beliefs, and perceptions toward the use of generative AI tools in the classroom. These attitudes encompass teachers’ evaluations of the usefulness, ease of use, and appropriateness of AI tools in enhancing their teaching practices and students’ learning outcomes. The attitudes may be positive or negative and are influenced by individual beliefs, past experiences, and contextual factors. In line with TAM, the attitude of teachers can be mainly demonstrated by the perceptions of usefulness and ease of use, as well as the judgements of the pedagogical and ethical suitability of the learning to the learners (Davis 1989; Granić 2022).

1.9.5 Differentiated Instruction

According to Ken et al. (2025), Differentiated Instruction is a pedagogical approach that involves proactively tailoring instruction to meet students’ diverse needs, readiness, and interests by modifying content, process, and product. In the context of this study, this concept refers to the specific application of Generative AI tools by ESL teachers to adapt lessons for mixed-ability secondary classrooms within the KSSM framework. This encompasses using AI to generate multi-level reading texts (content), provide personalized scaffolding and feedback (process), and design varied assessment tasks (product) that allow students of different CEFR proficiency levels to achieve the same learning standards.

1.9.6 Pre-active Teaching

According to Jackson (1968), the pre-active phase refers to the deliberate planning stage of teaching that occurs before the teacher enters the classroom. It involves cognitive processes such as selecting learning objectives, organizing content, and designing instructional strategies to meet curricular goals. In the context of this study, this phase is operationally defined as the teacher’s use of Generative AI tools to support instructional design and material preparation. This includes utilizing AI to brainstorm lesson ideas, generate KSSM-aligned reading texts, create lesson plans, and design differentiated worksheets or presentation slides prior to the actual lesson delivery.

1.9.7 Interactive Teaching

The interactive phase involves the spontaneous decisions and interactions that occur during the act of teaching (Jackson 1968). It is characterized by real-time decision-making where teachers must respond to student reactions, manage classroom dynamics, and deliver content. In the context of this study, this phase refers to the integration of Generative AI tools during live classroom instruction. This encompasses the teacher’s use of AI as a real-time pedagogical aid, such as using an AI interface on a smartboard to demonstrate language examples, verifying vocabulary or grammar in response to student queries, or facilitating AI-driven activities (e.g., students interacting with a chatbot) within the lesson timeframe.

1.9.8 Post-active Teaching

The post active phase was conceptualized by Clark and Peterson (1986), expanding upon Jackson’s (1968) earlier distinction between pre active and interactive teaching. It refers to the critical period of reflection and evaluation that occurs after the lesson has concluded, where teachers analyze their instruction and student performance to inform future planning. In the context of this study, this phase is operationally defined as the utilization of Generative AI tools to streamline assessment and reflective practices. Specifically, it refers to using AI to grade student essays, generate automated feedback rubrics, analyze student errors, or reflect on lesson outcomes to refine future teaching strategies.

1.9.9 Perceived Usefulness (PU)

Perceived Usefulness (PU) is a notion describing the degree to which ESL teachers think that the application of generative AI tools will positively affect their teaching performance and/or the English language learning outcomes of students (Davis 1989). In this research, PU incorporates perceived value of the tasks which include, lesson planning, language modelling, differentiated support, and timely feedback.

1.9.10 Perceived Ease of Use (PEOU)

Perceived Ease of Use (PEOU) is defined as how the educators of the English as a Second Language suppose that the generative AI tools are simple to study, operate and get involved in the regular teaching process with the little extra effort (Davis 1989). PESU in this study will comprise perceived ease of accessing the tool, writing prompts, reading outputs and classroom management.

1.9.11 Practice in the Classrooms (Integration of Generative AI)

The way ESL teachers document the application of generative AI tools in their professional practice throughout the teaching cycle. The classroom practice in this research is reviewed on the planning, instruction, and assessment-related practices showing how the teachers translate their beliefs and intentions into the practice pedagogical processes (Borg 2003; Clark and Peterson 1986).

1.9.12 Contextual Factors

Contextual factors are conditions at school and system levels that influence the capacity of teachers to embrace and implement generative AI tools. These aspects in this research are infrastructures and access (equipment, internet), pressure of curriculum and assessment, time and workload, institutional direction/policy, and professional learning support (Malaysian Ministry of Education 2023a).

1.10 SUMMARY

This chapter highlights the challenges faced by ESL education in Malaysia, where students’ English proficiency continues to fall short despite various efforts. It presents the integration of generative AI tools as a potential solution to improve both teaching practices and student outcomes. The study investigates ESL secondary school teachers’ attitudes towards AI tools and explores how these tools are incorporated into their classroom practices. Using the Technology Acceptance Model (TAM) and Teacher Cognition Framework, this research aims to understand the factors influencing AI adoption in ESL teaching. The study addresses a gap in existing literature, particularly in the Malaysian context, where limited research has focused on teachers’ perceptions and integration of AI. The findings will contribute valuable insights to enhance AI integration in ESL classrooms, aligning with national policies like the Digital Education Policy (DEP) to foster effective teaching and learning.

CHAPTER II

2.1 INTRODUCTION

Chapter II gives the literature-based context behind the emergence of the concept of generative artificial intelligence (GenAI) as a significant concern in Malaysian secondary ESL education, and the responses of teachers to it are critical to its successful uptake in any classroom setting. The chapter, which is a continuation of Chapter 1, in which the research has identified the ongoing proficiency disparity and the growing policy impetus toward digital transformation, places teachers at the interface point that could be used to mediate the technological possibility and classroom actuality. That is, GenAI can be available to a large number of teachers, yet its pedagogical usefulness depends on whether (i) the teachers feel that it is helpful and practical and (ii) can apply it in pedagogically acceptable manners within actual limitations such as examination pressure, teacher workload, school infrastructure, and school administrative guidance.

This chapter is a critical review of the literature on Generative Artificial Intelligence (GenAI) in ESL education, with a comprehensive overview of the Malaysian digital policy environment. It explains the role of teacher attitudes in determining the level of integration practice by teachers at pre-active, interactive and post-active levels by synthesizing the Technology Acceptance Model (TAM) and the Teacher Cognition Framework. The discussion points out major advantages, issues and situational barriers found in recent empirical research. Finally, the review identifies research gaps in the critical research on the Malaysian secondary school teachers, creating the theoretical and practical justification of the study.

In order to enhance the strength of the coherence and guiding toward a structured critical review (as opposed to descriptive listing of studies) the chapter is structured in such a way that it progresses to the context and then to theory then to evidence and finally to gaps. First, it surveys the Malaysian English language teaching situation (curriculum and policy) since technology choices of teachers are anchored in system demands (e.g., CEFR-based objectives) and assessment facts (e.g., SPM washback). Second, it talks about the theoretical background theoretical lenses upon which this study has built, Technology Acceptance Model (TAM) and Teacher Cognition, since they explain different yet complementary aspects of adoption: TAM elucidates why teachers are willing (or unwilling) to use GenAI, whereas Teacher Cognition elucidates how professional knowledge, beliefs, and situated judgement of teachers are relevant to what occurs in practice.

This chapter further aligns the discussion of GenAI-supported learning with key language acquisition theories, namely the social constructivist concept of scaffolding within the Zone of Proximal Development and Krashen’s Input Hypothesis. GenAI tools are often framed as supportive partners capable of modeling language and facilitating learner output, thereby providing the “comprehensible input” (i+1) and interactive feedback these theories emphasize. However, the literature cautions that such scaffolding is not inherently beneficial. Its efficacy is contingent upon deliberate teacher mediation, careful task design, and guided learner practices for critically evaluating and refining AI-generated output. Consequently, this chapter does not posit GenAI as a definitive solution to ESL challenges but rather examines it as a pedagogical tool whose effectiveness is ultimately mediated by teachers’ acceptance, their instructional reasoning, and the enabling conditions of their specific environments.

Lastly, the chapter wraps up empirical research (with some focus on recent research) to determine what is known regarding GenAI in the language education setting where evidence is both varied and insufficient, and what is not yet explored in the Malaysian secondary school settings, specifically, the attitudes of teachers and the extent to which it is integrated in planning, instruction, and assessment. The synthesis is what provides the path to the justification of the current study and the methodological decisions that are discussed in Chapter III.

2.2 ENGLISH LANGUAGE EDUCATION LANDSCAPE IN MALAYSIA

The use of English language teaching (ELT) as a means of teaching, globally, has resulted in a growing trend towards a communicative competence approach to teaching English, accompanied by more emphasis on the use of real language use, a pedagogical approach that is learner-centred, and the use of digital tools to facilitate interaction and feedback. These changes have been enhanced in Asia by policy pressures to graduate bilingual and digitally capable graduates to knowledge-based economies and has increased the rate of curriculum reform and the use of technology in education. In Malaysia, this global-regional trend can be seen in national reconfigurations in which English has been returned to the strategic competency, and more teachers have been positioned as the implementers of curriculum and digital transformation projects. This section examines the key national policies and curricular frameworks that shape the landscape of English language education in Malaysia. An overview of these frameworks provides insight into the structural demands placed on secondary school teachers, shedding light on the national shift towards communicative competence and digital transformation. Collectively, these policies provide the essential context for understanding the environment in which teachers are expected to integrate generative AI tools into their classroom practices.

2.2.1 National Education Philosophy (NEP)

At the core of the Malaysian education system lies the National Education Philosophy, which serves as the definitive guide for all educational activities. Formulated following the Cabinet Committee Report in 1979 and officially announced in 1988, the FPK defines education as a “continuous effort” to develop the potential of individuals in a holistic and integrated manner, producing citizens who are intellectually, spiritually, emotionally, and physically balanced (JERI) based on a firm belief in God (Malaysian Ministry of Education 2023b). In the era of the Industrial Revolution 4.0 (IR 4.0), this philosophy has become strategically critical; it serves as the essential mechanism to infuse academic supremacy with strong human values, ensuring that rapid technological advancement does not erode the nation’s cultural and ethical civilization (Hassan & Aziz 2024).

The successful translation of this philosophy into reality rests heavily on the professionalism of teachers, who are the primary agents responsible for internalizing and applying FPK principles. However, this holistic vision often conflicts with systemic realities, particularly the deep-seated “exam-oriented culture.” Empirical research indicates that the national assessment system focuses disproportionately on the “intellectual” element, often causing the emotional, spiritual, and physical aspects of student development to be neglected (Al-Hudawi et al. 2014). This creates a disconnect where students may achieve academic standards but lack the “holistic potential” envisioned by the NEP, as teachers are pressured to prioritize measurable outcomes over character building.

In the context of this study, the FPK serves as the essential ethical framework against which ESL teachers must evaluate the integration of Generative AI. While the philosophy encourages the development of “intellectual” elements such as critical and creative thinking, it simultaneously warns that physical technology is weak without the support of human values (Hassan & Aziz 2024). Therefore, the adoption of AI tools is not just a technical decision but a philosophical negotiation; teachers must determine if these tools support the FPK’s aspiration for a balanced human being or if they exacerbate the existing tendency to prioritize efficiency and exam results over genuine, holistic learning.

2.2.2 The Malaysia Education Blueprint (MEB) 2026-2035

As the successor to the Malaysia Education Blueprint 2013–2025, the Malaysia Education Blueprint (MEB) 2026–2035 serves as the new strategic intervention to ensure the national education system remains relevant in a post-pandemic global landscape (Malaysian Ministry of Education 2026). With the theme “Merapatkan Jurang, Meningkatkan Mutu, Meraih Kejayaan Bersama” (Closing the Gap, Raising the Bar, Together We Succeed), the MEB explicitly identifies technological challenges, specifically Artificial Intelligence (AI), as a critical focus area that requires urgent strategic alignment. Unlike previous policies that focused broadly on ICT, the RPM operationalizes this through Teras Strategik 4 (Strengthening Physical and Digital Infrastructure), which mandates the transformation of learning experiences through digital innovation to support the “Future-Ready School” framework.

Within this new framework, the integration of advanced technologies is no longer optional but a central requirement for developing “Insan Sejahtera” (holistic human capital). The MEB sets a rigorous Key Performance Indicator (KPI) requiring all students to achieve a minimum of an “Intermediate” level on the Digital Competency Score, moving beyond basic literacy to active digital fluency. Furthermore, the policy emphasizes Teras Strategik 1 (Ensuring an Inclusive, Dynamic, and Relevant Education System), which aims to produce students who are not only knowledgeable but capable of facing complex future challenges, including those posed by the rapid permeation of AI in the workforce (Malaysian Ministry of Education 2026).

However, the transition to this envisioned model is complicated by significant implementation gaps inherited from previous policies, a reality now explicitly validated by the RPM 2026-2035’s “deep diagnosis” of the education system. This diagnosis reveals interconnected pedagogical and infrastructural barriers that align closely with recent academic findings. At the pedagogical level, research indicates that the shift toward student-centered facilitation (PAK-21) is often misinterpreted by teachers who equate 21st-century learning with the mere presence of IT gadgets rather than fundamental pedagogical transformation (Muhamad & Seng 2022). The MEB 2026-2035 substantiates this observation, noting that while teacher performance evaluations often show high achievement, actual student outcomes remain below targets. This discrepancy has necessitated Teras Strategik 3 (Transforming Educators), which shifts the focus from basic proficiency to ensuring educators are “future-oriented” and capable of translating digital tools into meaningful learning outcomes.

This pedagogical challenge is compounded by persistent infrastructural inequalities. Critics have long argued that the promise of equitable digital access remains unfulfilled, with rural schools suffering from inadequate internet and device shortages that hinder the adoption of digital tools (Ismail 2025; Ahmad & Rathakrishnan 2025). The MEB 2026-2035 officially concedes this point, stating that “development imbalance between urban and rural areas remains glaring”. The policy document explicitly reports that access to internet and devices is still unequal in interior areas, confirming that the national digital transformation has been uneven.

Consequently, while the MEB 2026-2035 provides a robust policy framework for AI integration, it operates within a context of recovery. The policy’s introduction of Teras Strategik 4 (Strengthening Physical and Digital Infrastructure) is a direct response to these facilitating conditions that were previously absent. For ESL teachers, this establishes a critical context: their attempt to integrate Generative AI is not merely a matter of personal willingness, but of navigating a systemic gap between the strategic ambition of the new policy and the lingering reality of classroom infrastructure.

2.2.3 Secondary School Standard Based Curriculum (KSSM) AND Common European Framework of Reference for Languages (CEFR) Alignment

The implementation of the Secondary School Standard-Based Curriculum (KSSM) represents a strategic effort by the Ministry of Education to align Malaysian education with global standards, emphasizing the development of 21st-century skills such as critical thinking, communication, and collaboration (Malaysian Ministry of Education 2016). Being a national standards-based curriculum, KSSM gives the general framework that outlines learning standards and expectation of performance throughout the whole of secondary schooling, and thus how English is planned, taught, and assessed at the classroom level.

In the field of English language education in particular, these curriculum expectations are operationalised in the CEFR-aligned English curriculum. Central to this reform is the integration of the Common European Framework of Reference for Languages (CEFR), which provides a comprehensive descriptor scheme to measure language proficiency (Malaysian Ministry of Education 2020). Unlike previous curricula that often prioritized rote learning, the CEFR-aligned curriculum focuses on developing students’ communicative competence, requiring them to use language functionally in real-world contexts rather than merely knowing grammar rules. This shift necessitates a transformation in pedagogical approaches, urging teachers to move away from teacher-centered instruction toward more student-centered, interaction-based learning environments.

Yet even though the objectives of KSSM and CEFR are complementary, this idealistic model of the curriculum exists in the context of the powerful contextual reality of high-stakes tests and this creates a fundamental conflict to ESL teachers. The conflict is in the form of a pedagogical dilemma in which the pressure of the assessment and the goals of the curriculum conflict. Research indicates that the washback effect of high-stakes exams often mediates and can distort curriculum implementation. Contextual factors, particularly systemic pressure to produce grades, create a disconnect between the intended learning-oriented curriculum and actual exam-driven classroom practices, trapping teachers between formative assessment mandates and summative examination demands (Khan et al. 2025). Consequently, teachers face the practical challenge of balancing pedagogical innovation with exam readiness. Studies note that while educators recognize the value of interactive, technology-enhanced methods, they often deprioritize strategies like differentiation and formative assessment to focus narrowly on meeting exam-linked standards (Radi & Zabit 2025; Sadhasivam et al. 2023). This results in a critical gap, the KSSM/CEFR framework provides an ideal structure for holistic language development, but its practical integration is frequently subordinated to examination preparation.

This core dilemma directly informs the present study. It raises a pivotal question for understanding teacher attitudes towards Generative AI, “Do teachers perceive AI as a tool capable of bridging this gap by simultaneously enabling communicative, CEFR-aligned practice and building exam-relevant skills efficiently, or do they view it as another complicating factor within an already strained system?”. Examining this question is essential, as the perceived utility or threat of AI will be filtered through this pre-existing tension between pedagogical aspiration and systemic constraint.

2.2.3 Digital Education Policy (DEP)

To accelerate the digital transformation of the national education ecosystem, the Ministry of Education launched the Digital Education Policy (DEP) in 2023. While previous initiatives primarily focused on providing hardware, the DEP represents a strategic shift towards developing human capital, with the overarching goal of producing a digitally fluent generation capable of using technology in an integrated, creative, and ethical manner (Malaysian Ministry of Education 2023a).

Of relevance to this study is Core Thrust 2: Digitally Competent Educators. This thrust explicitly mandates that teachers must transcend basic digital literacy, such as using computers for administrative work to become digital innovators who can effectively integrate advanced technologies into teaching and learning (Malaysian Ministry of Education 2023a). To operationalize this, the Ministry has enhanced the DELIMa (Digital Educational Learning Initiative Malaysia) platform, which now serves as a central hub for AI-enabled tools and resources, signaling a clear institutional directive for teachers to adopt intelligent technologies in the classroom.

This policy landscape serves as a critical Contextual Factor within the framework of this study. As supported by the findings of Ahmad and Rathakrishnan (2025), national policies like the DEP create a supportive institutional environment that encourages teachers to integrate technology, thereby positively influencing their acceptance and usage. By formally embedding digital competence into professional standards, the government aims to reduce the technical barriers that previously hindered adoption, thereby encouraging ESL teachers to view generative AI not as a disruptor, but as an essential tool for meeting 21st-century educational demands.

2.3 UNDERPINNING THEORIES

This section examines the underpinning theories and models that inform and drive the study’s exploration of generative AI in ESL education. To provide a holistic understanding of the phenomenon, the research integrates two complementary frameworks, which are the Technology Acceptance Model (TAM) and the Teacher Cognition Framework. As well, the section also appeals to sociocultural theory (Vygotsky 1978) to describe the pedagogical logic of generative AI as a mediational tool and scaffold to learn a language, thereby providing coherence to the rationale of the learning theory presented in Chapter One. Together, these theories also shape the methodological design of the study, in the sense of the interview prompts and the first coding categories (attitudes, classroom practices, and contextual mediators) applied in the thematic analysis.

2.3.1 Technology Acceptance Model (TAM)

The Technology Acceptance Model (TAM), originally introduced by Davis (1989), provides the foundational framework for understanding the cognitive drivers behind technology adoption. For this study, TAM is essential for diagnosing ESL teachers’ willingness to integrate Generative AI, as it posits that adoption intent is primarily determined by two core beliefs: Perceived Usefulness (PU) and Perceived Ease of Use (PEOU).

TAM was first introduced within the information systems studies, but its main concepts have been broadly applied to study the acceptance of new tools by teachers and understand factors that drive this process based on the belief system (Granić 2022; Xue et al. 2025). In the research, TAM operates as a sensitizing and analytical prism of the qualitative inquiry instead of a statistical predictive model; PU and PEOU are employed to organize interview prompts and interpret the narratives of teachers regarding the usefulness and importance of generative AI in ESL work.

In the context of ESL teaching, Perceived Usefulness refers to a teacher’s belief that using Generative AI will tangibly enhance their pedagogical efficacy and job performance. For instance, by streamlining administrative tasks, generating personalized materials, or automating feedback (Nurjanah et al. 2025). Perceived Ease of Use, conversely, captures the anticipated effort required to master the technology, acting as a critical filter for its adoption. Empirical research supports the salience of these constructs in educational settings. A study by Mutammimah et al. (2024), which applied TAM to English language teaching, confirmed that both PU and PEOU are significant predictors of teachers’ intention to adopt tools like ChatGPT. Their findings underscore a key dynamic: teachers are more likely to integrate AI if they view it as a genuine enhancer of productivity and instructional quality (high PU), but only if they also find the technology intuitively manageable (high PEOU). The study further notes that PEOU can mediate perceptions of usefulness; a tool perceived as difficult may have its potential benefits disregarded regardless of its objective utility.

Therefore, applying TAM allows this study to systematically investigate the attitudinal foundations of AI integration among Malaysian ESL teachers. It moves beyond a simple inventory of tools used, to instead examine how teachers’ perceptions of pedagogical value and technical accessibility shape their openness to adopting Generative AI, thereby offering critical insights into the human factors that will determine the success of national digital transformation policies.

2.3.2 Teacher Cognition Model

To complement the attitudinal focus of the Technology Acceptance Model (TAM), this study employs the Teacher Cognition Framework (Borg 2003). While TAM helps predict adoption intention based on utility and ease, Teacher Cognition provides the essential structure for examining how teachers’ internal mental lives. Their beliefs, knowledge, and thoughts are translated into, or constrained by, actual classroom practice. This framework positions teachers as active decision-makers whose pedagogical choices are shaped by a personalized network of cognitions, acknowledging that the link between belief and action is often non-linear. For instance, a teacher may believe in AI’s potential for personalization yet ban its use over cheating concerns, with such contradictions frequently resolved in the real-time “interactive phase” of teaching where contextual realities force adaptation.

Crucially, the translation of cognition into practice is powerfully mediated by Contextual Factors. In the Malaysian educational landscape, these factors create a defining tension for innovation. Empirical research identifies macro-level pressures, such as the entrenched “exam-oriented culture,” and micro-level constraints, including excessive workload and insufficient time, which collectively compel teachers to prioritize exam preparation over innovative methods and severely limit their capacity to experiment with new tools like AI (Khan et al. 2025). This makes context a critical moderating variable, a teacher’s positive attitude as measured by TAM may not lead to integration if contextual barriers are prohibitive.

Furthermore, the advent of Generative AI necessitates a re-examination of teacher cognition for the digital age. Scholars argue that AI transforms the teacher’s role into that of a manager of “distributed cognition,” requiring pedagogical expertise to filter, adapt, and integrate AI-generated content meaningfully (Yıldız & Erçetin 2025). This underscores that successful integration depends not on the technology alone but on the teacher’s “human-in-the-loop” expertise. However, a significant gap persists, as noted by Wang and Zhang (2024), in empirical studies that explicitly link language teacher cognition to Generative AI adoption.

Therefore, by integrating the Teacher Cognition Framework, this study moves beyond cataloging if teachers use AI to investigate how their professional reasoning enables them to navigate the integration of AI within the specific, and often restrictive, context of Malaysian secondary ESL classrooms. It seeks to understand how teachers maintain pedagogical integrity and exercise agency amidst the conceptual and practical challenges posed by rapid technological change.

2.3.3 Sociocultural Theory of Learning (Vygotsky)

In addition to the belief in adoption and professional justification, the sociocultural theory offers the rationale of learning why generative AI can be significant in language classrooms. Vygotsky (1978) formulates the understanding of learning as a socially mediated process whereby people acquire higher-order abilities by interacting and using cultural tools. The core of this opinion is the Zone of Proximal Development (ZPD) whereby learners can better perform when assisted by More Knowledgeable Other (MKO) by scaffold. In language education, this means that any tools and interlocutors that make contingent support, modeling and guided practice possible can speed up development through making linguistic and cognitive processes visible and controllable to the learners (Cai et al. 2025).

Generative AI is discussed in this paper as a possible mediational device capable of operationalizing scaffolding in practice, such as the generation of leveled texts, exemplification, simulation of dialogue to practice speaking, and feedback on drafts by providing iteration. Nevertheless, regarding the sociocultural theory, it is also presupposed that mediation is not neutral, but the quality of support is determined by the suitability of guidance, needs of the learner, and staging of the activity by the teacher. This is consistent with the Teacher Cognition attention to professional judgment and with TAM attention to perceived pedagogical value; teachers tend to consider AI helpful when it actually aids in the ZPD-oriented learning and opposed to it when it endangers the agency of learners, their authenticity, or assessment integrity. In turn, when Vygotsky is included, the explanatory coherence of the study will be enhanced as the beliefs (PU/PEOU) and cognitions of teachers will be connected to a clear description of how the AI-mediated scaffolding can be performed during the planning, instruction, and assessment in Malaysian secondary ESL classrooms.

2.4 GENERATIVE ARTIFICIAL INTELLIGENCE (AI) IN EDUCATION

This section explores the technological phenomenon central to this study: Generative Artificial Intelligence (GenAI). While traditional educational technologies have long supported classroom instruction, the emergence of GenAI represents a paradigm shift, offering tools capable of creating novel content rather than merely analyzing existing data. The following sub-sections define this technology within the educational context, tracing its evolution from earlier forms of AI, and critically examining its specific applications and implications for English as a Second Language (ESL) teaching and learning. To strengthen the section’s argument, the discussion is anchored more explicitly in empirical studies already cited in Chapter 1, and it maintains a consistent emphasis on learner outcomes, while recognizing that classroom impact is mediated through teacher-guided use (Law 2024; Vera 2023; Wei 2023; Cai et al. 2025).

2.4.1 Evolution and Definition of Generative Artificial Intelligence (AI)

To situate the present study, it is essential to trace the technological evolution that has led to the current focus on Generative AI. Artificial Intelligence has progressed from early rule-based systems to sophisticated machine learning models, a trajectory historically dominated by Discriminative AI, systems designed to analyze, classify, and predict based on existing data patterns, such as identifying spam or recognizing faces (Kılınç & Keçeçioğlu 2024; Wang et al. 2025). While powerful for automation, these systems were inherently limited to interpreting pre-existing information.

The emergence of Generative Artificial Intelligence (GenAI) marks a paradigm shift within this trajectory. Defined as a subset of deep learning, GenAI utilizes architectures like Generative Adversarial Networks (GANs) and Transformers to create novel, human-like content including text, images, audio, and code by learning from vast datasets (Kılınç & Keçeçioğlu 2024; Reddy et al. 2025). This shift from classifying to generating is not merely technical but fundamentally alters the technology’s role in educational contexts. Whereas traditional AI served as an analytical assistant for tasks like automated grading, GenAI functions as a creative partner, capable of generating original lesson plans, dialogues, and instructional materials (Wang et al. 2025).

In education, this functionality is categorized into practical modalities like text-to-text and text-to-image generation, which offer direct applications for teaching (Law 2024). For ESL instruction specifically, this means AI is no longer just a tool for error detection but can actively scaffold learning by producing language in context, simulating conversations, and personalizing content. Understanding this distinction between discriminative and generative AI is therefore critical, as it frames the technology not as a simple efficiency tool but as a potentially disruptive agent of pedagogical innovation, whose adoption and integration will be heavily influenced by how teachers perceive its novel capabilities and associated challenges.

2.4.2 Generative Artificial Intelligence in English as a Second Language (ESL) Teaching and Learning

The integration of Generative AI into ESL education represents a paradigm shift from static instruction toward dynamic, personalized learning, as it introduces tools capable of generating novel content and adapting interactively to learner needs (Liao 2023). This functionality positions GenAI as a versatile on-demand tutor and scaffold, offering potential solutions to persistent challenges such as mixed-ability classrooms and limited teacher contact time (Shakri et al. 2025; Jamaluddin et al. 2025). The following synthesis examines the demonstrated applications and impacts of GenAI across core language skills, with particular attention to evidence from the Malaysian educational context. In line with the study’s broader framing, the emphasis here is placed primarily on how GenAI shapes student learning experiences and outcomes, rather than presenting the technology as a “teaching tool” in isolation.

In the domain of speaking, GenAI tools primarily function as conversational partners to address the critical lack of interactive practice. Research indicates these tools enhance communicative competence by providing instant, personalized dialogue practice (Muniandy & Selvanathan 2024) and play a significant psychological role by reducing foreign language speaking anxiety through non-judgmental, safe spaces for practice (Ding & Yusof 2025). This benefit is corroborated by studies on specific platforms that provide targeted feedback on pronunciation and fluency (Zainuddin & Mohamad 2024). Crucially, within the Malaysian secondary school context, AI integration has been linked not only to improved speaking proficiency but also to enhanced student creativity (Hassim et al. 2023), validating its role in facilitating the necessary output for language acquisition (Manaf et al. 2025; Sok & Shin 2025).

For writing instruction, GenAI’s impact extends beyond error correction to act as a dynamic cognitive scaffold. It supports complex writing processes, including idea generation, genre structuring, and iterative revision (Feng et al. 2025). In Malaysian secondary schools, where students often struggle with essay organization, tools like Google Gemini have successfully guided learners through pre-writing and drafting stages, reducing writer’s block and improving structural coherence (Jen & Salam 2025). This aligns with broader findings that combining AI with workshop models boosts student mastery and confidence through immediate feedback (Jaramillo et al. 2025). Importantly, the influence extends to learner psychology, with AI-assisted writing shown to positively shape students’ “Ideal L2 Self” and increase intrinsic motivation (Huang & Mizumoto 2025).

In reading comprehension, GenAI serves as an adaptive scaffolding agent by simplifying texts and creating personalized resources. While digital reading tools are common, there is a noted need for systems that intelligently adapt content to varying proficiency levels (Tessensohn et al. 2025). GenAI meets this need; for instance, students use tools like ChatGPT to deconstruct complex texts and generate summaries, significantly reducing cognitive load and improving engagement (Shah et al. 2025; Hutapea et al. 2025). This application shows particular promise for addressing equity gaps in Malaysia. Interventions such as the “Chat2Comprehend” method, which used ChatGPT to translate and simplify passages for rural primary students, led to marked improvements in comprehension and class participation (Henry 2025). Such studies confirm that AI-mediated scaffolding can effectively bridge proficiency divides in mixed-ability settings (Hidayat 2024).

Nevertheless, despite the positive learning results of several studies, these advantages should not be interpreted as automatic or universal. Experimental evidence indicates that language skills and motivation gains are most probable when the use of GenAI is organized as a kind of assistance instead of a substitute to the efforts of learners (Vera 2023; Wei 2023; Dikaprio & Diem 2024; Qiuyang, 2025). That is, GenAI is best thought of as a structure that could make practice opportunities and feedback feedback quicker, yet still it needs to be cognitively framed pedagogically to make learners think critically about language activities instead of just being outsoured into the production. This is in line with the already defined Vygotskian logic in Chapter 1, namely that scaffolding can only be effective when it helps the learners to move through the progression of their Zone of Proximal Development, and that scaffolding can and must be gradually decreased as the learners become competent (Vygotsky 1978; Cai et al. 2025).

The second problem is contextual fit, because of the pressures of curriculum and high-stakes testing, an ESL environment in Malaysia is predisposed to promoting efficiency-oriented applications of GenAI (e.g., the generation of model answers) rather than the nurturing of communicative skills. This implies that the same technology can be utilized to promote the processes of deeper language learning or superficial performance strategies, based on the way learning activities and assessment expectations are implemented (Khan et al. 2025; Dong 2024). Thus, the arguments concerning GenAI making people more skilled should be considered, referring to the realities of the classroom limitations and motivation of students, but not to the tool itself.

Collectively, this literature underscores GenAI’s transformative potential across language skills, particularly in addressing contextual challenges like large class sizes and proficiency diversity. However, the predominant focus remains on student outcomes and tool efficacy. This creates a pivotal gap for the present study, a lack of in-depth investigation into how ESL teachers perceive these benefits and, more importantly, how they navigate the integration of such versatile tools within the complex realities of their classrooms, curriculum demands, and assessment pressures.

2.4.3 Challenges and Ethical Concerns of Generative Artificial Intelligence in ESL Learning

Despite its transformative potential, the integration of Generative AI into ESL education introduces significant ethical and pedagogical challenges that shape its perceived viability. These concerns are interconnected, forming a complex landscape of risk that educators must navigate.

The foremost challenge is the profound threat to academic integrity. In Malaysia, stakeholders’ primary anxiety centers on AI-facilitated plagiarism and “contract cheating,” where students might bypass the learning process entirely (Tang & Chaw 2023). This concern is validated by findings that students may rationalize AI-assisted cheating as a legitimate survival strategy amidst academic pressure, normalizing behavior that undermines assessment validity (Giray et al. 2025). The problem is technically compounded by the difficulty of detection, as AI-generated texts can mimic high-proficiency human writing, making it challenging for teachers to discern authenticity without specialized tools (Mizumoto et al. 2024). More importantly, this difficulty also indicates that the solution to the issue of integrity cannot be found solely in detection, but it forces the reconsideration of the design of tasks and evaluation practice in such a way that classroom tasks do not focus on the polished final products but on the process of learning and its evidence (Dong 2024). It is in this sense that the integrity problem is not simply a student misconduct, but a tension inherent in the assessment systems together with new tools that reduce the cost of generating a fluent language output.

In addition to integrity, the credibility and correctness of AI outputs pose a pedagogical threat directly. Although fluent, technologies such as ChatGPT are not accurate; they may commit false corrections or overlook subtle mistakes, which may encourage learners to further accept the mistakes as correct patterns of language use (Alsaweed and Aljebreen 2024). Such hallucination effect is particularly harmful in the case of ESL settings, where students can accept the AI as one of the authoritative sources. Learning wise, the danger is not just the presence of the wrong answers, but the fact that learners adopt the false language models and it is only the metalinguistic awareness that allows them to identify and fix them. It supports the significance of teacher mediation and explicit training of learners on verification strategies particularly when GenAI is employed as an informer as opposed to a brainstorming helper (Cai et al. 2025; Tran et al. 2025b).

These concerns converge in the risk of cognitive dependency, which threatens the development of higher-order thinking skills. Passive, uncritical use of AI can degrade students’ critical thinking and problem-solving abilities (Nasr et al. 2025). Within the Malaysian context, research indicates a fine line between assistance and over-reliance, with students often using AI to bypass the essential cognitive struggle of writing (Ahmad et al. 2025). This suggests that the entrenched exam-oriented culture may incentivize students to prioritize AI’s efficiency over genuine learning, highlighting the need for carefully framed pedagogical implementation (Dong 2024). In a more objective way, the issue of dependency is not a reason to avoid GenAI, but a caution that scaffolding should be created to ensure that the learner agency is not taken away. AI assistance should not offer pre-written language that takes the decision-making part of the work done by the learner (Vygotsky 1978; Cai et al. 2025).

Lastly, the accessibility to infrastructure and devices is also an ethical consideration of GenAI in ESL learning, where the achievement gap may grow or be reproduced should GenAI-assisted learning become normalized without equal distribution. The latter is especially applicable in light of the ongoing inequality in digital access in school-related environments, particularly in rural ones (Ismail 2025; Ahmad and Rathakrishnan 2025). Thus, learner benefit itself depends not only on pedagogy and integrity protection, but also on the ability of the wider ecosystem to make the digital conditions that warrant meaningful GenAI-based practice fairly accessible.

Collectively, these challenges reveal that the integration of GenAI is not merely a technical adoption but a significant ethical and pedagogical negotiation. The extent to which teachers are concerned by these issues will critically influence their Perceived Usefulness and ultimately their willingness to integrate such tools, forming a key dimension of the attitudes explored in this study.

2.5 ENGLISH AS A SECOND LANGUAGE (ESL) TEACHERS’ ATTITUDES TOWARDS GENERATIVE ARTIFICIAL INTELLIGENCE IN TEACHING AND LEARNING

This section examines the current body of literature regarding ESL teachers’ attitudes toward the integration of Generative AI, addressing the study’s first research question. Grounded in the Technology Acceptance Model (TAM), understanding these attitudes specifically through the lenses of Perceived Usefulness and Perceived Ease of Use is critical, as teachers serve as the primary gatekeepers of educational innovation. The following sub-sections review empirical findings on the perceived pedagogical benefits that drive positive acceptance, as well as the specific challenges and resistance factors that may hinder the widespread adoption of these tools in the classroom.

2.5.1 Perceived Benefits of AI Adoption

In the global landscape of English Language Teaching (ELT), positive teacher attitudes toward Generative AI are strongly correlated with Performance Expectance., the belief that the technology will enhance job performance. Studies identify this construct, aligned with the Technology Acceptance Model’s Perceived Usefulness (PU), as a primary driver of adoption, with educators valuing AI for tangible efficiencies like automated grading and resource generation (Firdaus & Nawaz, 2024). This establishes a key premise, that is acceptance hinges on the tool’s perceived capacity to solve practical problems.

Within the Malaysian context, this perceived utility is uniquely filtered through pressing curricular and logistical demands. Here, AI’s benefit is critically evaluated against its ability to enable Differentiated Instruction (DI) within the KSSM framework. Teachers perceive AI not merely as an administrative tool but as a vital scaffold for personalizing content and processes for diverse learners, a task often unmanageable in overcrowded classrooms (Ken et al. 2025; Nasrudin & Hashim 2025). This represents a significant contextualization of PU, where usefulness is tied directly to achieving national inclusivity goals.

The perception of benefit further extends to pedagogical partnership in high-stakes skill development. For instance, secondary ESL teachers value AI as a cognitive scaffold in the writing process, utilizing it to help students overcome writer’s block and develop ideas during the crucial “ideation phase,” thus supporting the process-oriented approach mandated by the syllabus (Aineh & Ngui 2024). This nuanced view where AI is seen as augmenting instruction rather than replacing it is echoed in findings that teachers hold positive perceptions when the technology enhances, rather than supplants, their pedagogical role (Sivanganam et al. 2025). Consequently, the dominant narrative of AI’s benefit in Malaysia evolves from one of simple efficiency to a more complex appreciation of its role in addressing specific instructional challenges and aligning curricular priorities.

Most importantly, though, the purported advantages of AI in the literature are usually conditional: the perceived utility of AI would be maintained only when teachers are capable of checking AI outputs, aligning them to the results of CEFR/KSSM, and ensuring that they had the professional right to final decisions on instruction and assessment. In reality, the automation time savings may be partially compensated by the cognitive and ethical labour of timely design, output checking, and plagiarism prevention, which might restructure as well as cut down workload (Ling and Jan 2025; Tran et al. 2025b). This means that positive attitudes cannot be interpreted as blanket support of GenAI, but rather, as a negotiative position based on the pedagogical experience of teachers, school systems of support and the perceived fit between AI application and exam-based accountability.

2.5.2 Perceived Challenges and Resistance of AI Adoption

In the global discourse on educational technology, teacher resistance to Generative AI is rarely a simple rejection of innovation but stems from interlinked pedagogical and ethical deficits. Research framed by the Technological Pedagogical Content Knowledge (TPACK) model reveals that a critical barrier is not a lack of technological skill per se, but a gap in Technological Pedagogical Knowledge (TPK), the know how to integrate AI effectively into specific teaching contexts (Tran et al. 2025b). Compounding this is profound ethical anxiety; without training in AI ethics, teachers struggle to navigate the boundary between legitimate assistance and academic misconduct, leading them to perceive AI more as a threat to assessment validity than a pedagogical tool (Karaduman 2025; Chan & Tang 2025). This global landscape establishes that Perceived Ease of Use (PEOU) is undermined not by interface complexity alone, but by the lack of pedagogical frameworks for AI’s classroom management.

Within the high-pressure ecosystem of Malaysian secondary education, these universal concerns are intensified and transformed. Resistance here is deeply tied to the pedagogical paradox of the KSSM curriculum. ESL teachers fear that AI, rather than scaffolding learning, might facilitate intellectual indolence, allowing exam-focused students to bypass the cognitive struggle essential for developing the very critical thinking skills the curriculum promotes (Aineh & Ngui 2024). This fear directly erodes Perceived Usefulness (PU); if AI is seen to undermine core learning objectives, its value is negated. Furthermore, concerns about reliability and the loss of human interaction are heightened in contexts with mixed-ability learners, where teachers worry about error fossilization (Nasrudin & Hashim 2025).

Critically, this hesitation is structurally reinforced by a systemic “Training-Practice Gap.” While teachers may express openness to AI’s potential, they report a significant degree of uncertainty in implementation due to a lack of clear, practical guidelines (Sivanganam et al. 2025). Studies indicate that awareness of AI does not translate to readiness, as teachers lack the specific, hands-on professional development needed to navigate its ethical and pedagogical complexities autonomously (Shakri et al. 2025). Therefore, the resistance observed is not mere technophobia but a state of pragmatic caution: a willingness constrained by an ecosystem that has yet to provide the necessary scaffolding for teachers to move from apprehension to confident, pedagogically sound integration.

In general, the attitudinal data is evidence of an ambivalence trend and not an acceptance-rejection dichotomy. The very teachers who acknowledge the usefulness of GenAI in differentiation and process writing might at the same time oppose it when it endangers assessment validity, development of critical thinking or even the relationships aspect of language learning (Aineh and Ngui 2024; Karaduman 2025; Nasrudin and Hashim 2025). Notably, a significant portion of recent research is survey-based and is based on self-reported intentions, which may overstate preparedness under policy-directed digitalisation and do not reflect the real-life focus of teachers to implement the concept of human-in-the-loop in their classrooms (Sivanganam et al. 2025; Shakri et al. 2025). Consequently, a much more critical, qualitative explanation is needed of how ESL teachers in Malaysia arbitrate these opposing logics in everyday context, and how location influences the transfer of attitudes into pedagogically legitimate integration.

2.6 INTEGRATION OF GENERATIVE ARTIFICIAL INTELLIGENCE IN ESL CLASSROOM PRACTICES

The current section summarizes the existing empirical literature on the topic of the operationalization of Generative AI by ESL teachers, that is, it specifically answers the second research question of the study. Based on the Teacher Cognition Framework, the review will be structured on a functional basis to analyze how the teachers incorporate these tools into the pedagogical continuum, including the pre-active instructional design and resource development into the interactive classroom implementation and post-active assessment. Lastly, it is also a critical method of analyzing the mediating nature of contextual factors that influence the degree and effectiveness of such integration in the Malaysian educational ecosystem. Theoretically, this part of the paper considers integration as the enactive implementation by teachers of GenAI (intentional and context-specific) in planning, instructional and assessment processes, which are informed by professional judgement, classroom factors and curriculum-accountability priorities. More importantly, the evaluated studies are more likely to explain the purpose of GenAI, but less likely to address how teachers select specific applications, how they judge and modify AI productions and how conflicting demands (such as efficiency and pedagogical integrity) are decided in practice- an analytic gap that reinforces the relevance of the current study to understanding teacher cognition and the lived classroom realities.

2.6.1 AI for Instructional Design and Resource Creation

In the planning and preparation stage of teaching, Generative AI has emerged as a transformative tool for instructional design, shifting the teacher’s role from consumer of static resources to designer of bespoke learning materials (Kerr & Kim 2025). Globally, research identifies material design and planning as a primary use case, where educators leverage AI to manage cognitive load and enhance pedagogical precision. Studies note that teachers utilize GenAI to generate lesson frameworks and target-specific exercises, allowing them to focus on refining pedagogy rather than routine creation (Alzubi & Alelaiwi 2025; Kılıçkaya & Kic-Drgas 2025). This practice evolves from initial exploration towards more sophisticated, pedagogy-driven integration where prompts are crafted to ensure cultural and curricular relevance.

Within the Malaysian context, this utility is critically evaluated against systemic imperatives: overwhelming workload and the curricular mandate for differentiated instruction (DI). Research indicates that while teachers and trainees use AI to generate lesson ideas and visual aids, a significant challenge lies in adapting or customizing this content to fit the local cultural and linguistic context of Malaysian classrooms (Kussin et al. 2024). Despite this hurdle, the drive to support diverse learners fosters adoption.

In resource-constrained environments, this function transitions from convenience to necessity. For instance, teachers in low-enrolment schools use ChatGPT to expedite the creation of lesson plans and administrative documents, preserving instructional time despite limited support (Thong & Kamsin 2025). A concrete illustration is the “Chat2Comprehend” method, where a teacher used AI to translate and simplify complex texts for rural students, directly generating accessible “comprehensible input” that standard resources lack (Henry 2025).

Collectively, this literature reveals that AI integration in the pre-active phase is deeply influenced by contextual factors, such as workload, class composition, and resource availability that shape how teachers’ knowledge and beliefs about DI are translated into material design practice. However, it leaves unexplored the cognitive processes behind these decisions, that is how teachers judge the quality of AI output, adapt it for their students, and reconcile its use with their professional identity in an exam-driven system.

More importantly, the pre-active literature implies a conflict between efficacy-first integration (AI as a time-saving generator) and pedagogy-first integration (AI as a design partner that needs to be prompted, evaluated and localised repeatedly). Whereas global studies tend to establish sophistication in prompting and refinement as a measure of relevant integration (Alzubi and Alelaiwi 2025; Kılıçkaya and Kic-Drgas 2025), Malaysian studies preempt the translators work that teachers must perform in order to bring outputs to be culturally and linguistically relevant under real workload pressures (Kussin et al. 2024; Thong and Kamsin 2025). This comparison supports the necessity of the current research to explore the way educators make decisions concerning what is considered as good enough AI-generated content to be used by mixed-ability students and high-stakes-curriculum requirements.

2.6.2 AI as an Interactive Instructional Tool

During the interactive phase of instruction, Generative AI redefines the teacher’s role from a sole knowledge authority to a facilitator managing a dynamic learning environment where AI acts as a real-time pedagogical partner. Global reviews position GenAI as an intelligent interlocutor and co-creator, capable of providing immediate feedback and sustaining dialogue, thereby enabling a shift from lecture-based to interactive, student-centered models (Lee et al. 2025) but also as a vital instrument for classroom management. Yugandhar and Rao (2024) argue that AI tools significantly enhance instructional quality by automating routine monitoring and establishing structured routines, thereby fostering a positive and disciplined environment that encourages active student participation. This dual capacity enables a shift from lecture-based models to interactive, student-centered ecosystems. However, the translation of this potential into practice is uneven. Research suggests that in many contexts, including Southeast Asia, AI use remains skewed toward lesson preparation rather than live classroom engagement, indicating a significant implementation gap between technological access and pedagogical integration (Syafrayani et al. 2024).

Emerging studies from Malaysia, however, illustrate how teachers are actively navigating this gap by embedding AI into the flow of instruction to address specific contextual challenges. Here, integration is characterized by real-time scaffolding. For example, teachers use AI as an ideation partner during writing lessons, guiding students to prompt tools like Google Gemini for structures and vocabulary, thus providing immediate cognitive support that mitigates writer’s block, a critical hurdle in exam-focused classrooms (Jen & Salam 2025; Xiaofan & Annamalai 2025).

Furthermore, Malaysian educators leverage AI to simulate authentic practice and differentiate instruction instantaneously. To overcome the interaction deficit in large ESL classes, teachers employ ChatGPT as a conversational partner for students to practice speaking in a low-stakes setting before peer performance (Muniandy & Selvanathan 2024). In such scenarios, the teacher’s role evolves into a monitor and facilitator, curating the interaction to ensure pedagogical alignment (Robert et al. 2025). Similarly, teachers use AI to dynamically adjust task difficulty, such as simplifying complex reading passages in real-time based on student comprehension, ensuring continued engagement across proficiency levels (Henry 2025).

Collectively, these practices reveal a context-driven model of interactive AI use focused on scaffolding complex tasks and personalizing the learning experience as it unfolds. The literature thus begins to map the what of interactive integration. However, it provides limited insight into the teacher cognition underlying these moves: the in-the-moment decision-making, the assessment of AI’s suitability for a given interaction, and the management of the triadic relationship between teacher, student, and AI tool within the pressurized Malaysian classroom.

Critically, evidence of the interactive phase is quite varied: on the one hand, there are reports that GenAI allows facilitating communicative, student-centred engagements (Lee et al. 2025; Muniandy and Selvanathan 2024), and, on the other hand, it is stated that at present, the interactive use is still futuristic in most of the places where classroom integration is a step behind preparation use (Syafrayani et al. 2024). The Malaysian evidence suggests that interactive integration is most probable in the case when GenAI is discussed as scaffolding the immediate constraints (e.g., large classes, speaking anxiety, writer’s block), however, the literature does not specify the reasoning by which teachers will convert the risks (misinformation, over-reliance, mismatch with lesson objectives) into the professional risk management in real time. This reinforces the argument to investigate the on-the-spot decision rules and classroom management instructional approaches surrounding GenAI in teachers.

2.6.3 AI for Assessment, Feedback and Administrative Task

The post-active phase of teaching encompasses assessment, feedback, reflection, and follow-up actions. In this stage, Generative AI is increasingly utilized to automate routine tasks and enhance the quality of feedback, impacting both teacher workload and student learning outcomes. Global studies show that teachers use AI to draft feedback comments, generate rubrics, and create assessment items, thereby increasing efficiency and consistency (Kerr & Kim 2025). These tools are particularly valued for formative assessment, where rapid, personalized feedback can guide learning and reduce the burden on teachers managing large classes.

However, the integration of AI in grading reveals a profound pedagogical tension between formative support and summative trust. For formative feedback, especially in writing, AI is valued for providing “objective and consistent” suggestions that can complement or extend teacher input (Khamis & Yusof 2024). Yet, for high-stakes summative assessment, significant caution prevails. Teachers express reluctance to cede final grading authority to AI, citing concerns about its inability to evaluate nuanced elements like voice and authentic communicative competence (Aineh & Ngui 2024; Zainuddin & Mohamad 2024). This indicates that AI’s post-active role in Malaysia is defined by a “human-in-the-loop” model: fully embraced for generating materials and drafting feedback but rigorously supervised and ultimately subordinate to teacher judgment when definitive evaluation is required.

This literature effectively maps the terrain of practice, what AI is used for and the general boundaries of that use. It leaves a critical gap in understanding the teacher cognition that governs these boundaries, that is the decision-making processes, the internal criteria teachers use to determine when AI’s feedback is good enough, and how they negotiate the tension between efficiency and professional authority in an exam-driven culture.

Importantly, the post-active literature is characterized by a more defined boundary of what trust is and is not than the pre-active and interactive stage: teachers might accept GenAI as a productivity aid (crafting feedback, writing rubrics), but they do not accept it as an authority in high-stakes judgement (Aineh and Ngui 2024; Zainuddin and Mohamad 2024). This is not just a technical boundary but a professional and ethical one and is concerned with the issue of validity, fairness, and maintainability of teacher expertise. In the current research, this is significant as it implies that integration is a matter of phase and that the willingness of the teachers can vary between various tasks, though the same tool is utilized, which is why it is necessary to investigate the criteria according to which the teacher makes decisions about which tasks can be assigned to AI and which issues can be left to human attention.

2.6.4 The Mediating Role of Contextual Factors

The Teacher Cognition Framework underscores that the translation of pedagogical knowledge and positive attitudes into classroom practice is not direct but is powerfully mediated by contextual factors (Borg 2003). Within the Malaysian ESL landscape, these factors constitute an ecosystem of constraints and enablers that critically determine whether Generative AI integration progresses from intention to implementation. The literature reveals three interconnected dimensions of this ecosystem: technological infrastructure, institutional culture and support, and systemic curricular pressures.

The most foundational barrier remains the persistent digital divide. In low-enrolment and rural schools, the absence of stable internet access and adequate hardware renders advanced, cloud-based AI tools functionally unusable, creating a chasm between technological policy and classroom reality regardless of a teacher’s personal readiness (Thong & Kamsin 2025). This lack of basic “Facilitating Conditions” means that for many educators, AI integration remains an abstract concept rather than a practicable option. Where infrastructure exists, institutional leadership and professional development emerge as decisive mediators. A principal’s “digital vision” and active support through resource allocation and fostering a culture of innovation are significant predictors of successful teacher adoption (Zeng et al. 2025). Conversely, the absence of clear leadership manifests as a “support void,” leaving teachers to navigate ethical and technical complexities in isolation. Studies note that without specific, hands-on training and coherent institutional guidelines, teachers experience significant uncertainty, leading to hesitant and fragmented implementation practices (Shakri et al. 2025).

Ultimately, even with adequate tools and support, integration is tested against the overwhelming systemic pressures of the Malaysian secondary system. The exam-oriented culture, coupled with heavy administrative workloads and a packed syllabus, creates a context where efficiency is paramount (Ag-Ahmad 2025; Chandran 2022). Paradoxically, while AI promises time savings, its initial adoption often increases cognitive load as teachers must critically evaluate AI outputs for accuracy and suitability (Ling & Jan 2025). Consequently, innovative practices are frequently abandoned in favor of familiar, exam-focused methods that align with immediate performance demands, suggesting that AI must prove its ecological viability within this high-stakes environment to be sustained (Aineh & Ngui 2024).

The contextual evidence, on the whole, suggests that the integration of GenAI in Malaysian ESL is more of a negotiated practice, rather than a linear process of adoption, due to uneven infrastructure, institutional direction inconsistencies and pressure to focus on examination results. This is one of the reasons why the positive attitudes to the potential of GenAI do not necessarily result in the classroom adoption (Borg 2003; Shakri et al. 2025). In the current research the results would support the study of how teachers combine the use of GenAI or not, but how the predilection to situation and the enabling conditions respond to the belief of the teachers to become confident and pedagogically sound users or ambivalent and intermittent experimenters.

2.7 SUMMARY

This chapter provided a critical review of the theoretical and empirical literature surrounding Generative AI in ESL instruction, framed by the Technology Acceptance Model (TAM) and Teacher Cognition Framework. The analysis mapped the evolving role of AI across instructional design, interactive teaching, and assessment, while highlighting how uniquely Malaysian contextual factors, such as the digital divide and exam-oriented pressures powerfully mediate these integration practices. Ultimately, the review identified a critical gap in understanding the internal decision-making and cognitive negotiations of teachers within this constrained ecosystem, providing the theoretical imperative for the qualitative methodology detailed in Chapter III.

CHAPTER III

3.1 INTRODUCTION

This chapter outlines the research methodology employed to achieve the objectives of the study. It provides a detailed description of the research design, the population and sampling strategies, and the instruments utilized for data collection. Furthermore, the chapter explicates the specific procedures for data collection and data analysis, concluding with a discussion on the ethical considerations and the measures taken to ensure validity and reliability. As the name and purpose of the study suggest, the chapter also explains how the selected methodology will allow exploring the attitudes of ESL high school teachers to generative AI (through the prism of TAM constructs) and how these attitudes are reflected in classroom practice (through the Prism of Teacher Cognition Framework).  The chapter has been presented in such a way that it gives a clear research methodology channel between the research questions and the analytical outcomes. It starts by explaining why the overall qualitative case study design is valid, followed by the description of the method of participant selection, which is the narrow context of the Malaysian secondary ESL, the procedures of the semi-structured interviews, and the thematical analysis procedures to derive the meaning of the data. Lastly, it introduces ethical protection and trust measures that are employed in order to guarantee credibility and integrity of the findings.

3.2 RESEARCH DESIGN

The research design serves as the logical blueprint that aligns a study’s philosophical assumptions, research questions, and methods of inquiry. As Ansari et al. (2022) emphasize, a coherent design provides the specific procedures for collecting, analyzing, interpreting, and reporting data, ensuring methodological rigor and a direct pathway to answering the research problem.

3.2.1 Qualitative Research Approach

Guided by the study’s objectives to explore the nuanced attitudes and practices of ESL teachers, a qualitative research approach was adopted. Qualitative inquiry is characterized as an iterative process focused on understanding meaning and social phenomena from participants’ perspectives, seeking to get closer to the phenomenon studied (Lim 2025). This approach is indispensable for investigating the complex, internal dynamics of teacher cognition, including the beliefs, knowledge, and decision-making processes that underlie classroom practice which cannot be reduced to numerical data. In particular, the qualitative approach is appropriate because the study seeks to understand teachers’ evaluative beliefs about generative AI (e.g., perceived usefulness and perceived ease of use), and the professional reasoning through which teachers decide whether, when, and how to use these tools in English language teaching. The qualitative paradigm allows for a flexible, in-depth exploration of how teachers perceive Generative AI and why they integrate it in particular ways within their unique contexts.

3.2.2 Basic Qualitative Research Design

Within the qualitative paradigm, a basic qualitative research design has been selected as it is uniquely suited to answer this study’s “how” and “why” questions regarding the process of teacher mediation. According to Merriam and Tisdell (2016), the primary goal of such a design is to uncover and interpret the meanings individuals construct about their world and experiences. Unlike a case study, which is bounded by a specific system, or phenomenology, which seeks a universal essence of lived experience, a basic qualitative approach is most appropriate because it focuses explicitly on understanding a process. In this case, the process of how teachers interpret, negotiate, and mediate the integration of Generative AI (Tripathi et al. 2025.

This design choice is directly supported by the study’s Conceptual Framework, which positions “Teacher Cognition” and “Attitudes” as central to the adoption process. As Percy et al. (2015) note, a basic qualitative approach is ideal when seeking to understand a phenomenon across individuals without rigidly bounding the environment as a case. It therefore allows the researcher to prioritize teachers’ subjective reasoning and explore how they mentally filter AI tools through their pedagogical beliefs, rather than focusing primarily on institutional structures.

The suitability of this design is confirmed by its successful application in recent AI-education research. For instance, Farazouli et al. (2024) employed a similar qualitative approach to explore how university teachers navigated the uncertainty of integrating AI tools. By utilizing focus group interviews without confining the study to a single institution, they identified common cognitive themes, such as professional vulnerability and the renegotiation of trust in assessment across diverse disciplines. Similarly, Dincer and Bal (2024) used a qualitative methodology to examine EFL instructors’ beliefs about AI integration. Their design enabled a focus on pedagogical dynamics and shifting educator roles, highlighting how teachers internally negotiated their transition from knowledge providers to learning facilitators. These examples justify the use of a basic interpretive approach when the unit of analysis is the teacher’s perspective and internal negotiation, rather than the educational setting itself.

To ensure the findings capture the “Contextual Factors” outlined in the framework, this study recruits participants from secondary schools in one Northern state and one Southern state in Malaysia. These regions represent two distinct socio-economic ecosystems recognized in national development policy. The Northern region is characterized by significant urban-rural imbalances and a reliance on agriculture, with bridging these gaps identified as a primary strategic thrust (NCER Malaysia 2020). In contrast, the Southern region, specifically Iskandar Malaysia, is a metropolitan hub driven by a high-value service economy (Iskandar Regional Development Authority 2015). This distinction is reflected in a significant digital and economic divide: as of 2022, the mean household income in Johor (RM8,517) was nearly 53% higher than in Northern states such as Kedah (RM5,550) (Malaysian Ministry of Economy Department of Statistics 2023). This geographic selection is not a formal and replicative comparison but to ensure the study’s findings regarding systemic barriers and enablers are informed by the broadest possible range of socio-economic and infrastructural realities in Malaysia.

Consequently, the basic qualitative design is precisely suited to address all three research objectives. Its focus on meaning-making elucidates attitudes (RO1); its process-oriented nature reveals integration practices (RO2); and its flexibility within varied sites captures the contextual factors (RO3) that mediate them. This ensures the resulting conceptual framework reflects a nuanced understanding of how environmental factors shape, modify, or halt GenAI integration in the Malaysian ESL context.

3.3 SAMPLING AND POPULATION

To effectively address the research objectives, this study employs a purposive sampling strategy. Purposive sampling is a non-probability technique where participants are deliberately selected based on their specific characteristics and their potential to provide rich, relevant information about the phenomenon under study (Nyimbili & Nyimbili 2024). This approach is fundamental to qualitative inquiry, as it allows the researcher to identify and recruit information-rich individuals whose professional experiences offer the greatest potential for in-depth understanding.

Specifically, this study utilizes a maximum variation sampling approach, a type of purposive sampling to ensure the findings capture a wide spectrum of systemic realities and contextual feasibility (Nyimbili & Nyimbili 2024). To achieve this, the sampling frame is structured across two distinct socio-economic regions in Peninsular Malaysia, that is one state from the Northern region and one state from the Southern region. This regional selection, justified in the research design, is intended to capture maximum variation in the digital, economic, and infrastructural ecosystems that form the professional environment of ESL teachers. The goal is to recruit participants whose experiences can illuminate how such varied systemic contexts mediate attitudes and integration practices.

The primary units of analysis are ESL teachers. Recruitment will target in-service secondary school teachers from purposively selected schools in the Northern and Southern states. Participants must meet six inclusion criteria to ensure the data’s relevance and depth. First, they must teach at a National Secondary School. Second, they must be drawn from schools within both the Northern and Southern regions to facilitate analysis of how distinct regional ecosystems influence practice. Third, they must be currently teaching English under the KSSM curriculum within one of the two selected states. Fourth, they must teach English as their core subject. Fifth, they must have a minimum of two years of full-time teaching experience to ensure their perspectives are grounded in established practice. Sixth, and crucially, they must have direct experience using Generative AI tools in their professional practice for planning, instruction, or assessment. As demonstrated in related research, the purposeful selection of teachers with actual experience provides access to meaningful insights into integration challenges and opportunities (Tripathi et al. 2025). For this study, such first-hand experience is essential to exploring attitudes formed through actual use and examining authentic integration practices, moving beyond hypothetical opinions.

To strengthen sampling transparency and ensure replicability, the inclusion and exclusion criteria are summarized in Table 3.1.

Table 3.1: Summary of Inclusion and Exclusion Criteria

Within each selected state, three public secondary schools will be purposively selected to represent the administrative diversity of the state’s educational system. From each school, two in-service ESL teachers will be recruited, resulting in an initial target sample of twelve participants (six from the Northern state and six from the Southern state). This sample size is established as a baseline sufficiency criterion, informed by Hennink and Kaiser’s (2022) systematic review, which indicates that thematic saturation in qualitative interviews is typically achievable within a range of 9 to 17 participants. However, rather than adhering to a fixed quota, the final sample size will be determined by thematic saturation, defined as the point at which no new codes or themes emerge from the data (Naeem et al. 2024).

3.4 RESEARCH INSTRUMENT

The primary instrument for data collection in this study is a semi-structured interview protocol. This instrument is uniquely suited for qualitative inquiry as it provides a flexible structure, allowing the researcher to maintain focus on core research objectives while offering the freedom to probe deeper into participants’ unique experiences and emergent themes (Lim, 2025). In the context of this study, this flexibility is essential for two interrelated purposes.

First, to explore the internal and cognitive dimensions of teacher attitudes including Perceived Usefulness and Ease of Use as outlined by the Technology Acceptance Model (TAM); and second, to investigate how these attitudes translate into classroom practices and are mediated by contextual factors, in alignment with the Teacher Cognition Framework. By facilitating open-ended, reflective dialogue, the protocol enables the collection of rich, narrative data necessary to address the how and why underlying teachers’ engagement with Generative AI.

In addition, the protocol is developed as a theory-informed instrument, including TAM and Teacher Cognition serve as sensitising frameworks that guide systematic coverage of key constructs, while the semi-structured format preserves openness for unexpected themes (e.g., ethical tensions, institutional constraints, and evolving professional norms) that may not be fully captured by predefined constructs.

3.4.1 Interview Protocol

The interview protocol will be developed by the researcher based on the study’s research questions and theoretical frameworks. Its design is specifically informed by methodological literature on using Technology Acceptance Model (TAM) (Dincer & Bal, 2024; Kampookaew 2020; Huang et al. 2019) and the Teacher Cognition Framework to explore teachers’ perspectives (Chappell et al. 2015; Zaimoğlu & Dağtaş 2025), with questions adapted and contextualized to examine generative AI integration within the specific setting of Malaysian secondary ESL education. The protocol is organized into four thematic sections:

Section A: Demographic and Contextual Profile

This section gathers essential background information to contextualize the participant’s perspective, including years of teaching experience, qualifications, and prior exposure to digital or AI-related training.

Section B: Exploration of Attitudes

Grounded in the Technology Acceptance Model (TAM), this section explores the key constructs of Perceived Usefulness (PU) and Perceived Ease of Use (PEOU). Questions are designed to elicit teachers’ evaluative beliefs about AI. For example:

Section C: Examination of Integration Practices

Informed by the Teacher Cognition Framework, this section investigates the “Classroom Practice” component, probing how beliefs translate into action. It explores AI use across the teaching cycle:

Section D: Interrogation of Contextual Factors

This section explicitly explores the environmental mediators from the Teacher Cognition Framework, examining how infrastructure, institutional support, time constraints, and exam pressures shape the possibilities and limits of integration.

In order to make the development process explicit and methodologically rigorous, the protocol will be developed in the form of a systematic process of mapping: the researcher will (1) identify the key domains based on the literature review (Chapter 2) and two underlying frameworks (TAM; Teacher Cognition), (2) convert each domain into interview prompts and finalize the process of the protocol development with respect to its research objectives, (3) word item drafting and probes using local language that is relevant to the Malaysian secondary ESL setting, (4) revise the protocol with the validation of experts related to the field of study. This will help to keep the process of responding to each question traceable to either (a) TAM constructs framing the concept of attitudes (e.g., perceived usefulness and perceived ease of use) or (b) Teacher Cognition domains framing the concept of integration (planning, instruction, assessment, and contextual mediators), yet emerge as prompts when respondents mention experiences not covered by the domains.

In addition, to further evidence alignment between the literature review, theoretical framing, and the interview structure, Table 3.2 summarises how each protocol section links to the study objectives and constructs.

Table 3.2: Alignment of Interview Protocol Sections to Study Constructs and Objectives

3.4.2 Validity of the Instrument

In qualitative research, validity refers to the accuracy, meaningfulness, and trustworthiness of the findings derived from the data collection instrument (Arslan 2025). Ensuring validity is paramount as it confirms that the instrument, in this case, the semi-structured interview protocol truly measures the constructs it purports to investigate, such as teachers’ attitudes and integration practices. For this study, trustworthiness is secured by validating the data collection instrument itself and the subsequent analysis through two key procedures.

Content validity will be established through a two-stage process. First, expert review of the protocol’s alignment with theoretical constructs; and second, a preliminary study to assess clarity and flow. Content validity refers to the extent to which an instrument’s items sufficiently measure the targeted theoretical constructs (Mokkink et al. 2025). To achieve this, the protocol will undergo content validation by two experts in TESL with demonstrated experience in instrument review and qualitative interviewing.

Transparent evaluation criteria will guide the expert review. Each question will be rated based on three key aspects: (1) content relevance to the theoretical constructs of TAM and Teacher Cognition; (2) language clarity, including the use of non-leading phrasing and the capacity to elicit detailed, experience-based responses; and (3) suitability for context within Malaysian secondary ESL classrooms. Feedback from the experts will be used to revise question wording, eliminate ambiguity, and ensure the protocol comprehensively and appropriately addresses teacher attitudes, practices, and contextual influences.

Following expert review, a preliminary study will be conducted with two TESL teachers who meet the participant criteria but will not take part in the main study. This preliminary study serves as an important check of the instrument’s feasibility, timing, clarity, and flow in a realistic setting (Kimotho 2025). Interviews will be conducted under conditions similar to the main study, followed by a debriefing session to gather feedback on question comprehension, relevance, and overall interview experience. The debriefing will also specifically examine whether prompts inadvertently impose TAM language (e.g., “usefulness” or “ease”) at the expense of emergent concerns (e.g., ethics, workload, or assessment integrity), and whether probing questions effectively capture integration across planning, instruction, and assessment phases. Any issues related to phrasing, sequence, or duration will be refined before formal data collection begins, thereby enhancing the instrument’s face validity and practical reliability.

3.5 DATA COLLECTION PROCEDURES

The data collection for this study will follow a systematic, multi-phase procedure to ensure ethical compliance and the acquisition of rich qualitative data. The procedure is divided into three phases: preparation, recruitment, and fieldwork.

3.5.1 Preparation and Ethical Approval

Firstly, preparation and ethical approval. Prior to participant contact, formal approval will be sought to conduct research within the national school system. This will involve obtaining permission from the Educational Planning and Research Division (EPRD) of the Ministry of Education Malaysia and the relevant State Education Department. Concurrently, the semi-structured interview protocol will be finalized following the expert validation process

Moreover, the researcher will seek the permission of school-level access (where necessary) (e.g. by school principals/administrators) so that the study can be carried out not to harm teaching work and not being inconsistent with the institutional procedures. Before the data collection, the researcher will also develop the Participant Information Sheet and the consent form, the interview schedule, audio-recording equipment/softwares and a safe data management folder system (separating consent forms, recordings and transcripts) to reduce the chances of identification. In the case when the interviews are done online, the researcher will pretest the chosen secure video-conferencing software, including privacy settings (e.g., password-protected meeting links, waiting rooms) to avoid any unauthorised left access.

3.5.2 Participant Recruitment and Informed Consent

Next, participant recruitment. Upon receiving administrative clearance, the researcher will initiate recruitment using the purposive sampling strategy defined. The heads of the English panels at selected secondary schools will be contacted to help identify teachers who meet the inclusion criteria. Potential participants will be contacted via official email or professional channels. They will receive a Participant Information Sheet detailing the study’s purpose, voluntary nature, confidentiality measures, and data handling procedures. Written informed consent will be obtained from all willing participants prior to scheduling interviews. Following consent, participants will receive a formal appointment letter. This letter serves to formalize the arrangement and provide participants with clear documentation of their commitment. To reinforce ethical transparency, the participants will be expressly told that their involvement in the study is voluntary, and they do not have to answer any question, and they can drop out of the study without repercussions. In the event of withdrawal, the cut-off points of withdrawal (e.g., until the start of formal coding/analysis) will be made known to the participant (after which anonymised data will not be practically retractable). The participants will also be advised that the research is geared towards professional experiences and perceptions; hence, they must not disclose any identifiable details of any student in the process of the interviews, and the researcher will redirect the interviews where identifiable statements arise.

3.5.3 Execution of Interviews

Last but not least, execution of interviews. Data collection will center on conducting individual semi-structured interviews to ensure privacy and encourage candid responses. Interviews will be scheduled at times convenient for participants and will be conducted either face-to-face or online via a secure video-conferencing platform, based on participant preference and feasibility.

During each interview, the researcher will follow the validated protocol while using probing techniques to explore responses in greater depth. Interviews are anticipated to last 30 to 45 minutes. With explicit permission, all sessions will be audio recorded to ensure accuracy. Recordings will be securely stored and transcribed verbatim for analysis, with all identifiable information anonymized.

In case of face to face interviews, the researcher would arrange the interview in a private confidential place (e.g., a meeting room that is available) so that there are less disturbance and the interviewees will feel free to share their information. In case of the online interviews, the participants will be assisted to conduct the interviews in a private place and wear the headphones where possible to minimize the chances of third-party listening. In case there are technical problems during online interviews (e.g., poor connection), the interview will be stopped and resumed after the connection is stable or the interview can be rescheduled to prevent poor audio quality and discomfort of the participants.

Besides audio recording, the researcher will make brief reflective field notes just after every interview to record the contextual note to the interview (i.e., the points of emphasis, non-verbal communication in face-to-face interactions, and initial impressions of the analysis). Such notes are not going to substitute transcripts, as they will be used to assist in interpreting the meaning later in the thematic analysis.

3.5.4 Data Management, Confidentially and Security

To enhance confidentiality, every participant will receive an individual code (e.g., T1, T2) and it will be applied in transcripts and reporting. All direct identifiers (teacher, school, district, colleague and student names) will be eliminated on the transcription phase and substituted with neutral terms (e.g., school, district). To avoid any correlation with the identity and the answers, consent forms will be not stored with the interview data.

Every digital file (audio records, transcripts, and analytical files) will be stored in password-protected/encrypted storage available to the researcher only. In case of need of backup storage, the same will be kept in equally secured storage and no files will be exchanged via unsecured messaging applications. The retention period of data will be based on the institutional requirements and will be followed up by its secure destruction (e.g., the deletion of any digital records as well as the removal of files out of storage permanently).

3.5.5 Procedures of Transcription and Verification

To ensure the accuracy and richness of the accounts of the participants, audio-recordings will be transcribed word-by-word. The researcher will go through transcripts after transcription as she listens to the recordings and make corrections where necessary and completeness. To increase the credibility, the participants can be requested to check the correctness of their transcripts or a summary of essentials (member checking), especially in cases where the meaning might be delicate or subject to misunderstanding. Any clarification which is given by participants will be recorded and included into the final dataset without bringing any identifiable information.

3.6 DATA ANALYSIS PROCEDURE

The qualitative data gathered from the semi-structured interviews will be analyzed using Thematic Analysis as developed by Braun and Clarke (2006). Finlay (2021) defines thematic analysis as a method for systematically identifying, analyzing, organizing, and offering insight into patterns of meaning, or themes, across a dataset. This analytical method is suitable for this study due to its accessible and flexible nature. It allows the researcher to provide a rich, detailed, and complex account of data without being tethered to a rigid theoretical commitment, making it ideal for exploring the subjective lived experiences of participants. In the specific context of this study, thematic analysis is appropriate because it enables the researcher to move beyond surface-level descriptions of generative AI usage to uncover the underlying cognitive structures, such as specific fears, motivations, and contradictory attitudes that shape how ESL teachers integrate Generative AI into their practice.

This research will make use of thematic analysis as a versatile and iterative analytic approach which is both inductive and theory sensitised. This implies that the conceptualization of the study (TAM and Teacher Cognition) will help to focus on the main areas such as perceived usefulness, perceived ease of use, classroom practice at various phases (planning, instruction, assessment), and situation constraints/enablers will be based on the narratives of the teachers. The hybrid orientation fits in that the study is exploratory (it should be open to emergent themes), but also conceptually focused (it should cover empirically the constructs that provide responses to the research questions in a systematic manner).

3.6.1 Phase 1: Familiarisation with the Data

To ensure rigor and consistency, the analysis will follow the six-phase framework established by Braun and Clarke (2006, as cited in Ahmed et al. 2025). The process begins with familiarization with the data, during which the researcher will immerse in the data by transcribing interviews verbatim, reading and re-reading the transcripts, and noting initial observations.  During this phase, the researcher will also integrate reflective field notes (recorded after interviews) to preserve contextual meanings (e.g., emphasis, pauses, and practical classroom references) that may support later interpretation.

3.6.2 Phase 2: Generating Initial Codes

Following this, the phase of generating initial codes will involve systematically coding interesting features across the dataset. While remaining open to emergent ideas, this coding will be sensitized by the study’s conceptual framework. For example, data will be examined for instances related to Perceived Usefulness, such as time savings and effectiveness, and Perceived Ease of Use, such as technical barriers as well as for actions and decisions relating to planning, interactive teaching, and assessment. To maximize the quality of analysis, a researcher will keep a growing codebook containing (i) the code names, (ii) the operational definitions, (iii) the inclusion and exclusion criteria, and (iv) the sample excerpts. The codes will be used throughout the entire dataset to compare groups of teachers in various school settings (ex: urban vs. rural) as it is applicable in the second research objective of the study.

3.6.3 Phase 3: Developing Themes

Afterwards, the search in terms of the themes will entail collating and clustering codes under possible themes and a thematic map will be created that represent relationships between themes.  At this point, codes will be grouped into candidate themes that describe patterned meaning pertinent to the research questions, i.e. such themes might describe AI as workload relief, assessment integrity anxiety, human-in-the-loop checking, or infrastructure as a gatekeeper, which will depend on the dataset. Thematic mapping will be applied to test the relationship between themes with each other (e.g., how the contextual constraints influence the perceived usefulness, or how the assessment concerns influence the choice of integration).

3.6.4 Phase 4: Reviewing Themes

This results in the checking of themes in which the themes are compared to the coded data and the entire data set to determine whether they create a consistent pattern. At this stage, the themes are refined, subdivided or discarded to arrive at a strong thematic framework that is reflective of the data.  Two levels will be used to carry out the reviewing: (1) coherence within each of the themes (internal homogeneity) and (2) coherence between themes (external heterogeneity). Where overlap is present, themes will be combined or re-defined to enhance precision of analysis and clarity.

3.6.5 Phase 5: The Definition and Naming of Themes

This is followed by defining and naming theme whereby every theme is well defined and labeled to summarize the theme. In this case, the analysis is going to clearly state how each of the themes answers the research questions.  The phase will involve each theme being written in the form of an analytic claim (not topic label). The definitions of the themes will include (i) what is covered by the theme, (ii) what is not covered by the theme, (iii) what the theme explains about the attitudes of the teachers and/or their practices, and (iv) the connection of the topic to TAM (PU/PEOU) and Teacher Cognition (classroom practice/contextual factors) where necessary.

3.6.6 Phase 6: Producing the Report

Lastly, the producing the report stage will entail the choice of vivid and representative excerpts and the formulation of an analytic story that answers the two research questions. The focus in reporting will be on data-driven explanation more than description, demonstrating how teacher perceptions (i.e. usefulness, ease of use, trust, risk) relate to particular practices in integrating across planning, instruction and assessment and how these practices are influenced by realities in contexts (i.e. time, exam pressures, policy guidance and infrastructure). To protect participant anonymity, non-identifiable labels (e.g., T1, T2) and elimination of school- or person-identifying information will be used. The report will also provide comparison across the context of the participants where necessary to show how one can get similar attitudes to lead to different practices under different circumstances (such as high perceived usefulness still leading to low classroom use when the infrastructure or institutional support is weak).

3.7 TRUSTWORTHINESS

In qualitative inquiry, the rigor and credibility of the research are established through the concept of trustworthiness, which ensures findings are credible, authentic, and worth paying attention to (Lim 2025). This study adopts the four criteria proposed by Lincoln and Guba (1994), including credibility, transferability, dependability, and confirmability as a framework to demonstrate methodological rigor.

3.7.1 Credibility

Credibility refers to confidence in the accuracy of the findings and their alignment with participants’ realities (Lim 2025). To ensure credibility, this study will employ member checking (Arslan 2025). Following the initial analysis, a summary of the emergent themes will be shared with participants, who will be invited to confirm, clarify, or correct the interpretations. This iterative process ensures the final analysis remains grounded in the teachers’ lived experiences and minimizes researcher misinterpretation.

3.7.2 Transferability

Qualitative research does not seek statistical generalization but provides sufficient context for others to judge the applicability of findings to their settings, a concept known as transferability (Lim 2025). This study will achieve this through thick description (Arslan 2025). The final report will include rich, detailed accounts of the research context and participant profiles. This allows readers to assess the potential transferability of insights regarding AI integration to other similar ESL contexts.

3.7.3 Dependability

Dependability ensures the research process is logical, traceable, and consistent (Lim 2025). To establish this, the researcher will maintain a rigorous audit trail. This will comprise all raw data, processed data, and a detailed log of analytical decisions made during thematic analysis, including codebook evolution and theme development. This documentation creates a transparent decision trail that allows for the examination of the research process.

3.7.4 Confirmability

Confirmability addresses the neutrality of the findings, ensuring they are shaped by the data rather than researcher bias (Lim 2025). To achieve this, findings will be substantiated through data-driven evidence, with key themes supported by direct, anonymized participant quotes. As stated by Tariq (2025), findings and conclusions extracted are meaningful and accurate when they are grounded in participants’ actual responses and experiences, rather than the researcher’s personal biases or assumptions.

3.8 RESEARCH ETHICS

Ethical considerations are paramount in qualitative research to protect the rights, dignity, autonomy, and well-being of participants. This study will adhere to the highest ethical standards throughout the entire research process, from design and data collection to analysis and dissemination. The ethical framework for this study is guided by the principles of respect for persons, beneficence, and justice.

Firstly, prior to any contact with potential participants, the researcher will obtain formal ethical approval from the relevant bodies. Administrative permission will be sought from the Educational Planning and Research Division (EPRD) of the Ministry of Education Malaysia and the State Education Department of the selected state. No data collection will commence until written approvals from all necessary institutional and ethical review bodies are secured.

Next, the principle of informed consent is central to this study. Potential participants will be provided with a comprehensive “Participant Information Sheet” and a “Consent Form.” The information sheet will detail the study’s purpose, procedures, potential risks, benefits, confidentiality measures, and the voluntary nature of participation. It will also provide the researcher’s and supervisor’s contact information. Participants will be given time to review the documents and ask questions. Written, signed consent will be obtained from each participant prior to scheduling an interview. It will be explicitly stated that participation is entirely voluntary and that they have the right to withdraw at any point before, during, or after the interview without any penalty or negative consequence to their professional standing.

Furthermore, strict measures will be implemented to protect participant privacy and ensure data security throughout the study. To guarantee anonymity, all personally identifiable information, will be removed from the data. Participants will be referred to using non-identifiable pseudonyms in all transcripts, analytical notes, and the final report, and any potentially identifying contextual details will be generalized. Confidentiality will be upheld by treating all research materials, including audio recordings, transcripts, field notes, and signed consent forms with strict discretion. Digital files will be stored exclusively, accessible only to the researcher.

3.9 SUMMARY

This chapter outlines the research methodology employed to explore the attitudes and integration practices of Malaysian secondary ESL teachers regarding Generative AI. Adopting a basic qualitative research design, the research utilizes purposive sampling to recruit educators, continuing until data saturation is reached. Data is collected through validated semi-structured interviews and analyzed using Braun and Clarke’s six-phase thematic analysis framework. Finally, the chapter details the rigorous measures established to ensure trustworthiness, such as member checking and thick description, alongside strict adherence to ethical protocols regarding consent and anonymity.

REFERENCES

Ag-Ahmad, N., Mohamed, A.T.F.S. & Majilang, D.B. 2025. Voices from ESL classrooms: Overcoming challenges and enhancing English language teacher education in Malaysia. Studies in English Language and Education 12(1): 328-345.

Ahmad, A.L. 2025. Navigating challenges and strategies: Malaysian ESL secondary school teachers in differentiated reading instruction. Asian Conference on Education Official Conference Proceedings 1539-1549.

Ahmad, N.S. & Rathakrishnan, M. 2025. Digital technology integration in teaching and learning among teachers in Kedah, Malaysia. International Journal of Instruction, Technology, and Social Sciences 4: 83-94.

Ahmad, S.N., Khairuddin, Z., Shahabani, N.S., Yusof, F.H.M., Zamri, N.A., Ahmad, A.R. & Ibrahim, W.N.A.T. 2025. Students’ perceptions of the use of Artificial Intelligence (AI) in academic writing. Journal of Creative Practices in Language Learning and Teaching (CPLT) 13(3): 114-127.

Ahmed, S.K., Mohammed, R.A., Nashwan, A.J., Ibrahim, R.H., Abdalla, A.Q., Ameen, B.M.M. & Khdhir, R.M. 2025. Using thematic analysis in qualitative research. Journal of Medicine, Surgery, and Public Health 6: 100198.

Aineh, N.M.A. & Ngui, N.W. 2024. Teachers’ and students’ perceptions towards the use of ChatGPT to improve writing in the Malaysian secondary school context. International Journal on E-Learning Practices (IJELP) 7(1): 117-124.

Aldamen, H., Almashour, M., Al-Deaibes, M. & AlSharefeen, R. 2025. Testing Krashen’s input hypothesis with AI: A mixed-methods study on reading input and oral proficiency in EFL. Frontiers in Education 10: 1614680.

Alfarwan, A. 2025. Generative AI use in K-12 education: A systematic review. Frontiers in Education 10: 1647573.

Al-Hudawi, S., Fong, R.L.S., Musah, M. & Tahir, L.M. 2014. The actualization of the Malaysian national education philosophy in secondary schools: Student and teacher perspectives. International Education Studies 7(4).

Alsaweed, W. & Aljebreen, S. 2024. Investigating the accuracy of ChatGPT as a writing error correction tool. International Journal of Computer-Assisted Language Learning and Teaching (IJCALLT) 14(1): 1-18.

Alzubi, A.A.F. & Alelaiwi, A.S. 2025. Exploring EFL university teachers’ perceptions of AI-generative tools in Saudi Arabia: A mixed-methods study. Discover Computing 28(1): 282.

Ansari, M.R. et al. 2022. A study on research design and its types. International Research Journal of Engineering and Technology (IRJET) 9(7): 1132-1135

Arslan, E. 2025. Validity and reliability in qualitative research. Pamukkale University Journal of Social Sciences Institute 51(1): 395-407.

Babinski, L.M., Amendum, S.J., Carrig, M.M., Knotek, S.E., Mann, J.C. & Sánchez, M. 2024. Professional learning for ESL teachers: A randomized controlled trial to examine the impact on instruction, collaboration, and cultural wealth. Education Sciences 14(7): 690.

Bautista, A., Estrada, C., Jaravata, A.M., Mangaser, L.M., Narag, F., Soquila, R. & Asuncion, R.J. 2024. Preservice teachers’ readiness towards integrating AI-based tools in education: A TPACK approach. Educational Process: International Journal 13(3): 40-68.

Bayaga, A. & du Plessis, A. 2024. Ramifications of the Unified Theory of Acceptance and Use of Technology (UTAUT) among developing countries’ higher education staffs. Education and Information Technologies 29(8): 9689-9714.

Borg, S. 2003. Teacher cognition in language teaching: A review of research on what language teachers think, know, believe, and do. Language Teaching 36(2): 81-109.

Braun, V. & Clarke, V. 2006. Using thematic analysis in psychology. Qualitative Research in Psychology 3(2): 77-101.

Cai, L., Msafiri, M.M. & Kangwa, D. 2025. Exploring the impact of integrating AI tools in higher education using the Zone of Proximal Development. Education and Information Technologies 30(6): 7191-7264.

Celik, I. 2023. Towards Intelligent-TPACK: An empirical study on teachers’ professional knowledge to ethically integrate artificial intelligence (AI)-based tools into education. Computers in Human Behavior 138: 107468.

Chan, C.K.Y. 2023. A comprehensive AI policy education framework for university teaching and learning. International Journal of Educational Technology in Higher Education 20(1): 38.

Chan, K.K.W. & Tang, W.K.W. 2025. Evaluating English teachers’ artificial intelligence readiness and training needs with a TPACK-based model. World Journal of English Language 15(1): 129.

Chandran, V.N., Albakri, I.S.M.A., Shukor, S.S., Ismail, N., Tahir, M.H.M., Mokhtar, M.M. & Zulkepli, N. 2022. Malaysian English language novice teachers’ challenges and support during initial years of teaching. Studies in English Language and Education 9(2): 443-461.

Chappell, P., Bodis, A. & Jackson, H. 2015. Impact of teacher cognition and classroom practices on IELTS test preparation courses in the Australian ELICOS sector. IELTS Research Reports Online Series 6.

Clark, C.M. & Peterson, P.L. 1986. Teachers’ thought processes. In Wittrock, M.C. (ed.). Handbook of Research on Teaching. 3rd Ed., pp. 255-296. New York: Macmillan.

Davis, F.D. 1989. Technology acceptance model: TAM. In Al-Suqri, M.N. & Al-Aufi, A.S. (eds.). Information Seeking Behavior and Technology Adoption 205-219.

Dikaprio, V. & Diem, C.D. 2024. How effective is Talkpal.ai in enhancing English proficiency? Insights from an experimental study. Language, Technology, and Social Media 2(1): 48-59.

Dincer, N. & Bal, S. 2024. A qualitative journey on instructors’ perceptions of artificial intelligence in EFL education. In AI in Language Teaching, Learning, and Assessment, pp. 78-100. Hershey: IGI Global Scientific Publishing.

Ding, D. & Yusof, M.B. 2025. Investigating the role of AI-powered conversation bots in enhancing L2 speaking skills and reducing speaking anxiety: A mixed methods study. Humanities and Social Sciences Communications 12(1): 1-16.

Dong, L. 2024. ChatGPT in language writing education: Reflections and a research agenda for a ChatGPT feedback engagement framework. Language Teaching Research Quarterly 43: 121-131.

Dwivedi, Y.K., Rana, N.P., Jeyaraj, A., Clement, M. & Williams, M.D. 2019. Re-examining the unified theory of acceptance and use of technology (UTAUT): Towards a revised theoretical model. Information Systems Frontiers 21(3): 719-734.

Ekizer, F.N. 2025. Exploring the impact of artificial intelligence on English language teaching: A meta-analysis. Acta Psychologica 260: 105649.

Eusebio, E.J.G., Baldera, P., Patiam, A.M.C., Villanueva, E.R., Gaa, N.A., Solis, A.T., … & Ribon, A.L. 2025. AI in the classroom: A systematic review of barriers to educator acceptance. International Journal of Learning, Teaching and Educational Research 24(9): 126-147.

Eyal, L. 2025. Developing and validating an AI-TPACK assessment framework: Enhancing teacher educators’ professional practice through authentic artifacts. Education Sciences 15(11): 1452.

Farazouli, A., Pargman, T.C., Laksov, K.B. & McGrath, C. 2025. Navigating uncertainty: University teachers’ experiences and perceptions of generative artificial intelligence in teaching and learning. Studies in Higher Education: 1-16.

Feng, H., Li, K. & Zhang, L.J. 2025. What does AI bring to second language writing? A systematic review (2014-2024). Language Learning & Technology 29(1): 1-27.

Finlay, L. 2021. Thematic analysis: The ‘good’, the ‘bad’ and the ‘ugly’. European Journal for Qualitative Research in Psychotherapy 11: 103-116.

Firdaus, A. & Nawaz, S. 2024. Viewpoints of teachers about the usage of artificial intelligence in ELT: Advantages and obstacles. University of Chitral Journal of Linguistics and Literature 8(1): 82-93.

Giray, L., Jacob, J., Encanto, V. & Mansilungan, C.J. 2025. Cheating writing with generative AI: Exploring student motivations using the Theory of Planned Behavior. Journal of Academic Ethics 24(1): 1-23.

Granić, A. 2023. Technology acceptance and adoption in education. In Zawacki-Richter, O. & Jung, I. (eds.). Handbook of Open, Distance and Digital Education. Vol. 1, pp. 183-197. Singapore: Springer.

Hassan, W.Z.W. & Aziz, F. 2024. The application of human values through the National Education Philosophy towards the development of a prosperous nation in Malaysia in the era of Industrial Revolution 4.0. International Journal of Academic Research in Business and Social Sciences 14(3).

Hassim, H., Kassim, H., Kassim, A. & Kassim, M. 2023. Exploring the use of artificial intelligence-based technology to enhance creativity in ESL speaking classroom. In 2023 7th IEEE Congress on Information Science and Technology (CiSt) 512-517.

Hennink, M. & Kaiser, B.N. 2022. Sample sizes for saturation in qualitative research: A systematic review of empirical tests. Social Science & Medicine 292: 114523.

Henry, J. 2025. Chat2Comprehend: An AI-based approach to support reading comprehension in primary ESL classrooms. In The Agentive ELT Professional: Driving Innovation and Impact (33rd MELTA International Conference E-Proceedings), pp. 70-74. Kuala Lumpur: Malaysian English Language Teaching Association.

Hidayat, M.T. 2024. Effectiveness of AI-based personalised reading platforms in enhancing reading comprehension. Journal of Learning for Development 11(1): 115-125.

Huang, F., Teo, T. & Zhou, M. 2019. Factors affecting Chinese English as a foreign language teachers’ technology acceptance: A qualitative study. Journal of Educational Computing Research 57(1): 83-105.

Huang, J. & Mizumoto, A. 2025. The effects of generative AI usage in EFL classrooms on the L2 motivational self-system. Education and Information Technologies 30(5): 6435-6454.

Huang, Z., Fu, X. & Zhao, J. 2025. Research on AIGC-integrated design education for sustainable teaching: An empirical analysis based on the TAM and TPACK models. Sustainability 17(12): 5497.

Hutapea, E., Hutabalian, R. & Hartati, R. 2024. Summarizing AI application on student learning efficiency in understanding academic reading materials. Indonesian Journal of Education and Development Research 3(1): 737-745.

Iskandar Malaysia. 2015. Comprehensive Development Plan ii. https://iskandarmalaysia.com.my/comprehensive-development-plan.html.

Ismail, R.A.M. 2025. Bridging the digital divide in Malaysian education. Penang Institute Monographs.

Jackson, P.W. 1968. Life in Classrooms. New York: Holt, Rinehart, & Winston.

Jamaluddin, F., Jamaluddin, A.H., Jamaluddin, F. & Jamaluddin, F. 2025. Malaysia’s AI-driven education landscape: Policies, applications, and comparative insights for a digital future. arXiv preprint arXiv:2509.21858.

Jaramillo, J.J., Chiappe, A. & Delgado, F.S. 2025. From struggle to mastery: AI-powered writing skills in ESL education. Applied Sciences 15(14): 8079.

Jauhiainen, J.S. & Garagorry Guerra, A. 2025. Generative AI in education: ChatGPT-4 in evaluating students’ written responses. Innovations in Education and Teaching International 62(4): 1377-1394.

Javahery, P. & Alizadeh, M.J. 2025. Reimagining Krashen’s input hypothesis: The role of AI and multimodal strategies in language acquisition. Innovations in Pedagogy and Technology 1(3): 82-94.

Jen, S.L. & Salam, A.R. 2025. Teaching secondary school essay writing using generative AI. International Journal of Research and Innovation in Social Science IX(IIIS): 6065-6070.

Kampookaew, P. 2020. Factors influencing Thai EFL teachers’ acceptance of technology: A qualitative approach. ThaiTESOL Journal 33(2): 46-69.

Karaduman, C. 2025. Pre-service EFL teachers’ perceived AI literacy and competency: The integration of ChatGPT into English language teacher education. SAGE Open 15(3): 21582440251379712.

Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., … & Kasneci, G. 2023. ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences 103: 102274.

Ken, T.J., Zaini, M.H. & Nasir, M.K.M. 2025. ESL teachers’ perception of using artificial intelligence in differentiated instruction. International Journal of Academic Research in Business and Social Sciences 15(8).

Kerr, R.C. & Kim, H. 2025. From prompts to plans: A case study of pre-service EFL teachers’ use of generative AI for lesson planning. English Teaching 80(1): 95-118.

Khamis, N.Y.H. & Yusof, N. 2024. Artificial intelligence revolutionising writing assessments. International Journal of Language Education and Applied Linguistics 14(2): 1-6.

Khan, A., Hassan, N. & Cheng, L. 2025. Investigating the contextual factors mediating washback effects of a learning-oriented English language assessment in Malaysia. Language Testing in Asia 15(1): 20.

Kılıçkaya, F. & Kic-Drgas, J. 2025. Pre-service language teachers’ experiences and perceptions of integrating generative AI in practicum-based lesson study. Humanities and Social Sciences Communications 12(1): 1-11.

Kılınç, H.K. & Keçecioğlu, Ö.F. 2024. Generative artificial intelligence: A historical and future perspective. Academic Platform Journal of Engineering and Smart Systems 12(2): 47-58.

Kimotho, S. 2025. Piloting qualitative research: A critical examination of its role and practice. SSRN 5271340.

Kussin, H.J., Khalid, P.Z.M., Chaniago, R.H., Moneyam, S. & Hassim, H.E. 2024. The future of lesson planning: AI integration experiences among TESL teacher trainees in a Malaysian public university. AJELP: Asian Journal of English Language and Pedagogy 12(2): 178-189.

Law, L. 2024. Application of generative artificial intelligence (GenAI) in language teaching and learning: A scoping literature review. Computers and Education Open 6: 100174.

Lee, S., Choe, H., Zou, D. & Jeon, J. 2025. Generative AI (GenAI) in the language classroom: A systematic review. Interactive Learning Environments: 1-25.

Li, H. & Thien, L.M. 2025. Artificial intelligence (AI) anxiety and usage intention in English teaching: Insights from Chinese EFL pre-service teachers. Asian Education and Development Studies 14(5): 1022-1035.

Li, M. & Wilson, J. 2025. AI-integrated scaffolding to enhance agency and creativity in K-12 English language learners: A systematic review. Information 16(7): 519.

Liao, H., Xiao, H. & Hu, B. 2023. Revolutionizing ESL teaching with generative artificial intelligence: Take ChatGPT as an example. International Journal of New Developments in Education 5(20): 39-46.

Lim, W.M. 2024. What is qualitative research? An overview and guidelines. Australasian Marketing Journal (AMJ) 33(2): 199-229.

Lincoln, Y.S. & Guba, E.G. 1994. Competing paradigms in qualitative research. Handbook of Qualitative Research 2 105: 163-194.

Ling, Y. & Jan, J.M. 2025. Voices from the flip: Teacher perspectives on integrating AI chatbots in flipped English classrooms. Education Sciences 15(9): 1219.

Malaysian Ministry of Digital. 2025a. AI For Good (Educator) Conference. https://www.digital.gov.my/en-GB/siaran/AI-For-Good-(Educator)-Conference.

Malaysian Ministry of Digital. 2025b. National Artificial Intelligence (AI) Action Plan. https://cdnc.heyzine.com/files/uploaded/v3/70dc64987eea7443462d36807ac1daa6a0d584c2.pdf.

Malaysian Ministry of Economy Department of Statistics. 2023. Statistics on Household Income & Basic Amenities. https://www.dosm.gov.my/portal-main/release-content/household-income-survey-report–malaysia–states.

Malaysian Ministry of Education. 2016. Buku Penerangan Kurikulum Standard Sekolah Menengah (KSSM). https://anyflip.com/ubvja/apkf/basic.

Malaysian Ministry of Education. 2019. Rethink, Revamp Teaching and Learning of English. https://www.moe.gov.my/rethink-revamp-teaching-and-learning-of-english-new-straits-times-22-mei-2019.

Malaysian Ministry of Education. 2020. Teacher Guide Implementing the CEFR-Aligned Curriculum Planning and Managing Learning. 2nd Ed. https://anyflip.com/ojrv/ngtm/basic.

Malaysian Ministry of Education. 2023a. Digital Education Policy. https://www.moe.gov.my/storage/files/shares/Dasar/Dasar%20Pendidikan%20Digital/Digital%20Education%20Policy.pdf.

Malaysian Ministry of Education. 2023b. Falsafah Pendidikan Kebangsaan. https://www.moe.gov.my/falsafah-pendidikan-kebangsaan.

Malaysian Ministry of Education. 2025. Laporan Analisis Keputusan. https://lp.moe.gov.my/index.php/laporan-analisis-keputusan.

Malaysian Ministry of Education. 2026. Ringkasan Eksekutif Rancangan Pendidikan 2026-2035. https://bit.ly/DokumenRingkasanEksekutifRancanganPendidikanMalaysia2026-2035.

Manaf, M.B.A. & Azizan, M.B. 2025. Technology acceptance and academic willingness in supporting the offering of diploma in applied artificial intelligence at polytechnic Malaysia. In E3S Web of Conferences 664: 01015. EDP Sciences.

Merriam, S.B., Tisdell, E.J. & Stuckey-Peyrot, H.L. 2016. Qualitative Research: A Guide to Design and Implementation. John Wiley & Sons.

Mishra, P. & Koehler, M.J. 2006. Technological pedagogical content knowledge: A framework for teacher knowledge. Teachers College Record 108(6): 1017-1054.

Mizumoto, A., Yasuda, S. & Tamura, Y. 2024. Identifying ChatGPT-generated texts in EFL students’ writing: Through comparative analysis of linguistic fingerprints. Applied Corpus Linguistics 4(3): 100106.

Mokkink, L., Herbelet, S., Tuinman, P. & Terwee, C. 2025. Content validity: Judging the relevance, comprehensiveness and comprehensibility of an outcome measurement instrument-a COSMIN perspective. Journal of Clinical Epidemiology 111879.

Moorhouse, B.L. & Kohnke, L. 2024. The effects of generative AI on initial language teacher education: The perceptions of teacher educators. System 122: 103290.

Muhamad, M., & Seng, G. H. 2022. Issues in the implementation of 21st century learning skills in Malaysian ESL classrooms. Asian Journal of University Education (AJUE), 18(4), 1093-1104.

Muniandy, J. & Selvanathan, M. 2025. ChatGPT, a partnering tool to improve ESL learners’ speaking skills: Case study in a Public University, Malaysia. Teaching Public Administration 43(1): 4-20.

Mutammimah, H., Rejeki, S., Kustini, S. & Amelia, R. 2024. Understanding teachers’ perspective toward ChatGPT acceptance in English language teaching. International Journal of Technology in Education 7(2): 290-307.

Naeem, M., Ozuem, W., Howell, K. & Ranfagni, S. 2024. Demystification and actualisation of data saturation in qualitative research through thematic analysis. International Journal of Qualitative Methods 23: 16094069241229777.

Nasr, N.R., Tu, C.H., Werner, J., Bauer, T., Yen, C.J. & Sujo-Montes, L. 2025. Exploring the impact of generative AI ChatGPT on critical thinking in higher education: Passive AI-directed use or human-AI supported collaboration? Education Sciences 15(9): 1198.

Nasrudin, N.H. & Hashim, H. 2025. Exploring ESL teachers’ perceptions of AI integration: A systematic literature review (2019-2025). International Journal of Research and Innovation in Social Science IX(VII): 6078-6088.

Nawawi, N.M., Zuhaimi, N., Sabu, K., Mahamud, N.S.R. & Nasir, N.A.M. 2021. CEFR for Languages and its effective implementation in secondary schools in Malaysia. Asian Journal of Assessment in Teaching and Learning 11(1): 63-72.

NCER Malaysia. 2020. Northern Corridor Economic Region Strategic Development Plan (2021-2025). https://www.ncer.com.my/sites/default/files/node/report/files/2024-03/NCER_Strategic-Development-Plan_0.pdf.

Nguyen, L.T.H., Dinh, H., Dao, T.B.N. & Tran, N.G. 2025. Teachers’ perceptions and students’ strategies in using AI-mediated informal digital learning for career ESL writing. Education Sciences 15(10): 1414.

Nurjanah, L., Cahyono, B.Y. & Suryati, N. 2025. The successful use of AI for English teachers’ professional development. JALTCALL Trends 1(1): 2169.

Nyimbili, F. & Nyimbili, L. 2024. Types of purposive sampling techniques with their examples and application in qualitative research studies. British Journal of Multidisciplinary and Advanced Studies: English Lang., Teaching, Literature, Linguistics & Communication 5(1): 90-99.

Percy, W.H., Kostere, K. & Kostere, S. 2015. Generic qualitative research in psychology. The Qualitative Report 20(2): 76-85.

Pokrivcakova, S. 2023. Pre-service teachers’ attitudes towards artificial intelligence and its integration into EFL teaching and learning. Journal of Language and Cultural Education 11(3): 100-114.

Qiuyang, H., Wenling, L. & Yanmei, Z. 2025. Enhancing deep learning and motivation in university English education through AI technology: A quasi-experimental study. Asian Journal of Education and Social Studies 51(4): 452-463.

Radi, T.K., Zabit, M. & N., B. 2025. Balancing fun and exam readiness: Teachers’ perspectives on technology and interactive approaches in Malaysian primary English classrooms. International Journal of Advanced Research (IJAR) 13(9): 948-955.

Reddy, P., Ch, K., Sharma, K., Sharma, B. & Sharma, S. 2025. Evolution of generative artificial intelligence: A review of the developed and developing. Engineered Science 35: 1529.

Robert, J.D., Henry Joseph, A. & Apolonius, L.E. 2025. AI-assisted language learning in education: ESL learners’ perceptions and challenges using ChatGPT. Creative Practices in Language Learning and Teaching (CPLT) 13(1): 34-50.

Sadhasivam, S., Michael, M.V.P., Mohamad, M. & Yunus, M.M. 2023. The importance of innovative teaching and learning approaches in the implementation of CEFR: A literature review. International Journal of Academic Research in Progressive Education and Development 12(2): 1696-1705.

Seo, K., Tang, J., Roll, I., Fels, S. & Yoon, D. 2021. The impact of artificial intelligence on learner-instructor interaction in online learning. International Journal of Educational Technology in Higher Education 18(1): 54.

Shah, S.H.R., Altaf, A.R. & Mughal, K.H. 2025. Exploring the impact of AI tools on reading proficiency among undergraduate students. The Critical Review of Social Sciences Studies 3(3): 663-670.

Shakri, M.A.M., Peng, N.G., Saari, K.A. & Marliana, N.L. 2025. The application of generative artificial intelligence technology among language teachers in facing current educational challenges. Journal of ICT in Education 12(2): 138-148.

Şimşek, A.S., Cengiz, G.Ş.T. & Bal, M. 2025. Extending the TAM framework: Exploring learning motivation and agility in educational adoption of generative AI. Education and Information Technologies 30(15): 20913-20942.

Sivanganam, J., Yunus, M.M. & Said, N.E.M. 2025. Teachers’ perceptions in using artificial intelligence (AI) in ESL classrooms. International Journal of Academic Research in Progressive Education and Development 14(1).

Sok, S. & Shin, H.W. 2025. Do interactions with ChatGPT influence L2 learners’ oral speaking ability, summarization ability, and perceptions of generative AI tasks? TESOL Quarterly 59(S1).

Staller, K.M. 2021. Big enough? Sampling in qualitative inquiry. Qualitative Social Work 20(4): 897-904.

Sun, L. & Zhou, L. 2024. Does generative artificial intelligence improve the academic achievement of college students? A meta-analysis. Journal of Educational Computing Research 62(7): 1676-1713.

Syafrayani, P.R., Ayunda, R., Molina, M.L., Hutasuhut, M.L. & Siregar, M. 2024. Teachers’ perspectives on AI integration in EFL teaching: Perceived benefits and challenges. Linguistik Terapan 21(1): 27-40.

Tang, C.M. & Chaw, L.Y. 2023. What have people discussed about ChatGPT in Malaysian education? A qualitative content analysis of news articles. European Conference on e-Learning 22(1): 314-321.

Tang, Y. & Zhong, L. 2026. K-12 teachers’ adoption of generative AI for teaching: An extended TAM perspective. Education Sciences 16(1): 136.

Tariq, M.U. 2025. Ensuring trustworthiness and rigor in qualitative research. In Qualitative Inquiry in Doctoral Research: Pathways to Effective Design and Implementation, pp. 365-392.  IGI Global Scientific Publishing.

Tessensohn, T.C., Yunus, M.M. & Ismail, H.H. 2025. Using AI-powered tools in enhancing reading skills in the ESL classroom: A systematic review (2020-2024). International Journal of Academic Research in Progressive Education and Development 14(2): 57-70.

Thong, S.J. & Kamsin, I.F. 2025. The perception of low-enrolment school teachers about the use of ChatGPT. International Journal of Advanced Research in Education and Society 7(6): 371-385.

Tran, P.T.H., Huynh, L., Bien, T.A., Dang, B. & Nguyen, A. 2025a. Preparing preservice teachers for generative AI in lesson planning: A process mining study of AI mindset and tool-only training. Journal of Digital Learning in Teacher Education: 1-15.

Tran, T.L.N., Van Le, T. & Nguyen, T.L. 2025b. Teachers’ perspectives about the benefits, challenges, and strategies for using generative AI in English language teaching. The JALT CALL Journal 21(3): 102514.

Tripathi, T., Sharma, S.R., Singh, V., Bhargava, P. & Raj, C. 2025. Teaching and learning with AI: A qualitative study on K-12 teachers’ use and engagement with artificial intelligence. Frontiers in Education 10: 1651217.

United Nations. 2016. Goal 4 | Department of Economic and Social Affairs. https://sdgs.un.org/goals/goal4#targets_and_indicators.

Venkatesh, V., Morris, M.G., Davis, G.B. & Davis, F.D. 2003. User acceptance of information technology: Toward a unified view. MIS Quarterly: 425-478.

Vera, F.E.R.N.A.N.D.O. 2023. Enhancing English language learning in undergraduate students using ChatGPT: A quasi-experimental study. In Congreso Internacional de Aprendizaje Activo, 18-21.

Vygotsky, L. 1978. Mind in Society: Development of Higher Psychological Processes. Harvard University Press.

Wang, Y.Y. & Zhang, L.J. 2024. Charting the trajectory of language teacher cognition development: What 15 years of research in System informs us. System 127: 103508.

Wang, Y., Zhang, T., Yao, L. & Seedhouse, P. 2025. A scoping review of empirical studies on generative artificial intelligence in language education. Innovation in Language Learning and Teaching: 1-28.

Wei, L. 2023. Artificial intelligence in language instruction: Impact on English learning achievement, L2 motivation, and self-regulated learning. Frontiers in Psychology 14: 1261955.

Xiaofan, W. & Annamalai, N. 2025. Investigating the use of AI tools in English language learning: A phenomenological approach. Contemporary Educational Technology 17(2): 578.

Xuan, S.Y. & Yunus, M.M. 2023. Teachers’ attitude towards the use of artificial intelligence-based English language learning: A mini review. International Journal of Academic Research in Business & Social Sciences 13(5): 793-800.

Xue, L., Ghazali, N. & Mahat, J. 2025. Artificial Intelligence (AI) adoption among teachers: A systematic review and agenda for future research. International Journal of Technology in Education 8(3): 802-824.

Yıldız, S. & Erçetin, G. 2025. Teacher’s cognition. In The Palgrave Encyclopedia of Computer-Assisted Language Learning, pp. 1-8. Cham: Springer Nature Switzerland.

Yugandhar, K. & Rao, Y.R. 2024. Artificial intelligence in classroom management: Improving instructional quality of English class with AI tools. Educational Administration: Theory and Practice 30(4): 2666-2672.

Zaimoğlu, S. & Dağtaş, A. 2025. Teacher cognition and practices in using generative AI tools to support student engagement in EFL higher-education contexts. Behavioral Sciences 15(9): 1202.

Zainuddin, N.M. & Mohamad, M. 2024. Utilising Speechace to enhance speaking skills among English as a second language pre-university students. International Journal of Academic Research in Progressive Education and Development 13(2): 1206-1219.

Zeng, M., Cheah, K.S. & Abdullah, Z. 2025. The influence of school principals’ digital leadership on teachers’ competency in integrating artificial intelligence: A systematic thematic review. Frontiers in Education 10: 1655967.

Zhang, C. & Kang, S. 2022. A comparative study on lexical and syntactic features of ESL versus EFL learners’ writing. Frontiers in Psychology 13: 1002090.

Zhou, J. & Liu, D. 2026. From perceptions to beliefs: The mediating role of attitudes in pre-service English teacher cognition of Communicative Language Teaching in China. Frontiers in Psychology 17.

Zulkarnain, N.S. & Yunus, M.M. 2023. Teachers’ perceptions and continuance usage intention of artificial intelligence technology in TESL. International Journal of Multidisciplinary Research and Analysis 6(5): 2101-2109.

Answer

Request Answer of this Assignment