Use of Artificial Intelligence to Identify Invisible Barriers in Inclusive Education: An Exploratory Bibliographic Review (2020-2026)

Abstract

Inclusive education faces challenges that go beyond visible barriers related to physical access and infrastructure, involving subtle, recurrent, and often naturalized forms of exclusion. These barriers particularly affect people with disabilities, deaf students, autistic students, atypical learners, and other minority groups, and may appear as low participation, dropout, difficulties in retention, institutional silencing, and the lack of adequate pedagogical responses. This article aims to analyze recent scientific literature on Artificial Intelligence, inclusion, and education, discussing the possibilities and limits of intelligent systems in identifying invisible barriers. This is an exploratory bibliographic review inspired by the PRISMA protocol, conducted through the CAPES Journals Portal using CAFe/IFPR access. The initial search retrieved 147 records. After applying filters for publication period (2020-2026) and peer review, 55 records remained, of which 24 were considered relevant to the synthesis: 9 directly related to the core theme of Artificial Intelligence, inclusion, and people with disabilities, and 15 used as theoretical support. The findings suggest that AI can serve as a tool for reading institutional patterns of exclusion, provided that it is connected to ethics, data protection, community participation, and human validation. The study concludes that AI should not replace educational listening, but it can support the identification of less visible inequalities and strengthen policies for student retention, accessibility, and educational justice.