Discipline-Specific AI literacy (DiSAIL): a theoretical framework for situated engagement with generative AI in education

Abstract As generative AI (genAI) tools become increasingly integrated into educational settings, there is a growing need to understand how students engage with these technologies, not merely as users of information, but as participants in new forms of literacy practices. While existing frameworks for AI literacy often emphasise general competencies and technical understanding, they tend to overlook the disciplinary, linguistic, and epistemic dimensions of AI use in education. This article introduces the Discipline-Specific AI Literacy (DiSAIL) model, a conceptual framework that integrates perspectives from linguistics, literacy studies, and learning theory. The DiSAIL model derives from a model for technological literacy and consists of two interrelated facets: potential for DiSAIL , including disciplinary literacy, AI competencies, and AI acceptance, and enactment of DiSAIL , which captures how learners recognise needs, articulate problems, contribute to disciplinary reasoning, and analyse the consequences of AI use. By framing genAI interaction as a situated literacy practice, DiSAIL shifts the focus from tool adoption to pedagogical agency and from generic digital skills to meaningful disciplinary participation. The model offers a foundation for both empirical research and instructional design, supporting educators and researchers in developing AI-integrated learning environments that are epistemologically grounded, ethically aware, and pedagogically aligned.