Current research on generative AI use in education has predominantly focused on students with dominant middle-class identities in higher education (ΗΕ) settings. Vocational education and training (VET) remains an under-researched field, mirroring its lower social standing in Greece and elsewhere. Likewise, the potential of ChatGPT in fostering inclusive education for disabled migrant students has received little attention. Drawing on intersectionality, this study explores the ChatGPT learning experiences of six of these students, who are served yet unserved by the Greek upper secondary VET system. Emerging themes of in-depth interviews included: i) ChatGPT as an in-class participation facilitator, ii) ChatGPT as on-call support and, iii) Drawbacks. Findings explicate how intersections of cultural positioning, disability, class, age and gender shape marginalized students’ learning experiences with ChatGPT in the context of contemporary neoliberal educational policies in Greece. Attention to how diverse social positionings shape ChatGPT-assisted learning can add depth to the current discourse on contextualizing generative AI assisted pedagogies tuned to diverse local contexts and students’ needs.