Beyond Digital Transformation: A Disability-Inclusive Responsible AI Framework for Ghanaian Special Education 

Ghana's recent National Artificial Intelligence Strategy (2023-2033) and the Ministry of Communication, Digital Technology, and Innovation's continued rollout of AI initiatives signal a decisive national commitment to technology-driven development, including in education. Yet global and Ghanaian discussions of artificial intelligence (AI) in schooling have so far emphasized technological capability and workforce readiness far more than the specific needs of learners with disabilities. This conceptual manuscript examines how AI can be responsibly integrated into Ghana's special education system so that digital transformation strengthens, rather than undermines, inclusive education. Drawing on an integrative review of international literature on AI in special education and Ghana-specific research on inclusive education policy and implementation, and grounded in five established theoretical traditions, Universal Design for Learning, the Technology Acceptance Model, Responsible Innovation, Diffusion of Innovation Theory, and Implementation Science, the manuscript develops the Disability-Inclusive Responsible AI Framework for Ghanaian Special Education (DIRAI-Ghana). To our knowledge, this is among the first conceptual frameworks to integrate responsible AI governance, disability inclusion, and Ghana's emerging AI policy landscape specifically for special education. DIRAI-Ghana is best understood as a context-sensitive implementation and governance framework rather than a predictive or technology adoption model, intended to guide policymakers, educators, and technology developers in evaluating and implementing AI in ways that advance disability-inclusive education within Ghana's specific policy and resource environment. The framework is organized around five operational pillars: accessibility by design, teacher empowerment, contextual relevance, equity and infrastructure, and sustainability, all situated within an overarching commitment to responsible and ethical AI governance and presented both as a static structure and as a continuous, cyclical governance process linking inputs, implementation, outcomes, evaluation, and policy revision. The manuscript argues that existing scholarship offers extensive descriptions of AI applications in special education but comparatively limited theoretical guidance on disability-specific implementation in low-resource contexts, and that without deliberate attention to the dimensions the framework identifies, AI adoption risks reproducing the inequalities inclusive education is meant to overcome. It concludes by outlining the framework's implications for Ghanaian policy and practice, its limitations, and a staged agenda for future empirical research.