PurposeThis study critically examines how van Dijk's Digital Divide Theory applies to generative artificial intelligence (GenAI) access in regional, remote and rural (RRR) educational contexts. It interrogates whether a framework designed for standardised digital tools adequately captures GenAI-specific barriers, while mapping where existing research concentrates and where significant gaps remain across van Dijk's four dimensions of access.Design/methodology/approachA scoping review was conducted using the Population, Concept, Context (PCC) framework across four academic databases, with AI-assisted Boolean search development validated through pilot testing and supervisory oversight. Strict inclusion criteria focused on peer-reviewed publications examining GenAI in RRR educational settings. Twenty-nine articles were analysed using van Dijk's four-dimensional framework: material access, skills access, usage access and motivational access.FindingsMaterial and skills access emerged as the most researched dimensions, with material access representing a necessary but insufficient condition for equitable GenAI engagement and skills access requiring interventions across student, educator and policy levels simultaneously. Usage patterns cannot be assumed to follow naturally from material and skills access, requiring sustained time investment, pedagogical framing and institutional support. Motivational access remained critically under-researched, with only one study explicitly examining this dimension. The synthesis suggests sequential relationships across the four dimensions, where foundational precursors in material, skills and usage shape the conditions under which motivational engagement develops. This sequential pattern is presented as a provisional theoretical proposition warranting targeted empirical investigation.Research limitations/implicationsThe interconnected nature of access barriers creates systemic constraints extending beyond van Dijk's original conceptualisation. The synthesis suggests that sequential relationships between dimensions help explain why motivational access remains under-researched, with GenAI lacking foundational precursors necessary for widespread motivational engagement in RRR communities. Evidence-based policy addressing material and skills dimensions becomes essential for preventing systematic exclusion from GenAI's transformative educational potential.Practical implicationsThis research reveals that GenAI introduces a less visible form of digital exclusion in regional, remote and rural (RRR) educational contexts. Unlike previous digital divides characterised by observable access deficits, GenAI enables superficially sophisticated outputs from shallow engagement. This produces performative usage, defined here as the production of well-formed GenAI outputs that mask shallow understanding of subject content. Students in RRR communities may appear to use GenAI productively while developing neither subject mastery nor critical evaluation skills. Educational institutions serving RRR populations must therefore develop context-specific assessment frameworks capable of distinguishing between performative and meaningful GenAI engagement, ensuring that technology deployment addresses rather than obscures the educational disadvantages these communities already experience.Social implicationsThis research highlights that motivational disengagement from GenAI in regional, remote and rural (RRR) communities emerges not from technology rejection but from cascading structural barriers. When infrastructure deficits create skills gaps and repeated failure diminishes perceived effort-to-benefit ratios, self-reinforcing cycles of technological disengagement develop that persist even when access improves. Motivational access remains the least researched dimension of GenAI's digital divide, despite being central to the conditions under which other access dimensions translate into sustained engagement. Policymakers and educational institutions serving RRR communities must recognise that addressing material and skills barriers alone proves insufficient without concurrent attention to how these compounding disadvantages shape willingness and aspiration to engage with GenAI technologies.Originality/valueThis study provides the first application of van Dijk's Digital Divide Theory as an analytical framework for examining generative artificial intelligence education research in regional, remote and rural (RRR) contexts. Theoretically, it proposes a provisional reconceptualisation of van Dijk's sequential dimensions, suggesting that material access supports skills development, which supports usage patterns, which subsequently shape motivational engagement. This proposition offers a coherent explanation for why motivational access remains critically under-researched in RRR communities, namely that foundational precursors necessary for motivational engagement are themselves absent. Practically, by exposing significant gaps in understanding how cascading access barriers shape GenAI engagement in geographically disadvantaged settings, this review establishes the empirical justification for targeted research into GenAI adoption and implementation within RRR educational institutions.
Purpose This paper introduces ephegogy, a theory addressing the gap between traditional educational models and adolescents' developmental needs in artificial intelligence (AI)-augmented societies. Blending constructivism, connectivism and generativism, it explores how integrating generative AI (GenAI) and active learning can scaffold motivation, autonomy, and relevance for adolescent learners.Design/methodology/approach A mixed-methods case study conducted in a single secondary-school geography course during 2024–2025. Quantitative data came from matched pre- and post-intervention surveys measuring motivation and self-efficacy (N = 17; N = 11 for the self-efficacy composite) and locus of control (N = 12), using instruments adapted for an under-14, predominantly English-as-an-additional-language cohort. A separate matched engagement survey (N = 24) and an end-of-course co-creation survey (N = 33) provided further data. Qualitative data comprised 74 individual student reflections analysed thematically using the cultural-historical activity theory (CHAT) framework, alongside open-text survey responses from 74 students. The study was designed using three frameworks (self-determination theory, ephegogy and CHAT).Findings Findings are associational and exploratory, specific to this pilot. Composite motivation increased significantly (p = 0.043, d = 0.53), led by a significant rise in school enjoyment (p = 0.028). On a separate instrument, confidence in using AI tools showed the largest single increase (p = 0.016) and motivation relative to traditional courses also rose (p = 0.040); as several items were examined, these are treated as exploratory. Self-efficacy and locus of control showed no reliable change. Qualitatively, choice, the learning contract, and resource autonomy were dominant themes; most students disclosed and reflected on their AI use, and the modal student experience was ambivalent, valuing autonomy while finding it demanding.Research limitations/implications As a single pilot case study with small, matched samples, no control group and a single teacher-researcher as sole coder, findings are context-specific and cannot support causal claims. Future research should evaluate the framework with larger samples, control conditions, and across diverse cultural contexts and disciplines.Practical implications Ephegogy provides actionable strategies for scaffolding student autonomy and ethical AI use within project-based learning. Tools such as the Student-Author Voice (SAV) rubric support metacognitive reflection and responsible co-creation, assessing authored voice rather than relying on AI detection.Social implications Ephegogy provides actionable strategies for scaffolding student autonomy and ethical AI use within project-based learning. Tools such as the SAV rubric support metacognitive reflection and responsible co-creation, assessing authored voice rather than relying on AI detection.Originality/value Ephegogy reconceptualises adolescent education for the AI era as a distinct developmental stage requiring its own framework, operationalised through gradual immersion and the assessment of authored voice. This paper reports an initial classroom pilot offering preliminary, transferable insights rather than confirmatory evidence.
PurposeThis critical integrative synthesis examines generative AI integration in K-12 and higher education through an explicit decolonial lens. It focuses on how colonial mediations shape opportunities, challenges and strategies for equitable access and outcomes aligned with UN SDG 10.Design/methodology/approachFollowing a critical integrative review paradigm that fuses the integrative logic of Whittemore and Knafl (2005) with reflexive thematic analysis (Braun and Clarke, 2006) and decolonial theory, 58 sources (2020–2026) were identified through systematic search and purposive theoretical sampling. Sources prioritizing structural critique, Global South perspectives, K-12 evidence gaps, output-level neocolonial patterns and contextual agency were retained. Sources were appraised with the Mixed Methods Appraisal Tool and analyzed through a seven-phase relational mapping protocol that translated thematic codes into the layered architecture of the Decolonial Heuristic Layered Equity Framework (DHLEF). Robustness of core pathways was confirmed through sensitivity analyses.FindingsThree interdependent themes – transformative opportunities, interlinked challenges and enabling strategies – emerge when read through decolonial theory. Opportunities appear as outer potential outcomes mediated by challenges whose effects are conditioned by data colonialism, epistemic colonialism, infrastructure dependence, governance colonialism and output-level neocolonial patterns. Strategies achieve moderating effects most effectively when reframed as sovereignty practices, including contextual hacking and language sovereignty. The DHLEF renders these colonial mediations visible in hypothesized relational pathways and supplies a context-differentiated implementation roadmap grounded in empirical evidence from low- and middle-income countries (LMIC) and K-12 contexts.Research limitations/implicationsReliance on secondary and grey literature limits depth and generalizability. The DHLEF remains a synthesis-derived heuristic whose pathways constitute propositions requiring empirical validation through longitudinal and intersectional designs. The synthesis provides a transparent foundation for structural analysis rather than prescriptive protocols.Practical implicationsThe DHLEF supplies institutions and policymakers with a context-sensitive, non-prescriptive heuristic for navigating generative AI integration. It enables differentiated decision-making by educational level and geopolitical resource setting, positioning sovereignty-oriented practices – offline-first tools, community governance and Indigenous data protocols – as conditional moderators rather than sequential prerequisites. Educators and technology providers can apply the framework to diagnose infrastructure dependence, epistemic bias and accessibility barriers prior to deployment, while prioritizing institutional support structures empirically shown to strengthen equitable outcomes in low-resource and K-12 environments. Implementation remains iterative and locally calibrated rather than universally prescribed.Social implicationsBy rendering colonial mediations analytically central, the DHLEF advances structural justice rather than distributional inclusion alone, interrupting asymmetries in knowledge validation and educational opportunity. It positions sovereignty practices – contextual hacking, language sovereignty, community data governance and offline-first infrastructures – as necessary interruptions of data, epistemic and infrastructural extraction, with particular urgency for marginalized learners in Global South and low-resource contexts. The framework reframes equitable GenAI integration as an ongoing political and ethical project of refusal and resurgence, rather than a technical optimization problem, thereby aligning educational practice with the structural demands of UN SDG 10.Originality/valueThe DHLEF advances understanding by converting thematic synthesis into a relational structure that makes colonial mediations analytically central. It extends prior static ethics and education frameworks through decolonial integration and synthesis-derived pathways. It incorporates contextual differentiation by educational level and geopolitical setting, positions 2025–2026 trends as amplifiers of colonial dynamics and offers a tool for equitable GenAI integration that centers sovereignty rather than diffusion. The framework resolves three persistent tensions in existing scholarship: the absence of dynamic relational modelling, the dominance of distributional over structural equity framings and Northern-centric biases that have limited applicability to Global South and K-12 contexts.
PurposeThis paper proposes a pedagogical framework for cultivating inquiry-driven agency in student engagement with artificial intelligence (AI). It addresses a gap in the AI-in-education literature, in which substantial scholarly attention to material access and algorithmic fairness has been accompanied by comparatively less sustained attention to the learner-side capacities that determine whether AI is engaged with critically or passively. The paper offers three connected contributions: a cyclical model of inquiry-driven learning, a four-phase curricular sequence for student-AI engagement, and an operational assessment counterpart that remains diagnostically meaningful in AI-mediated environments.Design/methodology/approachThe paper employs an integrative conceptual approach in the tradition of model-paper scholarship (Jaakkola, 2020). It synthesises five bodies of literature: AI in education and AI literacy frameworks, critical scholarship on educational technology, foundational work on curiosity, dual-process cognition, and adaptive expertise, constructive-developmental theory, and research on question-asking in learning and authentic assessment. Sources were identified through targeted citation tracing supplemented by database searches. Theoretical resources are justified at the point of use rather than methodologically up front.FindingsThe analysis demonstrates that material access and algorithmic fairness, while necessary, do not on their own account for the learner-side capacities that determine whether AI is engaged with critically or passively. It argues that inquiry-driven agency, the capacity to ask productive questions of AI rather than receive its answers, is a worthwhile object of pedagogical design. The Student-AI Learning Trajectory Model presents this development as a four-phase curricular sequence (AI as Tool, Machine, Object, and Partner), with progression understood as a curricular design decision made by educators rather than a developmental trigger inside the learner.Research limitations/implicationsThe model is conceptual rather than empirically validated, and its applicability to younger learners, neurodiverse learners, and low-literacy contexts has not been fully worked out. The paper signposts several lines of empirical inquiry: longitudinal studies of SALT-informed curricula, methodological development of inquiry-sequence assessment instruments drawing on the Objective Structured Clinical Examination tradition, and comparative studies across institutional contexts and learner populations. The contribution is bounded to the pedagogical dimension and does not address the structural reforms that critical scholarship has identified as necessary for genuine educational equity.Practical implicationsThe paper has implications for curriculum design, educator development, and assessment practice. Curricula can sequence student-AI engagement across phases of increasing sophistication rather than treating AI literacy as content to be covered. Educator development requires sustained investment in pedagogical capacity, particularly the modelling of inquiry-driven engagement in front of students. Institutions should treat the redesign of assessment for the AI era as a pedagogical question rather than an integrity question, with inquiry-sequence assessment offering one operational direction. Equity-oriented investment should prioritise educator capacity over the mass distribution of adaptive learning software.Social implicationsThe cultivation of inquiry capacity in AI-mediated learning environments matters for the kinds of learners that schools and universities send into a world increasingly shaped by automated systems. Without deliberate pedagogical design, AI risks becoming a multiplier of existing educational divides rather than a neutralising force. The model locates responsibility with institutions and educators designing learning environments, not with individual students, and identifies the risk of a pedagogical Matthew Effect in which inquiry-driven pedagogy becomes available primarily to already advantaged students absent institutional investment in educator capacity.Originality/valueThe paper brings together three conversations that have developed largely in parallel: AI literacy as a set of competencies, evaluative judgement as a response to generative AI's destabilisation of assessment, and curiosity as a driver of self-directed learning. Its distinctive contribution is the articulation of phase progression as a curricular sequence informed by a cyclical model of inquiry learning, complemented by an operational assessment counterpart. The paper offers vocabulary and tools for educators, curriculum designers, and assessment scholars responding to the question of how AI should be integrated into the work of cultivating independent learners.
Blended Learning (BL) has become a widely adopted educational model, merging traditional face-to-face instruction with online learning tools and techniques. It offers many advantages, including flexibility and structured interaction. During the pandemic and post-pandemic era, there is a requirement to deliver robust, adaptable, and context-sensitive instructional design frameworks that will lead to a paradigm shift in the Teaching and Learning System (TLS). Despite the availability of established models like Analysis, Design, Development, Implementation, and Evaluation (ADDIE) and conceptual frameworks like Technological Pedagogical Content Knowledge (TPACK), there may be substantial research gaps available to fully support the dynamic demands of BL approaches. This research proposes a comprehensive instructional design framework integrating the existing available standards and models for enhanced learning experience, promoting motivation, engagement, and reflection among learners. The framework envisions a modular digital platform where TPACK informs the design phase, and Self-Regulated Learning (SRL) strategies are embedded throughout development, implementation, and evaluation. This proposed model, currently in the early stage of development, aims to empower instructors to design more coherent, adaptive, and learner- centered BL experiences.
