Research
My doctoral thesis, my publications and the projects I contribute to.
Doctoral thesis in progress
Surface properties of AI-generated text and adolescents' metacognitive regulation
University of Bern · students aged 13 to 15 · MASRL model (Efklides)
Answers from a language model arrive well written, in a confident tone, with hardly any sign of doubt. My thesis examines what these surface properties do to the judgement of students aged 13 to 15: does fluent, assertive text lull the monitoring they exercise over their own understanding?
The work builds on theories of self-regulated learning, in particular Anastasia Efklides's MASRL model, which separates what learners know about their learning from what they experience in the moment. It is this second level, metacognitive experience, that fluent text may short-circuit.
The practical stake: knowing when and how to teach vigilance in front of answers that all look right.
- 2026
Tschupp, G., & Barras, H. (2026). Évaluer un travail de bachelor rédigé par l'IA : réflexions critiques pour les pratiques pédagogiques des enseignants du supérieur [Communication]. Colloque AUPTIC, Lausanne, 5-6 janvier 2026. FREDI
Conference paper (in French) presented with Hervé Barras at the AUPTIC 2026 conference in Lausanne: assessing a bachelor's thesis written by AI.
- 2025
Tschupp, G., & Barras, H. (2025). Rédaction académique assistée par IA générative. Dans A. Baillifard & H. Carbonel (dir.), L'éducation à l'épreuve de l'intelligence artificielle (p. 91-135). Épistémé. FREDI
Book chapter (in French), written with Hervé Barras, on academic writing assisted by generative AI.
- 2022
Class, B., Balli, F., Enkerli, A., Favre, S., Gillet, D., Hensler, I., Mboa Nkoudou, T. H., Notari, M., Scherrer, Y., & Tschupp, G. (2022). Final report - A Roadmap for Open Education in Switzerland [Rapport]. Université de Genève. https://doi.org/10.5281/zenodo.7002915 DOI
Collective report proposing a roadmap for open education in Swiss higher education.