AI Could Transform Support for Students with Learning Difficulties, If Governance Catches Up

Artificial intelligence is quietly rewriting the rules of higher education. Adaptive learning platforms adjust the pace of a lecture to each student, intelligent tutoring systems offer round-the-clock feedback, and learning analytics dashboards flag students at risk of falling behind long before an exam reveals the problem. For students with learning difficulties such as dyslexia, dyscalculia, dysgraphia and attention-deficit/hyperactivity disorder, these technologies promise something that decades of accommodation policies have struggled to deliver: support that is built into the fabric of teaching rather than bolted on after a formal diagnosis. Yet a new conceptual study from Kuwait argues that without deliberate governance, the same tools could just as easily deepen the inequalities they are meant to close.

The research, published in Frontiers in Education by Abdullah F. Alenezi of the Public Authority for Applied Education and Training in Kuwait City, proposes a governance framework called the AI- and Learning-Analytics-Enabled Inclusive Support Framework for Learning Difficulties, or ISF-LD. Its central claim is deceptively simple: the question facing universities is no longer whether to adopt AI, but how to adopt it responsibly so that it promotes equitable and inclusive learning. The framework was developed through a structured conceptual methodology that synthesised four bodies of evidence, including research on learning difficulties and disability inclusion, scholarship on AI and responsible technology, national and international policy documents, and established educational and disability theories.

The urgency of the problem is rooted in a paradox familiar to disability researchers worldwide. Many learning difficulties are invisible disabilities, and students frequently choose not to disclose them because of stigma, fear of discrimination or worries about how faculty and peers will perceive them. In Kuwait, studies suggest that disclosure decisions are shaped by cultural expectations surrounding academic success, family reputation and societal attitudes towards disability. Because disclosure is typically the gateway to formal accommodations, students who stay silent often progress through their degrees without the specialist support they need. A nationwide survey by the Kuwait Dyslexia Association once estimated dyslexia prevalence at roughly six percent, though the authors caution that this figure has not been updated through recent peer-reviewed epidemiological research and should be treated as indicative.

Kuwait presents a striking case study in policy disconnection. On one side, the country has strengthened disability rights through Law No. 8 of 2010 and its ratification of the United Nations Convention on the Rights of Persons with Disabilities in 2013. On the other, Kuwait Vision 2035 and the national AI strategy have poured investment into digital transformation and AI-enabled education. What the policy analysis reveals, however, is that these agendas have evolved almost entirely in parallel. The national AI strategy sets targets for AI adoption across economic sectors but never mentions accessibility or learning analytics governance, while the Ministry of Education’s AI in education strategy focuses on curriculum integration and teacher training without referencing disability inclusion or universal design. The paper’s review of publicly available documents found no documented intersection between the two domains.

This governance gap is not unique to Kuwait. Across the Gulf Cooperation Council, including Saudi Arabia, the United Arab Emirates, Qatar, Bahrain and Oman, national disability legislation and ambitious AI strategies have advanced simultaneously but with little policy integration. Disability frameworks emphasise participation and reasonable accommodation, while AI strategies emphasise innovation and economic diversification. Few, if any, explicitly consider how AI should be governed to support students with learning difficulties, how ethical learning analytics should operate, or how algorithmic fairness and accessibility should be enforced in higher education. As universities across the region rapidly deploy AI-supported teaching and assessment, the authors warn that existing barriers risk being reproduced at scale rather than dismantled.

To close this gap, the ISF-LD rests on six theoretical pillars. The Social Model of Disability reframes the problem: barriers arise from inaccessible curricula, assessment practices and institutional cultures, not from individual impairment, so technologies must remove barriers rather than reinforce them. Vygotsky’s Sociocultural Theory positions AI tools such as intelligent tutoring and speech recognition as scaffolds within a learner’s zone of proximal development, augmenting rather than replacing human teaching. Self-Determination Theory insists that motivation depends on autonomy, competence and relatedness, meaning predictive systems must strengthen learner agency rather than make opaque, compulsory decisions. Universal Design for Learning pushes accessibility into mainstream curriculum design, reducing dependence on disclosure-based accommodation. Learning analytics ethics and responsible AI principles supply the safeguards of privacy, transparency, fairness, accountability and explainability.

Structurally, the framework comprises five interconnected implementation layers: national policy and governance, institutional leadership and disability services, faculty capability, AI-enabled inclusive pedagogy and assessment, and student self-determination. Crucially, these layers operate through an iterative, non-linear model rather than a top-down sequence. A university might begin by establishing an AI governance committee before national guidance exists, or through faculty-led accessibility initiatives, provided the missing layers are developed later. Student experience feeds back into institutional and national policy through advocacy and participation, ensuring that implementation remains responsive to the people it is meant to serve. Four cross-cutting principles, namely rights-based governance, stigma-aware disclosure, universal accessibility and responsible AI, apply across every layer, from procurement standards to classroom practice.

The technical risks the framework addresses are concrete rather than hypothetical. Algorithmic bias may arise where training data underrepresent students with disabilities or Arabic-language users, meaning AI systems can perform unevenly across linguistic and cultural contexts. Predictive learning analytics can infer disability status from behavioural data without explicit disclosure, raising acute concerns about privacy, informed consent and surveillance. Generative AI can produce hallucinated content, and proprietary platforms create commercial dependence without disability-specific data governance arrangements. The framework’s response includes data minimisation, systems that do not infer disability status without consent, explainable recommendations that students can understand and challenge, and meaningful human oversight of every AI-supported decision affecting learning or accommodation.

What distinguishes the ISF-LD from existing frameworks, the paper argues, is integration. Universal Design for Learning guides curriculum design but says little about governing AI systems; the SHEILA framework supports institutional learning analytics policy but pays little attention to disability; the OECD and UNESCO recommendations establish broad ethical principles without explaining how universities should operationalise them for students with learning difficulties. The new framework deliberately bridges these domains, positioning AI as a governed educational capability rather than a stand-alone innovation, and inclusion as a governance challenge rather than merely a pedagogical or technological one. It also generates five testable research propositions linking policy alignment, institutional capability, inclusive pedagogy, responsible AI practices and student-centred implementation to measurable outcomes such as accessibility, engagement, retention and trust.

The author is candid about the study’s limitations. The framework is conceptual and has not yet been empirically validated; it was developed without direct consultation with students with learning difficulties, which future participatory research should remedy under the disability movement’s principle of nothing about us without us; and the policy analysis relied on publicly available documents that may not capture internal institutional practice. A staged research agenda proposes qualitative case studies in Kuwaiti universities, expert Delphi panels, institutional surveys across the GCC, quasi-experimental studies of AI-enabled pedagogy and longitudinal tracking of student outcomes. Even so, the message travels well beyond the Gulf. As universities everywhere rush to embed AI in teaching and assessment, the Kuwaiti study offers a sobering reminder that technological innovation alone will not deliver educational equity. Sustainable inclusion requires governance models that place disability rights, ethical AI and learner participation at the centre of institutional decision-making, and the ISF-LD offers one blueprint for building them.

Subject of Research: AI and learning analytics governance for inclusive support of students with learning difficulties in Kuwaiti and Gulf higher education

Article Title: Toward an AI- and learning-analytics-enabled inclusive support framework for students with learning difficulties in Kuwait's higher education institutions

Article References: Alenezi, A. F. (2026). Toward an AI- and learning-analytics-enabled inclusive support framework for students with learning difficulties in Kuwait's higher education institutions. Frontiers in Education, 11, Article 1931442. https://doi.org/10.3389/feduc.2026.1931442

Image Credits: AI Generated

DOI: 10.3389/feduc.2026.1931442

Keywords: artificial intelligence, learning analytics, inclusive education, learning difficulties, disability inclusion, higher education, Kuwait, Gulf Cooperation Council, responsible AI, Universal Design for Learning, education policy, algorithmic bias