Volume 5 Nomor 2 Februari 2026 http://jurnal.
DIGITAL LEADERSHIP AND AI INTEGRATION IN HIGHER EDUCATION:
TOWARD A HOLISTIC FRAMEWORK FOR SUSTAINABLE QUALITY EDUCATION
Rokimin1.
Mona Novita2.
Sofwan Manaf3.
Fajar Suryono4.
Miftah Hur Rahman ZH5.
Zahid Zufar At Thaariq6.
Dudun Ubaedullah7 13 Universitas Darunnajah.
Indonesia 2 Universitas Nurul Jadid Paiton.
Indonesia 4 Universiti Sains Islam Malaysia 5 Universitas Negeri Malang.
Indonesia 6 Cukurova University.
Adana.
Turkiye 7Universitas Islam Negeri Syarif Hidayatullah Jakarta.
Indonesia rokimin@darunnajah.
ABSRACT
Digital transformation has reshaped debates on access, equity, and quality in higher education, with Artificial Intelligence (AI) as a key enabler.
This study synthesizes existing literature to examine the intersections between AI integration, digital leadership, pedagogical competency, and equityoriented practices.
Using a Systematic Literature Review (SLR) with the PRISMA framework, the study analyzed 69 peer-reviewed articles .
2Ae2.
from Scopus and Watase databases.
The findings identified three main themes: expanding digital access, advancing learning equity, and improving quality through AI-based pedagogy.
Significant regional differences exist.
uch as Indonesia.
Vietnam, and Thailan.
focus on infrastructure and teacher training, while developed countries emphasize ethics and curriculum innovation.
Methodologically, quantitative approaches dominate and theories such as TPACK and TAM are relied upon.
The study concludes that AI's impact is maximized when combined with human factors, such as teacher selfefficacy and institutional leadership.
AI is not a neutral tool, but rather a transformative force tied to ethics and inclusivity.
Future research should prioritize methodological pluralism and long-term evaluation to align technological advances with educational equity.
Keywords: Artificial Intelligence.
Digital Leadership.
Equality and Quality.
Higher Education.
Teacher Competence.
INTRODUCTION
The rapid pace of digital transformation has undeniably reshaped the landscape of higher education, bringing both opportunities and challenges(Samiyono et al.
, 2025.
Zubairi & Nurdin, 2.
Ensuring quality education is no longer merely a matter of providing physical infrastructure.
instead, it has become closely intertwined with questions of equity, accessibility, and sustainability at the global level(Mak et al.
, 2.
Interestingly, the rise of artificial intelligence (AI) has amplified these concerns, making its integration into educational systems a matter of urgency rather than an option.
At the same time, much of the earlier scholarship focused on traditional service quality in higher education, such as the development of sustainable measurement tools (Ozdemir et al.
, 2.
or the deployment of digital learning platforms (Linzalone et al.
, 2.
, more recent studies have Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
revealed a clear shift toward human capacity, digital competencies, and innovative pedagogical practices as critical dimensions of quality.
Regional evidence provides further nuance.
In Southeast Asia, for example, emphasized the adoption of digital platforms in ThailandAos post-pandemic recovery, while examined how AI integration could foster digital leadership among EFL teachers in Vietnam.
Meanwhile, research in Morocco and Pakistan highlighted the role of knowledge-based leadership and institutional reform in sustaining quality (Fahim et al.
, 2.
(Mansoor & Hussain, 2.
Notably, demonstrated that AI-driven professional training significantly boosted teachersAo self-efficacy and higher-order thinking skills in Indonesia and China.
More recently, (Biloshchytskyi et al.
, 2.
and (MuhammedZein & Abdullateef, 2.
underlined the importance of sustainable pedagogical approaches in Kazakhstan and Saudi Arabia, adding further weight to the argument that holistic frameworks are essential for equitable and future-proof education.
However, despite this growing body of research, several blind spots remain.
Earlier investigations, though valuable, were confined mainly to the physical or technical dimensions of education, leaving digital competencies and resilience underexplored (Y.
Liu et al.
, 2.
Similarly, although AI-driven initiatives have been shown to improve teacher confidence and classroom innovation (Yang et al.
, 2.
, their interplay with leadership practices and sustainability agendas has received limited attention.
Prior reviews, whether systematic (Wang et al.
, 2.
, bibliometric or scoping (Ahmad et al.
, 2023.
Sofwan Manaf, 2.
, were informative but often thematically narrow, geographically uneven, and lacking integrative synthesis.
Importantly, there is still little evidence regarding long-term impacts, particularly in the Global South, where infrastructural and cultural barriers remain acute.
This study, therefore, positions itself at the intersection of these debates.
proposing a holistic review that integrates technological capacity, pedagogical competence, and digital leadership, it addresses a gap left by previous works that tended to isolate these The novelty of this research lies precisely in its attempt to weave together these dimensions into a single framework, thereby offering both conceptual and practical contributions(Hoet et al.
, 2020.
Sofwan Manaf 2.
On the one hand, it advances theory by proposing a model that recognizes the interdependence of technology, pedagogy, and leadership in shaping sustainable education.
On the other hand, it informs practice by offering insights for policymakers and institutions seeking to design professional development and digital strategies that are both inclusive and resilient(Veinovic, 2.
The justification for this review is threefold.
First, it aligns with the global commitment to the United NationsAo Sustainable Development Goal 4 (Quality Educatio.
, situating the discussion within the broader agenda of equity and inclusivity.
Second, it extends the geographic coverage of prior reviews by incorporating evidence from underrepresented contexts such as Indonesia.
Malaysia, and Vietnam, which often face unique institutional and infrastructural challenges.
Third, it explicitly incorporates recent debates on generative AI tools .
ChatGPT), which are beginning to redefine pedagogical innovation but remain underexamined in existing reviews.
Taken together, these elements provide a strong rationale for conducting a systematic literature review that not only maps current trends but also develops new conceptual linkages(Rahimi & Teimouri, 2.
To frame the scope of inquiry, this review formulates several interrelated research First, it examines the diverse contexts in which AI-driven interventions and Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
digital platforms have been deployed to address pressing challenges of access, equity, and educational quality.
Second, it investigates the mechanisms that mediate their effectiveness, ranging from teacher self-efficacy and digital literacy to institutional leadership, while considering how these factors interact across settings.
Third, it explores the reported outcomes with respect to equity, learning achievement, and sustainability, highlighting variations across geographical and institutional contexts.
Fourth, it analyzes prevailing research trends, particularly in terms of publication focus, methodological approaches, and theoretical frameworks, to reveal persistent gaps that limit cumulative knowledge building.
Finally, it identifies potential future directions, with particular attention to the long-term consequences and ethical implications of integrating AI into educational systems.
combining elements of research profiling with thematic analysis, these questions ensure coherence between the reviewAos novelty and its objectives.
The article unfolds in four main The Methodology explains the systematic review process, including database selection and screening protocols.
The Findings map the key interventions, mechanisms, and outcomes identified across the literature.
The Analysis section integrates these insights into a holistic framework that connects technology, pedagogy, and leadership.
Finally, the Conclusion highlights theoretical contributions, practical implications, and future research This structure is deliberately designed to ensure both analytical depth and conceptual clarity.
METHOD
This research adopted a Systematic Literature Review (SLR) guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework (Moher et al.
, 2.
The decision to employ PRISMA was not merely procedural but also strategic, as the framework is widely recognized for improving both the transparency and the methodological rigor of systematic reviews (Siddaway et al.
, 2.
Notably, the process unfolded across four interconnected stages: identification, screening, eligibility, and Each stage contributed to narrowing the evidence base in a structured and replicable manner, and the overall flow of this process is presented in Figure 1.
During the identification stage, the search strategy relied on carefully chosen keywords such as AuTeachers in Digital Era,Ay AuDigital Learning Platforms,Ay AuProfessional Teachers with AI,Ay AuContinuing Education for Teachers,Ay AuSustainable Quality Education,Ay and AuTeacher Professionalism in Schools.
Ay These terms were deliberately selected to capture the breadth of scholarship addressing teacher competence in a digitalized educational landscape, particularly the integration of Artificial Intelligence (AI).
The primary database was Scopus, which has the advantage of stringent indexing and quality assurance (Lasda Bergman, 2.
and (Rocha et al.
, 2.
Interestingly, this choice was also pragmatic because, unlike Google Scholar.
Scopus minimizes duplicate records and helps avoid predatory journals (Hariningsih et al.
, 2.
To enrich the dataset, supplementary searches were carried out using the Watase database (Wahyudi, 2.
Altogether, the search initially generated 799 records.
The subsequent screening stage aimed to refine this corpus.
At this point, 38 duplicate entries were removed, along with 117 publications outside the target period .
5Ae2.
, 89 studies from journals that did not meet the required tier standards, and 18 papers without adequate abstracts.
After this process, 537 records remained.
These were then Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
screened manually for thematic relevance.
Interestingly, none were excluded, so the full set progressed to the next step.
The eligibility assessment involved attempting to retrieve complete versions of the 537 records.
However, access barriers meant that only 69 articles could be fully obtained, while 468 remained unavailable.
Each of the retrieved studies was then evaluated in detail for methodological rigor and alignment with the reviewAos objectives, and all were deemed appropriate for inclusion.
Finally, in the inclusion stage, the 69 eligible studies formed the evidence base of this Their findings were analyzed thematically, a strategy that allowed recurring patterns and key themes to emerge organically.
To support this process, the Watase Uake System (Wahyudi, 2.
was used, providing a systematic structure for categorizing and interpreting the literature.
Taken together, this staged process ensured that the review was not only methodologically rigorous but also transparent and replicable, thereby strengthening the trustworthiness of its conclusions.
Figure 1.
Prisma Flow Chart
RESULTS AND DISCUSSION
Thematic Trends in AI and Education Research Three overarching themes dominate the literature: access to education, equity of opportunity, and the role of AI in strengthening educational quality.
These are not isolated rather, they are interconnected, with access acting as a prerequisite for equity, and Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
quality operating both as a target and a measure of success (UNESCO, 2.
In lowerincome regions, unsurprisingly, access is foregrounded.
studies emphasize technological infrastructure and affordability (World Bank, 2.
Meanwhile, research from highincome contexts gravitates toward innovation: digital literacy, personalization of learning through AI, and professional development of teachers (OECD, 2.
Interestingly, citation analysis reveals that articles explicitly incorporating AI tools into their frameworks receive disproportionately higher attention.
This suggests that the academic community increasingly regards AI not as an auxiliary theme but as a central driver of educational Moreover, subthemes such as gender inclusion, accessibility for learners with disabilities, and teacher capacity-building reinforce the idea that AI in education must be understood in relation to broader social justice concerns.
Figure 2: Thematic Trends in AI and Education Research Regional Disparities and Research Contexts One of the most striking patterns concerns geography.
Of the 20 reviewed studies, 12 originate from developing contexts such as Indonesia.
Vietnam.
Thailand, and several African nations (Songkram et al.
, 2.
(Nguyen Huu, 2.
and (Zulianti et al.
, 2.
These studies concentrate on pressing challenges uneven digital infrastructure, limited teacher training, and affordability gaps.
By contrast.
European and North American contributions often pivot toward qualitative inquiries into ethics, curriculum integration, or advanced applications such as machine learning in education (Christen et al.
, 2.
and (Silhol, 2.
This asymmetry is not accidental.
Instead, it reflects global socio-economic divides: while advanced economies experiment with optimization, many developing regions still struggle to establish baseline accessibility.
However, this very imbalance creates fertile ground for comparative inquiry, pushing scholars to ask how global frameworks might bridge regional disparities without exacerbating them (Ebrahimi et al.
, 2.
Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
Figure 3.
Regional Disparities and Research Contexts Temporal Dynamics in Publication Output The temporal distribution of studies between 2020 and 2025 reveals notable surges in 2022 and 2024, years shaped by both global disruption and rapid innovation.
In 2022, five studies centered on the immediate need for AI-enabled remote learning, a clear response to the COVID-19 pandemic (Christen et al.
, 2.
and (Y.
Liu et al.
, 2.
By 2024, attention had shifted toward consolidation, with researchers interrogating how these tools could be embedded into long-term pedagogical practices (Silhol, 2.
What is particularly noteworthy is how the pandemic functioned as a catalyst: it not only accelerated the adoption of AI but also laid bare systemic weaknesses, from infrastructural fragility to teacher preparedness.
The trajectory suggests that subsequent research will increasingly examine long-term consequences, rather than emergency responses.
Methodologically, the literature leans heavily on quantitative designs, particularly those deploying survey data analyzed through Structural Equation Modeling (SEM) or quasi-experiments (Songkram et al.
, 2.
Nine out of twenty studies adopt this approach, generating results that are replicable and generalizable but often stripped of contextual Mixed methods studies (Nguyen Huu, 2.
attempt to counterbalance this by triangulating statistical models with semi-structured interviews, offering a richer texture of Purely qualitative works, though fewer, contribute depth in areas such as ethics and cultural adaptation (Muis & Huda, 2.
Notably, the privileging of quantitative approaches reveals a disciplinary bias: the pursuit of generalizability often overshadows the subtle, socially embedded dimensions of AI adoption.
This imbalance underscores the necessity of methodological pluralism in future research.
Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
Figure 4.
Methods Classification Data Collection and Analytical Tools Instruments and tools also reveal telling patterns.
Questionnaires dominate as the preferred means of data collection, used in eight studies (Songkram et al.
, 2.
Interviews and hybrid instruments appear less frequently but are nonetheless critical for contextual Analytical tools mirror this distribution: SPSS is ubiquitous in statistical modeling, while NVivo serves as the mainstay for qualitative coding.
More recent work has begun incorporating Python-based machine learning or TensorFlow, signaling a gradual methodological shift (Wang et al.
, 2.
Interestingly, this diversification reflects not only technological innovation but also the growing willingness of education researchers to collaborate with computer scientists.
Such interdisciplinary convergence strengthens empirical rigor but simultaneously demands new forms of expertise that many educational institutions have yet to cultivate fully.
Theoretical Foundations: Toward Integration The studies reviewed draw from a diverse yet convergent set of theoretical The Technological Pedagogical Content Knowledge (TPACK) model remains a favored lens for examining classroom-level AI adoption (Zulianti et al.
, 2.
Social Cognitive Theory, meanwhile, highlights the role of self-efficacy and social learning in shaping teachersAo and studentsAo engagement with AI (Yang et al.
, 2.
The Technology Acceptance Model (TAM) is often paired with leadership theories to analyze institutionallevel adoption (Nguyen Huu, 2.
What is particularly compelling is how these frameworks, despite their strengths, tend to be applied in isolation.
This compartmentalization suggests the need for a more integrative theoretical scaffold, one capable of addressing technological, pedagogical, and socio-psychological dimensions High-Impact Journals and Authorial Strategies Citation analysis further illuminates the landscape.
Q1 journals such as Education and Information Technologies and Sustainability dominate as publication outlets (Olthuis et al.
, 2.
Authors who secure space in these journals tend to balance practical Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
applicability with theoretical advancement, thereby appealing to multiple audiences, from researchers to policymakers.
Studies that focus on developing regions (Maltagliati et al.
have garnered particular attention, arguably because they align novelty with This indicates that research framed around equity is not only ethically compelling but also strategically impactful in terms of scholarly visibility(Saidel & Smith, 2.
Despite this robust body of work, critical gaps remain.
Ethical considerations are still underexplored, particularly in contexts where AI risks amplifying inequality(Ladyshewsky & Taplin, 2.
Methodological imbalance, domination by quantitative paradigms, has also limited the depth of contextual insights(Kim et al.
, 2.
Moreover, the uneven geographical distribution of research outputs underscores the need for comparative designs that bridge both North-South and intra-regional divides(Valle et , 2.
Future research would benefit from three shifts: first, the integration of ethical dimensions into empirical analysis.
second, the embrace of methodological pluralism that values both statistical robustness and contextual richness.
and third, the expansion of geographic scope to ensure that global narratives about AI in education are not skewed by overrepresentation from specific regions(Chibuye & Singh, 2.
In short, the path ahead lies not in multiplying studies of the same kind but in deepening, diversifying, and contextualizing them.
Disscusion The findings of this systematic literature review (SLR) offer a broad yet nuanced account of how Artificial Intelligence (AI) has been integrated into education, particularly in relation to access, equity, and quality.
Beyond merely cataloguing previous works, this review critically reflects on how these themes converge, diverge, and occasionally leave essential questions unanswered.
Unlike earlier syntheses, which tended to concentrate narrowly on technological functionalities, this review situates AI within a broader sociopedagogical and ethical landscape, thereby providing fresh insights into its transformative yet contested role in global education.
Revisiting the Dominant Themes in Light of Prior Studies A consistent observation across the reviewed literature is the prominence of three interrelated themes: access to education, equity in learning opportunities, and quality These emphases are primarily in line with earlier analyses by (X.
Liu et al.
and (Lei & Tang, 2.
, who pointed to the ways AI can personalize learning and automate assessment to raise educational outcomes.
Yet this review extends the findings of such studies by stressing the systemic implications of AI adoption, including its potential to reshape the broader architecture of education systems rather than optimize classroom Interestingly, recent works focusing on teacher self-efficacy and student adaptability (Tolentino et al.
, 2.
add nuance to older debates that primarily emphasized infrastructure and access (OECD, 2.
What emerges here is the recognition that the successful uptake of AI is not merely a technical achievement but a profoundly humancentered process, mediated by trust, training, and the professional identities of educators.
This represents a notable shift from earlier research, where the human dimension was often treated as secondary to technological capability.
Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
Contextual Variations: Developed and Developing Worlds The geographical distribution of studies provides another essential layer of In developing contexts such as Indonesia.
Vietnam, and Thailand, the literature prioritizes teacher training and infrastructural readiness (Songkram et al.
, 2.
By contrast, in advanced economies such as the United States and the United Kingdom, the focus shifts toward sophisticated pedagogical innovations and curriculum redesigns(Alenezi & Alfaleh, 2.
This contrast is in line with institutional reports by UNESCO 2021 and the World Bank 2020, which similarly highlight stark digital divides(UNESCO, 2.
However, what distinguishes this review is its empirical emphasis.
Synthesizing peer-reviewed studies rather than policy documents alone it adds nuance to our understanding of how these divides materialize in practice.
Strikingly, the evidence suggests that while technology can indeed mitigate inequities, its effectiveness is highly contingent on local ecosystems of policy, culture, and institutional support.
This observation runs contrary to earlier arguments by (Moher et al.
, 2.
, who contended that technology by itself cannot narrow educational gaps.
The findings here demonstrate that, under the right conditions.
AI-enabled platforms can reduce barriers to access, although admittedly not uniformly across all contexts.
Methodological and Theoretical Contributions From a methodological perspective, the heavy reliance on quantitative approaches most notably Structural Equation Modeling (SEM) and Exploratory Factor Analysis (EFA) is not surprising and remains in line with prior meta-analyses (Priscilla Josierra Liora et al.
Yet this review highlights a less-discussed issue: such methodological homogeneity risks obscuring the complex social realities of AI integration.
Notably, the relative absence of longitudinal and qualitative designs constrains our ability to assess not only causal relationships but also the lived experiences of students and teachers navigating AI-driven Theoretically, the field continues to be anchored in frameworks such as the Technological Pedagogical Content Knowledge (TPACK) model, the Technology Acceptance Model (TAM), and Social Cognitive Theory.
These models, while undoubtedly influential (Zulianti et al.
, 2.
, reveal certain blind spots when scrutinized in light of this reviewAos findings.
For instance.
TAMAos reliance on constructs such as Auperceived usefulnessAy and Auease of useAy seems insufficient to explain adoption patterns in contexts shaped by socio-cultural hierarchies and infrastructural constraints.
Likewise, while TPACK helps connect pedagogy, content, and technology, it says little about the ethical challenges of algorithmic bias or the inclusivity of marginalized learners (Brown et al.
, 2.
This review, therefore, extends the findings of prior theoretical applications by suggesting that hybrid models are urgently required models that merge technological frameworks with equityoriented and ethical perspectives.
Such integration resonates with (Freire, 1.
critical pedagogy, which foregrounds empowerment and justice, and aligns with (Selwyn, 2.
call for more holistic approaches to educational technology.
Ethical and Equity-Oriented Challenges Another essential dimension concerns the ethical tensions inherent in AI adoption.
The reviewed studies consistently document concerns around algorithmic bias, privacy, and inequitable access.
These findings are clearly in line with ongoing global debates about AI Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
ethics (Guidelines, 2.
and (Brown et al.
, 2.
warning that, without careful governance.
AI may exacerbate rather than alleviate educational inequality.
Yet this review adds nuance to such concerns by showing how ethical issues are unevenly distributed.
In developed contexts, privacy and data protection dominate discussions, while in developing settings, the most pressing matters revolve around infrastructural scarcity and culturally insensitive This geographical differentiation, rarely highlighted in prior syntheses, underscores the importance of situating AI ethics within specific socio-political contexts rather than treating them as universally identical challenges.
Novelty and Contribution of This Review This review offers a novel perspective by systematically connecting the dots between access, equity, and quality in AI-based education.
Unlike earlier reviews, which often examined these dimensions in isolation, the present synthesis demonstrates how they are deeply intertwined and mutually reinforcing.
Particularly understudied in previous scholarship is the role of socio-cultural and ethical considerations, which this review brings to the forefront.
Moreover, the attention to human-centered variables such as teacher confidence and student adaptability represents a significant departure from technologically deterministic accounts.
By highlighting these factors, the review reframes AI adoption not as an inevitable march of progress but as a negotiated process shaped by human agency.
Finally, the comparative scope covering both developed and developing regions addresses a persistent imbalance in the literature, where technologically advanced settings tend to dominate scholarly narratives.
This inclusivity strengthens the novelty of this review and enhances its contribution to the field.
Theoretical and Practical Implications Theoretically, the findings call for a broader conceptual repertoire in the study of AI in education.
Existing frameworks remain valuable, yet they do not fully account for the ethical, cultural, and political complexities revealed here.
Future research should therefore experiment with integrating critical theory, complexity theory, and equity frameworks to produce richer and more context-sensitive accounts.
Such theoretical expansion is not only academically desirable but also practically necessary if research is to guide responsible AI On the practical front, several implications emerge.
At the institutional level, teacher professional development should not be limited to technical training but should also cultivate ethical literacy and cultural responsiveness.
At the policy level, governments must design supportive infrastructures that prioritize equitable access, particularly in underserved communities.
For technology developers, the challenge lies in producing AI tools that are transparent, bias-aware, and adaptable to diverse educational realities.
Notably, this review underscores the importance of cross-national collaboration.
The sharing of best practices and the co-creation of ethical standards can mitigate risks while amplifying the benefits of AI.
Furthermore, longitudinal evaluations will be critical in tracking both the intended and unintended consequences of AI integration over time.
CONCLUSION
This systematic literature review, employing the TCCM (Theory.
Context.
Characteristics.
Methodolog.
framework, has drawn together major patterns and gaps in Tadbiruna: Jurnal Manajemen Pendidikan Islam Creative Commons Attribution 4.
0 International License.
the scholarship on educational quality in the age of artificial intelligence (AI).
The findings show that theoretical perspectives such as Technological Pedagogical Content Knowledge (TPACK).
Social Cognitive Theory, and the Technology Acceptance Model (TAM) remain the most frequently adopted.
These frameworks have helped explain how technology can be embedded into pedagogy, yet they also reveal certain blind spots.
Notably, the literature has yet to provide a comprehensive framework that addresses pressing issues of ethics, equity, and algorithmic accountability.
In this sense, the present review extends prior work by highlighting the need for conceptual innovation that transcends purely technical or pedagogical concerns and embraces the broader social consequences of AI in education.
Contextual analysis further illustrates how research agendas are unevenly Studies in developing countries such as Indonesia.
India, and Vietnam tend to emphasize access and equity, while those in more advanced contexts focus on innovation and applications in higher education.
Interestingly, this divergence reflects global inequalities in digital infrastructure and priorities, suggesting that meaningful comparisons across regions could yield fresh insights into how AI can contribute to more inclusive educational futures.
Methodologically, the predominance of quantitative approaches, particularly Structural Equation Modeling (SEM) and Exploratory Factor Analysis (EFA), demonstrates the fieldAos empirical rigor.
However, the relative scarcity of qualitative inquiry limits our understanding of the lived experiences of teachers, students, and institutions navigating AI integration.
Mixed-methods research, which is beginning to gain traction, shows promise in bridging this gap, though much remains to be done.
Building on these observations, several avenues for future research stand out.
Scholars should first design and test integrative frameworks that weave together pedagogical, technological, ethical, and social dimensions.
Second, comparative studies across regions with different levels of digital readiness are essential for identifying both shared challenges and context-specific solutions.
Third, more in-depth qualitative and mixed-methods studies are needed to capture the nuances of classroom practice, user perception, and institutional change.
Finally, future work should move beyond short-term evaluations to consider the long-term implications of AI adoption for equity, teacher identity, and institutional culture.
In conclusion, this review not only consolidates existing evidence but also underscores the need for research that is inclusive, critical, and forwardlooking.
By foregrounding ethics and sustainability alongside innovation, it offers a novel perspective on how AI can transform education.
Ultimately.
AI in education should be viewed not simply as a technological upgrade but as an ongoing global experiment that requires collaborative inquiry, careful reflection, and a commitment to educational justice.
BIBLIOGRAPHY