Research · Research topic

AI and inclusion in schools

Schools are adopting AI quickly. My research asks what that means for students with disabilities, and how leaders can make sure new tools serve every learner.

Leer esta página en español

Overview

Generative AI has moved into schools faster than most policies could follow. Districts are writing guidance, teachers are trying tools for lesson planning and feedback, and companies are marketing AI products for special education, including tools that draft IEP goals. The promise is real: AI can make text easier to read, turn speech into text and back again, and offer extra practice. The risks are real too.

For students with disabilities, the questions are sharp. A tool that is not accessible to a screen reader excludes some students entirely. A system trained on biased data can misjudge students who communicate or learn differently. A tool that drafts IEP content can speed up paperwork while making plans less individual. And student data protected under privacy and disability law can flow to vendors in ways families never see.

My research studies how school leaders and teachers choose, use and set rules for AI, and whether students with disabilities are part of those decisions. I am also tracking what official AI guidance for schools says, and does not say, about disability, accessibility and inclusion. As a visually impaired researcher, I come to these questions as a user of assistive technology as well as a scholar.

Questions I am pursuing

  1. How do official AI guidance documents for schools address students with disabilities, accessibility and disability law?
  2. How do school and district leaders decide which AI tools to adopt, and whose needs count in those decisions?
  3. What competencies do K–12 leaders need to lead AI adoption in inclusive ways?
  4. How can AI tools support Universal Design for Learning without replacing individual planning for students with IEPs?

Scholars this work builds on

Neil Selwyn and Ben Williamson have argued that claims about AI in education need careful, critical scrutiny, with attention to who benefits and who bears the risks (Selwyn, 2019; Williamson & Eynon, 2020). Meredith Whittaker and colleagues showed how AI systems can encode and amplify bias against disabled people (Whittaker et al., 2019).

Policy guidance from the U.S. Department of Education (2023) and UNESCO (2023) calls for keeping humans in charge of decisions, protecting data and attending to equity. CAST’s Universal Design for Learning Guidelines (CAST, 2024) offer a framework for judging whether a tool widens or narrows access to learning.

My work on this topic

  • Rath, S. (in press). A generation without a first rung: Gen Z, generative AI and entry-level work in India. NHRDN Journal.
  • Presentation: Developing an AI leadership competency framework for K–12 educational leadership: A mixed-methods study (with J. He, C. R. McCray and F. Stromsland). American Educational Research Association, Toronto, April 2027.
  • Presentation: Bridging AI and transformational leadership in K–12 education (with J. He, C. R. McCray and F. Stromsland). University Council for Educational Administration, Pittsburgh, November 2026.
  • Presentation: Trust but verify: Navigating hallucinations and the ethics of AI choice (with H. A. Jackson and S. Yang). Teacher Education Division, Council for Exceptional Children, Denver, November 2026.
  • Presentation: A multi-paradigmatic perspective towards the adoption of artificial intelligence in education (with H. Jackson). American Educational Research Association, Los Angeles, April 2026.
  • Technology and Inclusion in Schools: a tracker of what official rules on AI and other school technology say about students with disabilities.

Common questions

Can AI help students with disabilities?

It can. Text-to-speech, speech-to-text, simplified reading levels, image descriptions and extra practice are all useful supports, and many build on long-standing assistive technology. The benefit depends on whether the tool is accessible, accurate and chosen with the student’s needs in mind.

What are the main risks of AI for students with disabilities?

Inaccessible tools, biased outputs, inaccurate information presented with confidence, less individualized planning, and the sharing of sensitive student data with vendors. Schools should check accessibility and privacy before adoption, not after.

Should AI write IEPs?

AI can help with drafting and organizing, but IDEA requires that an IEP be developed by a team, including the family, based on the individual student. Any AI-drafted content should be reviewed and rewritten by people who know the student, and families should be told when AI was used.

What should school leaders ask before adopting an AI tool?

Is it accessible to students who use assistive technology? What student data does it collect, and where does that data go? Has it been tested with students with disabilities? Who reviews its outputs? Were students, families and special educators part of the decision?

References

CAST. (2024). Universal Design for Learning Guidelines (Version 3.0).

Selwyn, N. (2019). Should robots replace teachers? AI and the future of education. Polity.

UNESCO. (2023). Guidance for generative AI in education and research. UNESCO.

U.S. Department of Education, Office of Educational Technology. (2023). Artificial intelligence and the future of teaching and learning: Insights and recommendations.

Whittaker, M., Alper, M., Bennett, C. L., Hendren, S., Kaziunas, L., Mills, M., Morris, M. R., Rankin, J., Rogers, E., Salas, M., & West, S. M. (2019). Disability, bias, and AI. AI Now Institute.

Williamson, B., & Eynon, R. (2020). Historical threads, missing links, and future directions in AI in education. Learning, Media and Technology, 45(3), 223–235.

All research topics →