"I Would Share It, But..." Exploring Ways to Optimize the Privacy-Personalization Trade-Off in Intelligent Tutoring Systems

Adrienn Toth, Neele Roch, Noé Zufferey, Verena Zimmermann

Abstract: Personalized learning is increasingly improving through AI-enhanced intelligent tutoring systems (ITS). However, ethical and privacy aspects of ITS, such as the privacy-personalization trade-off, are under-researched. We conducted an interview study with \(N=32\) university students and found that students were not largely concerned about privacy as such but implications of data collection for social aspects, their education, and society at large. Students preferred ITS to supplement rather than replace human teachers in certain tasks, leveraging the benefits of both. We provide recommendations on addressing the identified concerns for ITS design and successful integration into a curriculum, for which AI literacy and student autonomy emerge as crucial factors.

External IDs:dblp:conf/aied/TothRZZ25