From learner autonomy to hybrid autonomy: A research agenda

Abstract

This conceptual article proposes the construct of hybrid autonomy to describe how control, agency, and reflection are increasingly distributed between learners and artificial intelligence systems in language learning. While learner autonomy has traditionally been understood as the capacity to take charge of one's own learning, the growing integration of generative AI, conversational agents, adaptive systems, and learning analytics reshapes how autonomous learning is manifested. Rather than viewing AI as either enhancing or diminishing autonomy, the article argues that AI transforms the ecology of control by redistributing cognitive, evaluative, and regulatory functions across human and non-human actors. Drawing on research in learner autonomy, computer-assisted language learning, distributed cognition, hybrid intelligence, and AI-mediated language learning, this article synthesizes recent empirical and theoretical developments to examine how AI influences the cognitive, affective, social, and technological dimensions of autonomy. Particular attention is paid to emerging tensions between empowerment and dependency, personalization and standardization, and agency and algorithmic guidance. The article further examines implications for pedagogy, assessment, teacher roles, and institutional policy, arguing that autonomy in AI-rich environments depends not simply on access to intelligent tools but on learners' ability to critically negotiate, interpret, and regulate their interactions with them. The article concludes by outlining a research agenda for investigating hybrid autonomy in language education, including directions for studying shared agency, evaluative judgement, AI-supported self-regulation, learner identity, and the distribution of cognitive labour in human-AI interaction.