Generative AI as a Pedagogical Co-Pilot in Higher Education: A Review of Adaptive Learning and Teacher Support Systems
Journal of Technological Pedagogy and Educational Development, vol. 2, no. 1, pp. 28-42, 2025, https://doi.org/10.59247/jtped.v2i1.8
Abstract
Generative Artificial Intelligence (GenAI) has emerged as a transformative force in higher education, offering dynamic support for adaptive learning and instructional practices. This review explores the integration of GenAI in pedagogical environments, focusing on its dual function as a facilitator of personalized learning and a support system for educators. As its adoption expands, concerns arise regarding algorithmic bias, data privacy, overdependence, and the erosion of critical thinking. This paper synthesizes findings from a range of recent studies to assess GenAI's capabilities, limitations, and ethical implications. Employing a thematic literature analysis, the review categorizes GenAI applications into adaptive learning systems, teacher support tools, and student engagement platforms. The results indicate that GenAI enhances individualized instruction, reduces administrative burden, and fosters student interaction through co-creation and chatbot-assisted learning. However, challenges such as content hallucination, contextual insensitivity, academic integrity concerns, and equity gaps persist. Comparative analysis of general-purpose and education-specific GenAI tools further highlights trade-offs between adaptability, accessibility, and user control. The study concludes with recommendations for developing teacher-AI collaboration frameworks, improving AI literacy, and establishing inclusive and ethical governance models to ensure responsible GenAI integration in education. This review contributes to the field by providing a structured taxonomy of GenAI tools, a comparative analysis of their pedagogical functions, and policy-oriented recommendations for ethical and inclusive integration in higher education.
Keywords
Generative AI; Adaptive Learning; Teacher Support; Pedagogical Technology; AI Ethics
1. Introduction
In recent years, Generative Artificial Intelligence (Generative AI or GenAI) has been widely adopted in various sectors, including the field of education [1]. GenAI refers to an artificial intelligence system that is capable of automatically generating new content, such as text, images, audio, and video, based on complex data processing [2]. In the context of higher education, this technology has been used to compile teaching materials tailored to individual needs, provide real-time feedback, and automate various administrative tasks [3]. GenAI's potential in improving operational efficiency and learning quality has begun to be recognized by educational institutions in various countries [3].
As its application becomes more widespread, normative and ethical challenges begin to be identified in various studies. Risks related to algorithmic bias, data privacy violations, and the legitimacy of automatically generated content are a major concern in policy development [4], [5]. In addition, changes in the role of educators due to increased reliance on automation systems are considered to affect the quality of pedagogical interaction and reflective processes in learning [6]. The absence of an adequate regulatory framework is also referred to as a structural obstacle that needs to be responded to immediately. GenAI's implementation strategy in education needs to be directed to strengthen the learning ecosystem holistically, by emphasizing the principles of responsibility, inclusivity, and sustainability [7], [8].
The development of Artificial Intelligence in education has undergone a significant paradigm shift, from being a traditional tool to a more complex role as an intelligent co-pilot in the learning process. In the early stages, the application of AI in the classroom is generally focused on limited functions such as automated assessment, content management, and student performance tracking [9] [10], [11]. Advances in generative technology and more adaptive machine learning, the role of AI has evolved into a collaborative partner that is able to respond to learning needs in contextual and real-time. Functions such as personalized feedback, data-driven learning material development, and pedagogical decision-making support can now be dynamically executed by AI systems [12], [13], [14]. This transformation shows that AI is no longer positioned as a passive tool, but rather as an active supporting entity in improving the quality and effectiveness of the educational process.
As the adoption of GenAI increases in the education sector, the need for a structured review of the role of this technology in the teaching and learning process is becoming more urgent. Various studies have been conducted separately to evaluate the benefits and challenges of GenAI, but until now there have been no comprehensive studies that systematically review the impact of GenAI on the role of educators and the learning dynamics of students. Without an integrated understanding, it is difficult to formulate appropriate policies, ethical guidelines, and implementation strategies. Therefore, a review approach is needed that not only examines GenAI's potential in improving learning efficiency and personalization, but also considers its pedagogical, ethical, and institutional implications as a whole.
While prior reviews have explored GenAI in education, few have systematically examined its dual function as both a pedagogical co-pilot and teacher support system in higher education settings. This gap necessitates a focused review to assess both its transformative potential and the pedagogical risks that may arise in implementation.
This article was compiled to comprehensively examine the role of GenAI in supporting the teaching and learning process in higher education environments. In recent years, GenAI has come under widespread concern due to its potential to transform educational practices [3]. Therefore, the main focus is given to the analysis of GenAI's contribution as a teacher support systems and adaptive learning facilitator, as well as on the identification of ethical, pedagogical, and institutional challenges that arise along with its implementation. GenAI is seen not only as an aid, but as a collaborative entity that is able to interact with learning needs dynamically and contextually.
The review is guided by a number of key questions designed to guide the process of collecting and analyzing the literature. These questions include:
1) How GenAI is applied in supporting the teaching-learning process in higher education;2) The extent to which this technology is able to strengthen the role of educators without reducing the dimension of human interaction in learning;
3) How the personalization mechanism of teaching materials is facilitated by GenAI; and
4) What challenges are faced, both in terms of algorithmic bias, data protection, and institutional unpreparedness in adopting this technology in a sustainable manner.
These questions are positioned as a conceptual framework for evaluating the findings presented in this article.
This article contributes by filling a critical gap in the literature through the development of a pedagogically grounded taxonomy of GenAI tools, comparative evaluation of educational versus general-purpose applications, and the formulation of strategic insights for higher education stakeholders regarding ethical and inclusive AI integration.
The structure of this article is divided into several main sections to answer the focus of the study systematically. The first part presents a conceptual background describing the transformation of AI in education, including the transition from traditional AI systems to GenAI's role as a pedagogical co-pilot. The second part describes the literature review methodology used, including study selection criteria, thematic analysis approaches, and study limitations. The third section outlines key findings grouped into two dimensions, namely support for educators and facilitation of personalized learning. Furthermore, the fourth part discusses critical issues such as ethics, data security, regulatory challenges, and institutional readiness to integrate GenAI into the education ecosystem. The article concludes with a concluding section that summarizes the conceptual contributions of this study, and offers practical recommendations for policy development, learning design, and future research directions.
2. Method
This review adopted a qualitative thematic analysis approach to synthesize recent developments and perspectives on the use of Generative Artificial Intelligence (GenAI) in higher education. The method involved three key stages: literature identification, selection based on eligibility criteria, and thematic categorization aligned with the study’s guiding questions.
A comprehensive literature search was conducted across Scopus, ScienceDirect, and Google Scholar, chosen due to their extensive coverage of peer-reviewed research in educational technology and AI applications. The search queries utilized Boolean combinations of keywords such as “Generative AI in education,” “adaptive learning,” “teacher support systems,” “AI co-pilot,” “pedagogical technology,” and “AI ethics in higher education.” The initial search yielded 158 studies.
The inclusion criteria were as follows: (1) articles published between 2019 and 2025, (2) peer-reviewed journal articles, conference proceedings, or book chapters, (3) studies explicitly examining GenAI applications within educational contexts, especially in higher education. The exclusion criteria included: (1) non-English publications, (2) editorials or opinion pieces without empirical or theoretical grounding, (3) studies focused solely on traditional AI without a generative component, and (4) duplicate or inaccessible full-texts.
To ensure rigor and transparency, two authors independently screened the titles and abstracts of the retrieved articles. Eligible full texts were then assessed for thematic relevance. Discrepancies were resolved through collaborative discussion to minimize bias and enhance analytical reliability. Each study was then coded thematically based on recurring patterns related to GenAI applications in adaptive learning, teacher support systems, student engagement, and pedagogical implications.
The review process adhered to a semi-systematic strategy, allowing flexibility for thematic exploration while maintaining clear documentation of the review pathway. Although not fully PRISMA-compliant, the review methodology emphasized replicability and transparency in source selection, theme development, and synthesis of findings.
A structured review process was employed to ensure transparency and replicability. Records were retrieved from Scopus, ScienceDirect, and Google Scholar, then screened based on predefined inclusion and exclusion criteria. Fig. 1 presents the PRISMA-based flowchart detailing the number of records identified, screened, excluded, and those that met the eligibility criteria for final inclusion.
Fig. 1. PRISMA Flow Diagram of Study Selection Process
3. Theoretical Foundations and Background
GenAI's capabilities differ from conventional AI systems, as they are adaptive, interactive, and data-driven at scale [15], [16]. The development of GenAI began with the use of a rigid and limited rule-based system [17]. The next evolution was marked by the presence of more flexible machine learning models, reaching the stage of Large Language Models (LLMs), such as GPT, which are able to understand and generate natural languages in a complex manner [18]. These advances are driven by the availability of large amounts of data as well as increased computing power, thus enabling the application of GenAI in more sophisticated and realistic learning scenarios.
The application of GenAI in education is based on pedagogical principles such as constructivism and adaptive learning. The constructivist model emphasizes that knowledge is actively constructed by learners, and GenAI plays a role in providing relevant stimuli and responses to the process [19]. GenAI's adaptive capabilities allow personalized learning based on individual achievement and speed, which was previously difficult to realize at scale [20].
Along with this technological transformation, the role of educators has also undergone fundamental changes. Teachers or lecturers no longer function solely as material presenters, but instead turn into learning facilitators, critical thinking competency development coaches, and partners in designing meaningful learning experiences [21], [22]. GenAI integration requires educators to build synergy between human interaction and artificial intelligence in an ethical, directed, and oriented manner towards strengthening the learning process.
4. A Taxonomy of Generative AI Applications in Education
4.1. Adaptive Learning Systems
One of GenAI's key contributions to education lies in its ability to deliver material in a personalized manner. Based on the analysis of individual learning data, this system can adjust the content, difficulty level, and order of presentation of material according to the characteristics and needs of each student [23], [24]. This approach not only increases the relevance and efficiency of the learning process, but also encourages the optimization of individual potential through contextually tailored learning experiences. Fig. 2 shows that the process in Adaptive Learning Platforms that is the foundation for the adaptive use of GenAI involves three main components that are integrated with each other, namely the learner model, the domain model, and the adaptation model [25]. The learner model collects and updates data on learner characteristics, the model domain organizes the content systematically, and the adaptation model adjusts the delivery of material based on data from the learner model. The integration of these three components is the main foundation for the creation of a personalized, dynamic, and student-centered learning system.
GenAI's ability to provide real-time feedback also strengthens the implementation of formative assessments [26]. Through automatic detection of errors or understanding gaps, learning interventions can be delivered directly and specifically [27]. This supports the involvement of students in the process of continuous learning reflection [28]. On the other hand, AI-based scaffolding strategies allow learning support to be delivered gradually and responsively, with the level of assistance adjusted based on the development of student performance [29]. As comprehension levels increase, support is systematically reduced to encourage independent learning. These approaches show how GenAI can be integrated as a pedagogical instrument that is adaptive and responsive to learning dynamics.
Fig. 2. Schematic diagram representing the overall structure of an ALP
4.2. Teacher Support Systems
The use of GenAI as a teacher support system has brought significant changes in instructional management. One of the applications can be found in automatic learning planning, where the design of learning activities, time allocation, and selection of teaching materials is prepared quickly based on predetermined pedagogical parameters. This process allows for efficiency in the preparation of learning tools, while also providing space for educators to focus on strategic aspects of teaching [30].
GenAI's ability to speed up the assessment and reporting process also makes a tangible contribution to reducing administrative burdens [31]. Quantitative and qualitative assessments can be carried out with precision through automatic analysis of student learning outcomes, which can then be used as the basis for the preparation of learning progress reports [32]. Various curriculum development and material differentiation features have been provided to assist in the preparation of content that is responsive to the diversity of learners' characteristics [23]. The existence of these features makes GenAI a pedagogical tool that strengthens the role of teachers in designing adaptive, measurable, and inclusive learning experiences.
4.3. Student Engagement Tools
The involvement of students in the learning process is a key element in improving the quality of education. GenAI has been leveraged to develop a variety of tools specifically designed to enhance students' active participation and cognitive engagement. One approach that is starting to be widely applied is the use of AI-based collaborative platforms that allow students to play the role of co-creators in activities such as narrative writing, visual content creation, and digital project development [33]. This kind of approach encourages the formation of a constructive learning environment, where individual creativity and expression are given space to develop autonomously.
Intelligent chatbot integration has also been used to support exploratory and inquiry-based learning [34]. This system is designed to be able to respond to questions, provide instructions, and facilitate the process of self-discovery of concepts by students. The interaction that is built is dialogical and oriented towards the development of critical thinking, thus encouraging students to not only receive information, but also to evaluate, examine, and actively build understanding. GenAI shows great potential in supporting deeper and meaningful engagement in the learning process.
5. Survey of Tools and Models
5.1. General-Purpose Models
The development LLMs has become the foundation of various GenAI applications used in education. These models are categorized as general-purpose models because they are designed to perform a wide range of tasks across domains, including natural language processing, semantic reasoning, and content generation in various forms [35]. Some of the key models that stand out in this category include GPT-4 (OpenAI) [36], Claude (Anthropic) [37], Gemini (Google DeepMind) [38], and LLaMA (Meta AI) [39]. Each is developed with advanced architecture and trained using data at massive scale to optimize linguistic performance and contextual adaptability.
The advantages of these models lie in their ability to flexibly adjust the output based on user prompts, support context understanding, and respond with a high level of cohesion and relevance. The model has been used to support various activities such as academic writing, explanation of concepts, conversation simulations, and the preparation of adaptive teaching materials. Because they are generalist, these models also allow the development of learning systems that can be adjusted across subjects and educational levels. The presence of general-purpose models is considered the main technological foundation in the use of GenAI in the higher education environment.
5.2. Education-Specific Applications
As the adoption of GenAI in education increases, a number of applications specifically designed for learning purposes have been developed to address more specific pedagogical needs. In contrast to the general model that is cross-domain, this application is categorized as education-specific GenAI tools because it is built with pedagogical principles, curriculum, and classroom dynamics in mind.
One example of an innovative application is Khanmigo, a GenAI-based virtual tutor developed by Khan Academy to assist students in understanding subject matter personally [40]. Curipod is used to instantly create interactive presentations and classroom activities based on student participation [41], while Diffit is designed to simplify and adjust reading or lesson texts according to the level of literacy ability of students [42]. The MagicSchool app gives teachers access to generate high-efficiency worksheets, practice questions, and lesson plans [43], while GRAIDE is used to help with the essay grading process and provide automated feedback on student writing [44].
Compared to education-specific GenAI tools, general-purpose models such as ChatGPT, Gemini, or Claude offer a broader range of capabilities across multiple domains, with high adaptability and versatility in prompt-based content generation. However, they often lack pedagogical alignment with specific curricula and require more contextual framing by educators. In contrast, education-specific tools like Khanmigo, MagicSchool, and Diffit are intentionally designed to meet instructional objectives, incorporate curriculum standards, and integrate classroom analytics. While these tools offer more targeted support for educators, they tend to be limited in cross-domain functionality and are often constrained by platform ecosystems or subscription models. Therefore, selecting between these tools depends heavily on instructional goals, user expertise, and institutional readiness for integration.
5.3. Comparative Analysis
Comparative analysis of various applications of GenAI in education needs to be carried out to understand the differences in characteristics, effectiveness, and implementation challenges in various learning contexts. Each application exhibits unique strengths and limitations, depending on the design of the technology, the purpose for which it is used, as well as the extent to which pedagogical integration has been designed from the beginning. The evaluation is conducted based on a number of key parameters, namely the level of adaptability, user accessibility, teacher involvement, and the extent to which control is given to the end user.
In terms of adaptability, apps like Diffit and Khanmigo show a high ability to adapt learning content to individual needs, especially in supporting instructional differentiation. However, the degree of adaptation often depends on the initial input provided by the teacher or student, as well as the accuracy of the system's processing of the context. Accessibility is also a determining factor, with platforms like Curipod and MagicSchool tending to be more accessible because they have interfaces designed to be used quickly by educators, although connectivity and device availability are still constraints in some regions.
Teacher involvement is an important element in keeping GenAI integration in line with learning goals. Apps that provide collaborative spaces, such as MagicSchool and GRAIDE, allow educators to stay on top of the learning process, even if they are aided by automated systems. Meanwhile, user control over the learning process, both from the student and teacher sides, is an important parameter that distinguishes between a transactional system and a system that supports reflective decision-making. The results of this analysis show that no single application is completely superior in all aspects, so the selection and implementation of GenAI needs to consider the context of use carefully and based on clear instructional objectives.
To provide a clearer understanding of the current implementation of Generative AI in the context of education. Table 1 presents a comparative overview of various GenAI tools that are currently widely used and that are specifically developed for learning and teaching purposes. Table 1 summarizes key features, target users, and limitations of various GenAI tools currently used in educational contexts. For example, ChatGPT and Gemini represent general-purpose platforms with extensive flexibility but face challenges in curriculum alignment. Meanwhile, tools like Khanmigo and MagicSchool demonstrate stronger classroom integration with features such as personalized feedback, worksheet generation, and lesson planning tailored to teacher needs. However, these specialized platforms often come with limited scalability or accessibility for institutions with lower digital capacity.
| Tool / Platform | Developer / Provider | Type | Main Educational Use Cases | Target Users | Key Features | Limitations |
|---|---|---|---|---|---|---|
| ChatGPT (GPT-4) | OpenAI | General-purpose LLM | Tutoring, essay writing, Q&A, code explanations | Students, Teachers |
Conversational AI, context-aware responses, multilingual | May hallucinate, lacks curriculum alignment |
| Khanmigo | Khan Academy (built on GPT-4) |
Domain-specific LLM | Socratic tutoring, feedback on math & science problems | Students, Teachers |
Curriculum-aligned feedback, guided prompts, classroom analytics | Limited to Khan Academy ecosystem |
| Curipod | Curipod AS | Presentation & lesson builder | Interactive slide generation, polls, formative questions | Teachers | AI-assisted lesson planning, visual storytelling | Content quality varies; limited customization |
| Diffit | Diffit.me | Content differentiation tool | Automatic reading passage simplification, quiz generation | Teachers | Level-based output (Lexile), multilingual support | Best for language arts, not STEM content |
| MagicSchool AI | MagicSchool.ai | Educator-focused assistant | Lesson plans, IEP drafts, parent emails, rubric creation | Teachers, Admins |
50+ AI templates for common tasks | Requires verification, subscription tiers |
| GRAIDE | University of Bath (Research tool) | Grading assistant | Automated feedback on math and logic-based answers | Teachers | Detailed grading rubric support, error identification | Limited to specific subject domains |
| Google Gemini (in Workspace) | General-purpose + integrated | Brainstorming, writing assistance, document feedback | Students, Staff |
Integration with Docs, Sheets, Slides | Requires Workspace access, data privacy issues | |
| Copilot for Microsoft 365 | Microsoft | Integrated LLM assistant | Drafting feedback, lesson formatting, presentation generation | Teachers, Staff |
Embedded in MS Word/PowerPoint | Subscription-based, institutional licensing |
In a real-world classroom scenario, a mathematics teacher employing Khanmigo can use the tool to facilitate one-on-one tutoring for students struggling with algebraic equations. The system provides Socratic-style guidance, offering hints rather than direct answers, thereby encouraging students to engage in reflective problem-solving. Simultaneously, the teacher receives real-time analytics on student progress, enabling targeted intervention and differentiated instruction without disrupting the flow of the lesson.
6. Pedagogical and Ethical Implications
6.1. Changing Teacher-Student Dynamics
The application of GenAI in learning has significantly affected the dynamics of the relationship between educators and learners. This change is characterized by a shift from a one-way interaction model, in which teachers are the main source of information, to a more collaborative and dialogical pattern of relationships [45]. In an environment supported by smart technology, learners are given greater space to explore knowledge independently, while educators play the role of facilitators, metacognitive trainers, and reflective partners in the learning process [46].
With the presence of the GenAI system that is able to provide explanations, feedback, and guidance automatically, some of the teacher's instructional functions have been divided into machines. However, the role of teachers is not replaced, but transformed. Interactions that were previously focused on delivering material are now diverted to strengthening understanding, managing discussions, and fostering critical values and ethics in learning. These changes demand new skills from educators, including the ability to strategically integrate GenAI, interpret the results the system provides, and maintain a balance between digital interaction and human relationships in the learning process [47]. This new dynamic creates space for a more balanced pedagogical partnership, in which learning agents of both humans and machines contribute to supporting educational goals holistically.
6.2. Academic Integrity Concerns and Plagiarism Detection
The application of GenAI in higher education has raised growing concerns regarding academic integrity. The ability of generative systems to automatically generate text, essays, quiz answers, and projects poses new challenges in distinguishing students' original work from content that is assisted or fully generated by machines. These concerns include the potential for increased plagiarism practices, a decline in academic originality, as well as a loss of authenticity in the process of scientific thinking and reasoning [48], [49].
In response to these problems, a number of technology-based approaches have been developed to detect plagiarism and irregularities in the use of AI. Detection systems such as Turnitin, GPTZero, and other authenticity verification tools have been widely used to analyze sentence structure, syntactic patterns, and the probability of text authenticity [50]. The effectiveness of these tools is still a matter of debate, especially as advances in advanced generative models such as GPT-4 and Claude are increasingly capable of generating texts with linguistic characteristics that resemble human writing [51].
This problem shows that technology-based supervision needs to be balanced with strong institutional policies and strengthening academic ethical literacy. It should be emphasized that the inculcation of the value of intellectual honesty and academic responsibility cannot be completely replaced by automated detection systems. Higher education needs to develop a dual strategy, namely through increasing ethical awareness among students and educators, as well as updating evaluation policies that take into account new challenges in the GenAI era.
6.3. Limitations: Hallucinations, Misinformation, and Context Sensitivity
Although GenAI offers a variety of advantages in supporting the learning and teaching process, a number of technical and epistemological limitations still need to be critically addressed. One of the main issues is the tendency of generative models to generate hallucinations, i.e., information that sounds convincing but is factually inaccurate or even completely false [52]. This phenomenon occurs because the model does not really "understand" the truth, but rather only predicts the next word or phrase based on statistical patterns in the training data. In the field of education, this can cause misinformation that confuses students and has the potential to weaken the validity of learning materials.
Besides generating hallucinated content, another challenge is the model's limited ability to accurately interpret contextual nuances. GenAI models often have difficulty distinguishing different nuances of meaning, cultural interpretations, or conceptual frameworks, especially when questions are complex or require cross-disciplinary understanding [53], [54]. As a result, the responses given may appear superficially relevant but not in accordance with the pedagogical intent or scientific principles that are expected. This is a serious challenge in its application to ethical, historical, or philosophical materials, where interpretive precision is very important.
Therefore, the use of GenAI in education cannot be separated from the function of human supervision and verification. Active involvement of educators is needed to review and filter the results produced by the system, as well as guide students in evaluating the credibility of information [55], [6]. This approach is important to ensure that technology is used wisely and does not replace critical thinking processes, but rather supports its development.
6.4. Overdependence and the Erosion of Critical Thinking
The widespread use of GenAI in learning activities has raised concerns about the potential for overdependence on this technology. When learners rely too often on automated systems to formulate answers, formulate arguments, or complete academic assignments, the risk of decreased critical thinking capacity and intellectual independence becomes significant [56], [57]. Instead of developing the ability to reflectively analyze, evaluate, and build knowledge, students may be encouraged to passively accept GenAI's output results without questioning the validity or underlying logic.
This situation is compounded by the ease of access and speed of response offered by GenAI, which often replaces the learning process that requires perseverance and deep thinking. In the long run, these kinds of habits have the potential to hinder the formation of the high-level cognitive skills that should be the primary goal of higher education [58]. In addition, the role of educators can also undergo a detrimental shift if the learning system relies too much on automated tools without a balance between human intervention and artificial intelligence [59].
A pedagogical approach oriented towards strengthening learning autonomy and analytical reflection needs to be put forward to prevent the erosion of critical thinking. GenAI should be positioned as a supporting tool, not a substitute for the learning process itself. Learning strategies that encourage open-ended questions, argument-based discussions, and verification of information sources must still be maintained as the core of quality education. The balance between the ease of technology and the formation of thinking character is key in ensuring that GenAI really plays a role in educating, not just making it easier.
6.5. Bias, Equity, and Digital Divide Issues in GenAI Deployment
The issues of algorithmic bias, equal access, and digital divide are structural challenges in the application of GenAI in the education sector [60], [61]. GenAI's models are built on large amounts of training data sourced from the internet and public repositories, which do not always reflect proportionate diversity of cultures, languages, or pedagogical perspectives. As a result, the inherent bias in training data can be replicated and amplified in the outputs generated by the system, thus risking marginalizing certain groups, especially learners from minority backgrounds or regions with low representation in global data.
In addition to the issue of bias, the problem of inequality of access to GenAI technology is also a major concern [62]. In many regions, particularly in developing countries or resource-constrained institutions, the digital infrastructure required to operate advanced AI systems is not yet evenly available. This widens the digital divide between institutions that are able to fully adopt GenAI and those that are not. This inequality has a direct impact on equitable and inclusive learning opportunities, while potentially creating new forms of technology-based education exclusion.
Efforts to ensure equal access, algorithm inclusivity, and fairness in the use of technology are critical to this phenomenon. Evaluation of training datasets, involvement of diverse stakeholders in system design, and strengthening inclusive policies need to be prioritized in the development and implementation of GenAI in the world of education. Thus, technology can be adopted ethically and responsibly, without deepening pre-existing inequalities.
7. Future Research Directions
Future research directions in the use of GenAI for education need to be focused on developing a collaborative framework between educators and artificial intelligence systems. So far, the role of AI has tended to be functional and technical, while the pedagogical involvement of educators has not been fully strategically integrated. Therefore, a collaborative framework is needed that allows for the role of teachers as the main drivers in the use of GenAI, while clarifying the boundaries between human intervention and machine-based automation. Research in this area can include designing human-AI cooperation models that are oriented towards value-based learning, shared decision-making, and improving the quality of learning interactions.
In addition, it is important to conduct research that focuses on developing AI literacy for educators and learners, including skills in prompt engineering as a form of effective communication with GenAI systems. This capability is not only relevant for optimizing the use of technology, but also for building critical awareness of the processes and limitations of AI systems. Longitudinal research is also needed to evaluate the impact of GenAI use on long-term learning outcomes, 21st century skill development, and changes in learners' learning behavior holistically.
Policy aspects and ethical governance are important agendas in the direction of future research. The formulation of a regulatory model that is able to ensure data security, access fairness, and algorithmic transparency needs to be comprehensively studied so that the application of GenAI does not conflict with the principles of humanistic and inclusive education. Finally, the potential for GenAI's integration with new technologies such as augmented reality (AR), virtual reality (VR), and learning analytics also needs to be further explored to create an adaptive and data-driven multimodal learning environment. Interdisciplinary research in this area will greatly determine the future direction of transformative and sustainable digital education.
8. Conclusions
This review underscores the emerging role of Generative AI as a pedagogical co-pilot in higher education. GenAI demonstrates significant potential in enhancing adaptive learning environments, automating instructional tasks, and supporting reflective engagement among students. However, its deployment must be guided by ethical considerations, robust governance structures, and continuous human oversight. Institutions are encouraged to invest in AI literacy training for educators and administrators, develop policies for responsible GenAI use, and implement clear guidelines to support ethical classroom integration. Providing infrastructure and equitable access is essential, especially in resource-constrained environments.
This review is limited by its reliance on secondary literature and the lack of empirical validation across diverse educational settings. The thematic analysis did not include quantitative meta-analysis or region-specific policy impact, which future studies could explore. Future research should focus on classroom-based empirical studies, the design of collaborative human-AI frameworks, and the long-term influence of GenAI on cognitive development, academic integrity, and equity in education. Integrating GenAI with technologies such as AR/VR and learning analytics also presents opportunities for more personalized and immersive educational experiences.
Declaration
Author Contribution: All authors contributed equally to the main contributor to this paper. All authors read and approved the final paper.
Funding: This research received no external funding
Acknowledgment: The authors would like to express their sincere appreciation to the academic and technical teams who provided valuable insights and support during the development of this review. Special thanks are extended to the editorial board of the Journal of Technological Pedagogy and Educational Development for their guidance and constructive feedback. The authors also acknowledge the contributions of colleagues and peer reviewers whose critical input helped improve the quality and clarity of this article.
Conflicts of Interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References
[1] N. Gupta et al., “Exploring prospects, hurdles, and road ahead for generative artificial intelligence in orthopedic education and training,” BMC Med Educ, vol. 24, no. 1, p. 1544, Dec. 2024, https://doi.org/10.1186/s12909-024-06592-8.
[2] M. L. tala, C. N. Muller, I. A. Nastase, O. State, and G. Gheorghe, “Exploring University Students'; Perceptions of Generative Artificial Intelligence in Education,” Amfiteatru Economic, vol. 26, no. 65, p. 71, Feb. 2024, https://doi.org/10.24818/EA/2024/65/71.
[3] A. Gupta, “Impact of Generative AI in Transforming Higher Education Pedagogy,” Leveraging ChatGPT and Artificial Intelligence for Effective Customer Engagement, 2024, pp. 285–300, https://doi.org/10.4018/979-8-3693-0815-8.ch017.
[4] B. Sam Paul and A. Anuradha, “Artificial Intelligence in Different Business Domains,” 2024, pp. 13–33, https://doi.org/10.4018/979-8-3693-1565-1.ch002.
[5] J. Gomathi Sankar and A. David, Eds., Generative AI and Implications for Ethics, Security, and Data Management., IGI Global Scientific Publishing, 2024, https://doi.org/10.4018/979-8-3693-8557-9.
[6] Q. Tang, W. Deng, Y. Huang, S. Wang, and H. Zhang, “Can Generative Artificial Intelligence be a Good Teaching Assistant?—An Empirical Analysis Based on Generative AI‐Assisted Teaching,” J Comput Assist Learn, vol. 41, no. 3, Jun. 2025, https://doi.org/10.1111/jcal.70027.
[7] F. Yuan and J. Li, “Opportunities, Challenges, and Future Directions for Generative Artificial Intelligence in Library Information Literacy Education: A Scoping Review,” Journal of Library and Information Science in Agriculture, vol. 36, no. 9, pp. 44–57, 2024, http://journals.caass.org.cn/nytsqb/EN/10.13998/j.cnki.issn1002-1248.24-0614.
[8] FX. R. Baskara et al., “Redefining educational paradigms: Integrating generative AI into society 5.0 for sustainable learning outcomes,” Journal of Infrastructure, Policy and Development, vol. 8, no. 12, p. 6385, Nov. 2024, https://doi.org/10.24294/jipd.v8i12.6385.
[9] S. C. Vetrivel, V. P. Arun, R. Ambikapathi, and T. P. Saravanan, “Automated Grading Systems,” in Adopting Artificial Intelligence Tools in Higher Education, pp. 41–61, 2025, https://doi.org/10.1201/9781003470304-3.
[10] R. Rajput, “Use of Artificial Intelligence to Solve Problems in the Classroom,” in New Technological Applications in the Flipped Learning Model, pp. 137–168, 2025, https://doi.org/10.4018/979-8-3373-0437-3.ch005.
[11] E. Mupaikwa, “The Use of Artificial Intelligence in Education,” Emerging Technology-Based Services and Systems in Libraries, Educational Institutions, and Non-Profit Organizations, pp. 26–50, 2023, https://doi.org/10.4018/978-1-6684-8671-9.ch002.
[12] M. Guettala, S. Bourekkache, O. Kazar, and S. Harous, “Generative Artificial Intelligence in Education: Advancing Adaptive and Personalized Learning,” Acta Informatica Pragensia, vol. 13, no. 3, pp. 460–489, Aug. 2024, https://doi.org/10.18267/j.aip.235.
[13] G. S. Divya, “Impact of GenAI on Student Outcomes,” Generative Artificial Intelligence Empowered Learning, pp. 139–175, 2025, https://doi.org/10.1201/9781003422433-7 .
[14] D. A. Popescu, N. Bold, and M. Stefanidakis, “A Systematic Model of an Adaptive Teaching, Learning and Assessment Environment Designed Using Genetic Algorithms,” Applied Sciences, vol. 15, no. 7, p. 4039, Apr. 2025, https://doi.org/10.3390/app15074039.
[15] E. Yehia, “Developments on Generative AI,” AI and Emerging Technologies, pp. 139–160, 2024, https://doi.org/10.1201/9781003501152-9.
[16] A. Al-Dahoud, M. Fezari, A. Al-Dahoud, D. Aqel, H. Mimi, and M. Sh. Daoud, “Revolutionizing Space: The Potential of Artificial Intelligence,” WSEAS Transactions on Computer Research, vol. 12, pp. 404–414, Sep. 2024, https://doi.org/10.37394/232018.2024.12.40.
[17] D. Gupta, V. Swami, D. Shukla, and K. Nimala, “Design and Implementation of an AI-Driven Mental Health Chatbot: A Generative AI Model,” 2024 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES), pp. 1–7, Dec. 2024, https://doi.org/10.1109/ICSES63760.2024.10910363.
[18] N. Anderson, A. McGowan, L. Galway, P. Hanna, M. Collins, and D. Cutting, “Implementing Generative AI and Large Language Models in Education,” 2023 7th International Symposium on Innovative Approaches in Smart Technologies (ISAS), pp. 1–6, Nov. 2023, https://doi.org/10.1109/ISAS60782.2023.10391517.
[19] H. Guo, W. Yi, and K. Liu, “Enhancing Constructivist Learning: The Role of Generative AI in Personalised Learning Experiences,” Proceedings of the 26th International Conference on Enterprise Information Systems, pp. 767–770, 2024, https://doi.org/10.5220/0012688700003690.
[20] A. P. Gervacio, “Exploring how generative AI contributes to the motivated engagement and learning production of science-oriented students,” Environment and Social Psychology, vol. 9, no. 11, Dec. 2024, https://doi.org/10.59429/esp.v9i11.3194.
[21] R. Chhatrala and D. Chaudhari, “Asking Right Questions: Towards Engaging and Inclusive Learning Environment to Enhance Creativity Using Generative AI,” General Aspects of Applying Generative AI in Higher Education, pp. 187–206, 2024, https://doi.org/10.1007/978-3-031-65691-0_10.
[22] R. Sandhu, H. K. Channi, D. Ghai, G. S. Cheema, and M. Kaur, “An Introduction to Generative AI Tools for Education 2030,” Integrating Generative AI in Education to Achieve Sustainable Development Goals, 2024, pp. 1–28, https://doi.org/10.4018/979-8-3693-2440-0.ch001.
[23] R. Yadav, G. Huzooree, M. Yadav, and D. S. K. Gangodawilage, “Generative AI for Personalized Learning Content Creation,” Transformative AI Practices for Personalized Learning Strategies, 2025, pp. 107–130, https://doi.org/10.4018/979-8-3693-8744-3.ch005.
[24] A. Weiand, E. Reategui, and R. Motz, “Using Artificial Intelligence to Guide the Learning Analytics Process,” Proceedings of the 19th Latin American Conference on Learning Technologies (LACLO 2024), 2025, pp. 481–491, https://doi.org/10.1007/978-981-96-3698-3_34.
[25] L. Y. Tan, S. Hu, D. J. Yeo, and K. H. Cheong, “Artificial intelligence-enabled adaptive learning platforms: A review,” Computers and Education: Artificial Intelligence, vol. 9, p. 100429, Dec. 2025, https://doi.org/10.1016/j.caeai.2025.100429.
[26] B. Hawkins, D. Taylor-Griffiths, and J. M. Lodge, “Summarise, elaborate, try again: exploring the effect of feedback literacy on AI-enhanced essay writing,” Assess Eval High Educ, pp. 1–13, Apr. 2025, https://doi.org/10.1080/02602938.2025.2492070.
[27] A. Hadyaoui and L. Cheniti-Belcadhi, “IntelliFrame: A Framework for AI-Driven, Adaptive, and Process-Oriented Student Assessments,” Proceedings of the 20th International Conference on Web Information Systems and Technologies, pp. 441–448, 2024, https://doi.org/10.5220/0013070800003825.
[28] A. P. Gervacio, “Exploring how generative AI contributes to the motivated engagement and learning production of science-oriented students,” Environment and Social Psychology, vol. 9, no. 11, Dec. 2024, https://doi.org/10.59429/esp.v9i11.3194.
[29] K. I. Katonane Gyonyoru and J. Katona, “Student Perceptions of AI-Enhanced Adaptive Learning Systems: A Pilot Survey,” 2024 IEEE 7th International Conference and Workshop Óbuda on Electrical and Power Engineering (CANDO-EPE), pp. 93–98, Oct. 2024, https://doi.org/10.1109/CANDO-EPE65072.2024.10772884.
[30] W. Powell and S. Courchesne, “Opportunities and risks involved in using ChatGPT to create first grade science lesson plans,” PLoS One, vol. 19, no. 6, p. e0305337, Jun. 2024, https://doi.org/10.1371/journal.pone.0305337.
[31] D. DiSabito, L. Hansen, T. Mennella, and J. Rodriguez, “Exploring the frontiers of generative AI in assessment: Is there potential for a human‐AI partnership?,” New Directions for Teaching and Learning, vol. 2025, no. 182, pp. 81–96, Jun. 2025, https://doi.org/10.1002/tl.20630.
[32] S. K. Sajida Sultana, R. Renugadevi, M. Bhargavi, and S. A. Afzal Biyabani, “AI-Driven Evaluation Techniques,” Adopting Artificial Intelligence Tools in Higher Education, pp. 1–22, 2025, https://doi.org/10.1201/9781003470304-1.
[33] A. Ayari and L. Ouerfelli, “Generative Ai in the Classroom: Balancing Innovation, Fear, and Necessity,” 2025 IEEE Global Engineering Education Conference (EDUCON), pp. 1–6, Apr. 2025, https://doi.org/10.1109/EDUCON62633.2025.11016392.
[34] S. Virkus et al., “Chatbots Scenarios for Education,” Information and Software Technologies, pp. 207–221, 2024, https://doi.org/10.1007/978-3-031-48981-5_17.
[35] B. Dong, J. Bai, T. Xu, and Y. Zhou, “Large Language Models in Education: A Systematic Review,” 2024 6th International Conference on Computer Science and Technologies in Education (CSTE), pp. 131–134, Apr. 2024, https://doi.org/10.1109/CSTE62025.2024.00031.
[36] N. Kerimbayev, Z. Menlibay, M. Garvanova, S. Djaparova, and V. Jotsov, “A Comparative Analysis of Generative AI Models for Improving Learning Process in Higher Education,” 2024 International Conference Automatics and Informatics (ICAI), pp. 271–276, Oct. 2024, https://doi.org/10.1109/ICAI63388.2024.10851491.
[37] H. Huang and H. W. Chuang, “Assessing the Effectiveness of Easy-to-Read Principles in Different Generative AI Models: Enhancing Knowledge Acquisition for People with Disabilities,” 2024 International Conference on Orange Technology (ICOT), pp. 1–5, Dec. 2024, https://doi.org/10.1109/ICOT64290.2024.10936939.
[38] S. Das, D. Santra, T. Chhari, S. Roy, and S. Mukherjee, “Prompt Driven Image Creation: A Comparative Evaluation of Generative AI Tools,” 2025 8th International Conference on Electronics, Materials Engineering & Nano-Technology (IEMENTech), pp. 1–6, Jan. 2025, https://doi.org/10.1109/IEMENTech65115.2025.10959444.
[39] Y. Lee, “Development of Open Large Language Models for Artificial Intelligence Digital Textbooks,” TEM Journal, pp. 2773–2783, Nov. 2024, https://doi.org/10.18421/TEM134-14.
[40] J. I. Alvarez and J. R. Angeles, “Khanmigo in the Virtual Classroom: A Strategic Evaluation through SWOT and Acceptability Analysis,” Educational Process International Journal, vol. 16, no. 1, 2025, https://doi.org/10.22521/edupij.2025.16.272.
[41] S. P. S. Deshmukh, V. Kulkarni, and P. Malathi, “Leveraging Interactive Learning by Integrated Assessment Software in EdTech: Enhancing Sustained Learning Outcomes with Mentimeter,” Journal of Engineering Education Transformations, vol. 38, no. IS2, pp. 14–22, May 2025, https://doi.org/10.16920/jeet/2025/v38is2/25003.
[42] L. Javourey-Drevet, S. Dufau, T. François, N. Gala, J. Ginestié, and J. C. Ziegler, “Simplification of literary and scientific texts to improve reading fluency and comprehension in beginning readers of French,” Appl Psycholinguist, vol. 43, no. 2, pp. 485–512, Mar. 2022, https://doi.org/10.1017/S014271642100062X.
[43] H. Fan, G. Chen, X. Wang, and Z. Peng, “LessonPlanner: Assisting Novice Teachers to Prepare Pedagogy-Driven Lesson Plans with Large Language Models,” Proceedings of the 37th Annual ACM Symposium on User Interface Software and Technology, pp. 1–20, Oct. 2024, https://doi.org/10.1145/3654777.3676390.
[44] A. Sharma and D. B. Jayagopi, “Modeling essay grading with pre-trained BERT features,” Applied Intelligence, vol. 54, no. 6, pp. 4979–4993, Mar. 2024, https://doi.org/10.1007/s10489-024-05410-4.
[45] P. Belim, N. Bhatt, A. Lathigara, and H. Durani, “Enhancing Level of Pedagogy for Engineering Students Through Generative AI,” Journal of Engineering Education Transformations, vol. 38, no. IS2, pp. 463–470, May 2025, https://doi.org/10.16920/jeet/2025/v38is2/25057.
[46] X. Wang and L. Lei, “A Path Study of Generative Artificial Intelligence Enabling Online Education Platforms in Colleges and Universities,” Proceedings of the 2024 International Symposium on Artificial Intelligence for Education, pp. 332–338, Sep. 2024, https://doi.org/10.1145/3700297.3700354.
[47] J. Velander, N. Otero, F. Dobslaw, and M. Milrad, “Eliciting and Empowering Teachers’ AI Literacy: The Devil is in the Detail,” Methodologies and Intelligent Systems for Technology Enhanced Learning, 14th International Conference, pp. 138–152, 2024, https://doi.org/10.1007/978-3-031-73538-7_13.
[48] N. Guillén-Yparrea and F. Hernández-Rodríguez, “Unveiling Generative AI in Higher Education: Insights from Engineering Students and Professors,” 2024 IEEE Global Engineering Education Conference (EDUCON), pp. 1–5, May 2024, https://doi.org/10.1109/EDUCON60312.2024.10578876.
[49] A. Dhruv, S. Saha, S. Tyagi, and V. Jain, “Investigating the Transformative Impact of Generative AI on Academic Integrity Across Diverse Educational Domains,” 2024 2nd International Conference on Advancement in Computation & Computer Technologies (InCACCT), pp. 87–92, May 2024, https://doi.org/10.1109/InCACCT61598.2024.10551108.
[50] A. M. Elkhatat, K. Elsaid, and S. Almeer, “Evaluating the efficacy of AI content detection tools in differentiating between human and AI-generated text,” International Journal for Educational Integrity, vol. 19, no. 1, p. 17, Sep. 2023, https://doi.org/10.1007/s40979-023-00140-5.
[51] D. Weber-Wulff et al., “Testing of detection tools for AI-generated text,” International Journal for Educational Integrity, vol. 19, no. 1, p. 26, Dec. 2023, https://doi.org/10.1007/s40979-023-00146-z.
[52] N. Nananukul and M. Kejriwal, “HALO: an ontology for representing and categorizing hallucinations in large language models,” in Disruptive Technologies in Information Sciences VIII, p. 9 Jun. 2024, https://doi.org/10.1117/12.3014048.
[53] R. Qadri, R. Shelby, C. L. Bennett, and E. Denton, “AI’s Regimes of Representation: A Community-centered Study of Text-to-Image Models in South Asia,” 2023 ACM Conference on Fairness, Accountability, and Transparency, pp. 506–517, Jun. 2023, https://doi.org/10.1145/3593013.3594016.
[54] T. R. Mcintosh, T. Liu, T. Susnjak, P. Watters, and M. N. Halgamuge, “A Reasoning and Value Alignment Test to Assess Advanced GPT Reasoning,” ACM Transactions on Interactive Intelligent Systems, vol. 14, no. 3, pp. 1–37, Sep. 2024, https://doi.org/10.1145/3670691.
[55] D. T. K. Ng, E. K. C. Chan, and C. K. Lo, “Opportunities, challenges and school strategies for integrating generative AI in education,” Computers and Education: Artificial Intelligence, vol. 8, p. 100373, Jun. 2025, https://doi.org/10.1016/j.caeai.2025.100373.
[56] M. El Samaty, “Beyond AI-Enabled Classrooms: Fostering Critical Thinking in the Age of Generative Artificial Intelligence,” Prompt Engineering and Generative AI Applications for Teaching and Learning, pp. 527–546, 2025, https://doi.org/10.4018/979-8-3693-7332-3.ch030.
[57] C. Gonsalves, “Generative AI’s Impact on Critical Thinking: Revisiting Bloom’s Taxonomy,” Journal of Marketing Education, Dec. 2024, https://doi.org/10.1177/02734753241305980.
[58] J. Liu et al., “A bibliometric analysis of generative AI in education: current status and development,” Asia Pacific Journal of Education, vol. 44, no. 1, pp. 156–175, Jan. 2024, https://doi.org/10.1080/02188791.2024.2305170.
[59] J. L. Pallant, J. Blijlevens, A. Campbell, and R. Jopp, “Mastering knowledge: the impact of generative AI on student learning outcomes,” Studies in Higher Education, pp. 1–22, May 2025, https://doi.org/10.1080/03075079.2025.2487570.
[60] L. R. Salazar, S. F. Peeples, and M. E. Brooks, “Generative AI Ethical Considerations and Discriminatory Biases on Diverse Students Within the Classroom,” The Role of Generative AI in the Communication Classroom, pp. 191–213, 2024, https://doi.org/10.4018/979-8-3693-0831-8.ch010.
[61] Y. Assefa, M. M. Gebremeskel, B. T. Moges, S. A. Tilwani, and Y. A. Azmera, “Rethinking the digital divide and associated educational in(equity) in higher education in the context of developing countries: the social justice perspective,” The International Journal of Information and Learning Technology, vol. 42, no. 1, pp. 15–32, Jan. 2025, https://doi.org/10.1108/IJILT-03-2024-0058.
[62] M.-S. Ramírez-Montoya, L. M. Oliva-Córdova, and A. Patiño, “Training Teaching Personnel in Incorporating Generative Artificial Intelligence in Higher Education: A Complex Thinking Approach,” Proceedings of TEEM 2023, pp. 163–175, 2024, https://doi.org/10.1007/978-981-97-1814-6_16.
Refbacks
- There are currently no refbacks.
Copyright (c) 2025 Purwono Purwono
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Journal of Technological Pedagogy and Educational Development
ISSN:
Organized by Peneliti Teknologi Teknik Indonesia
Published by Peneliti Teknologi Teknik Indonesia
Website: https://ejournal.jtped.org/ojs/index.php/jtped
Email: jtped@ptti.web.id
Address: Yogyakarta, Indonesia