The rapid development of generative AI is reshaping core assumptions about writing, learning, and independence in higher education [1]. The project is grounded in an academic literacies perspective, which views writing as a socially and institutionally situated practice shaped by norms, values, and disciplinary expectations. Complementing this, Science and Technology Studies (STS) is used to conceptualize AI as a sociotechnical actor that actively participates in reshaping assumptions about authorship, independence, and pedagogical practice. AI literacy serves as an analytic lens, capturing how students interpret, critique, and negotiate the role of AI technologies as part of their emerging professional identities as teachers.

The study addresses the following research question: How is writing constructed and understood in relation to AI literacy and independence by student teachers, and what forms of writing actions, cognitive processes, and ideals of independence are expressed in the teacher’s task instructions and assessment criteria?

Methodologically, the project employs classroom observation and textual analysis. Observation of one English-language lesson in teacher education provides insight into how students discuss and negotiate the role of AI in writing and how they position themselves in relation to expectations of independence. This is complemented by analysis of written task instructions and assessment documents, which reveals how institutional ideals surrounding writing, cognition, and autonomy are communicated.

Preliminary findings indicate a clear tension between institutional ideals of independent authorship and students’ uncertainty about the role AI should play in academic writing, in line with earlier research [2].

Students expressed hesitation and ambiguity regarding what forms of AI use are permissible and how independence ought to be interpreted in an AI-mediated environment. This uncertainty generates an ongoing negotiation as students attempt to align institutional expectations with evolving technological possibilities.

Analysis of the teacher’s task instructions and assessment criteria shows that independence, originality, and reflective engagement are strongly emphasized, and that the use of AI tools is (currently) explicitly prohibited. Expectations of cognitive effort and individual responsibility thus remain firmly anchored in pre-AI models of writing, creating a pronounced misalignment between institutional framings and the technological realities students encounter.

The study contributes to international research on writing, AI literacy, and teacher education by offering empirical insights into how student teachers negotiate authorship and independence in an AI-saturated context. The findings underscore the urgency of developing responsible, pedagogically grounded approaches to AI in teacher education—approaches that acknowledge both the transformative potential and the epistemic challenges introduced by generative AI [3].

References:[1] Kasneci, E. et al. (2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and individual differences, 103, p.102274.[2] Josefsson, P., Hüttenrauch, H., & Jacobsson, M. (2023). AI-driven plagiarism in higher education: Impact on learning and examination. Forskning om högre utbildning, 54–55.[3] Miao, F. and Shiohira, K., 2024. AI competency framework for students. UNESCO Publishing.Keywords: AI, academic writing, higher education, teacher education.