Reflexive AI literacies: conceptualizing a model for learners’ critical engagement with generative AI

Despite the rapid worldwide increase in students’ use of AI, reflexivity remains underdeveloped in both research and practice. To address this gap, we analyse two case studies involving students in the United States and in Denmark, who utilized generative AI tools for creative tasks involving writing and visual presentations. Our findings reveal a significant gap between students’ recognition of ethical issues and their actual practices, highlighting burgeoning but superficial agency and limited reflexive engagement. Drawing on research on videogaming and literacies, we propose a new model for reflexive AI literacies based on three interconnected concepts: design and context, cybernetic relationships, and I-positions. This model emphasizes recursive, socially situated meaning making facilitated by reflexive pedagogy. Advocating for greater attention to reflexivity, we use this model to offer entry points for critically evaluating AI outputs, understanding biases, and assuming moral and ethical responsibility–key components for fostering sophisticated, responsible AI literacies essential for future education and society.