The Dizzy Spells of AI: Enacting Intellectual Independence in AI Literacy and AI-Mediated Learning

Abstract The integration of artificial intelligence (AI) into education and learning risks leading to intellectual dependence among users of AI tools. Meanwhile, intellectual dependence remains underexplored in AI-mediated learning and is addressed ambiguously in AI literacy programmes, which claim to foster critical thinking (CT) and critical consciousness (CC) among learners, helping them move from naïve to more critical ways of thinking about and using AI. However, such approaches to AI literacy often presume and reinforce a hierarchy of thought: teachers are seen as critical thinkers about AI, while students are viewed as less so. This hierarchy arguably leads to intellectual dependence of the latter on the former. In this paper, we problematise AI literacy programmes that embrace CT and CC as educational objectives, since, as we argue, they remedy intellectual dependence on AI tools (e.g. ChatGPT, Claude…) by subjecting AI users/learners to another intellectual dependence, but on AI teachers. By drawing on Jacques Rancière’s ideas of universal intellectual equality, we argue that enacting intellectual independence is not an educational or learning goal to be achieved in the future, but a starting point—a condition necessary for teaching and learning about and with AI. Hence, we apply this assumption of equality of intelligence to both AI literacy and AI-mediated learning, and illustrate this application with two practical scenarios: (1) AI-centred pedagogy (thing-centred AI literacy) and (2) ignorant AI-tutors (thing-centred AI-mediated learning). Both approaches aim to reduce dependence on human teachers in AI literacy or on AI tutors in AI-mediated learning. Ultimately, this paper presents design principles for these two solutions, offers a philosophical premise for rethinking education, teaching, and learning in the age of AI, and concludes with contributions, limitations, and a research agenda.