Reframing AI governance in education: insights from the social model of disability

Artificial Intelligence (AI) offers potential benefits for students with hidden disabilities like dyslexia but also risks exacerbating inequities due to digital poverty and commercial bias. This paper explores how commercial interests shape access to AI tools for students with dyslexia. Using the social model of disability, we critique how commercial influences shape access to assistive AI technologies through four speculative scenarios. The first scenario highlights exploitative commercial practices driving inequities, the second inadequate institutional policies failing to address digital poverty, the third insufficient accountability among developers prioritizing profit over fairness, and fourth weak government oversight lacking safeguards against commercial bias. We advocate for inclusive AI governance that recognizes the interconnected dynamics of human, technological, and commercial power to promote equitable access for students with disabilities, drawing on the social model of disability and Australia's Voluntary AI Safety Standards. While situated in the Australian context, the analysis contributes to global discussions on the role of AI in creating accessible learning environments. We argue that inclusive AI governance must address the entangled roles of human, computational, and commercial power to ensure accessibility for students with disabilities.