The use of generative AI by students with disabilities in higher education

The use of generative AI is controversial in education largely because of its potential impact on academic integrity. Yet some scholars have suggested it could be particularly beneficial for students with disabilities. To date there has been no empirical research to discover how these students use generative AI in academic writing. Informed by a prior interview study and AI-literacy model, we surveyed students regarding their use of generative AI, and gained 124 valid responses from students with disabilities. We identified primary conditions affecting writing such as ADHD, dyslexia, dyspraxia, and autism. The main generative AI used were chatbots, particularly ChatGPT, and rewriting applications. They were used in a wide range of academic writing tasks. Key concerns students with disabilities had included the inaccuracy of AI answers, risks to academic integrity, and subscription cost barriers. Students expressed a strong desire to participate in AI policymaking and for universities to provide generative AI training. The paper concludes with recommendations to address educational disparities and foster inclusivity. • Students with disabilities are using generative AI to overcome barriers they face in academic writing. • The main generative AI used by students with disabilities are chatbots, rewrite applications, and translation software. • The main concerns are the inaccuracy of answers, the risk of academic integrity breaches, and the digital divide created by subscription costs. • There is a strong desire among students to be involved in generative AI policymaking and for universities to provide training on generative AI.