From annotation to reflection: how participatory AI training enhances critical thinking

Abstract This article challenges widespread concerns that artificial intelligence (AI) in higher education undermines critical thinking. Instead, we show how the process of annotating academic texts to be later used for training AI tools can create new opportunities for reflection and learning. Drawing on an interdisciplinary teaching and research project in International Relations (IR)—a subfield of political science—we demonstrate how students became active participants in developing an AI tool intended to trace the migration of arguments across scholarly texts. Through their involvement in creating annotation schemes and in annotating academic journal articles, students engaged with the underlying assumptions, ambiguities, and value judgments that shape AI systems. Their reflections reveal that annotation is not just a technical task but a socially situated practice that brings questions of language, culture, and power to the surface. Working within a community of practice, students learned to challenge established categories, interrogate expert knowledge, and consider whose perspectives are included or excluded in the development of AI tools. Rather than adopting AI uncritically, they developed the capacity to ask how it works, what it assumes, and what it leaves out. Our findings suggest that embedding AI into participatory, reflexive pedagogies can support the development of critical thinking by making the cultural and political dimensions of technology more visible. We argue that such an approach not only strengthens students' analytical skills but also empowers them to become more thoughtful and responsible users and shapers of AI in higher education.