AI in Higher Education Marginalizes Disabled Knowledges, Study Finds
A new study published on arXiv (2605.26769) argues that generative artificial intelligence systems in higher education reinforce epistemic coloniality by marginalizing non-hegemonic knowledges. Drawing on educational sciences, critical technology studies, and disability studies, the research shows that training datasets remain predominantly Anglophone and Western-centric. The situation of persons with disabilities illustrates this phenomenon: technological architectures confine them to reductive stereotypes or exclude them from design processes, causing double marginalization. The article examines whether hybridization between researcher and machine could preserve epistemic plurality, while acknowledging structural limitations.
Key facts
- Generative AI restructures scientific knowledge production in higher education.
- AI systems are not neutral and marginalize non-hegemonic epistemologies.
- Training datasets are predominantly Anglophone and Western-centric.
- Persons with disabilities face reductive stereotypes or exclusion from AI design.
- The study draws on educational sciences, critical technology studies, and disability studies.
- The article is published on arXiv with identifier 2605.26769.
- It questions whether researcher-machine hybridization can preserve epistemic plurality.
- Structural limitations inherent in AI architectures are acknowledged.
Entities
Institutions
- arXiv