Knowledge-Based Design Requirements for Generative Social Robots in Higher Education

Synthesis: Vonschallen, Oberle, Schmiedel, and Eyssel (2026) adopt a knowledge-based design perspective to investigate what information tutoring-oriented generative social robots (GSRs) require to function responsibly and effectively in higher education. Recognizing that GSRs powered by large language models enable adaptive, conversational tutoring but introduce risks such as misinformation, overreliance, and privacy violations, they conducted twelve semi-structured interviews with university students and lecturers, identifying twelve design requirements across three knowledge types: self-knowledge, user-knowledge, and (a third type concerning the domain/context).

Key Findings

Study Design & Method

This is a qualitative interview study. The researchers conducted twelve semi-structured interviews with university students and lecturers to identify the knowledge prerequisites that tutoring-oriented generative social robots need to function responsibly and effectively in higher education. Using a knowledge-based design perspective, they analyzed the interviews to derive twelve design requirements organized across three knowledge types (self-knowledge, user-knowledge, and domain/context knowledge), addressing the gap between responsible-AI behavior frameworks and the informational requirements generative agents need.

Implications for AI in Education

The study provides a design foundation for building generative social robots as tutors in Higher Ed, addressing the risks of misinformation, overreliance, and privacy violations. It connects to Generative AI, large language models, tutoring, and responsible AI design. For designers and educators, it specifies the knowledge a tutoring robot must hold (about itself, the user, and the domain) to behave responsibly and effectively, informing the development of trustworthy AI tutoring agents.

Limitations

The findings derive from twelve interviews in a specific higher-education context, so they may not generalize across disciplines, institutions, or learner populations. The design requirements are identified but not yet fully validated through implementation and evaluation. The focus is on the knowledge prerequisites of tutoring robots rather than on broader social-robot design or measured learning outcomes.

Connected Concepts

Connected Articles

Citation

Vonschallen, S., Oberle, D., Schmiedel, T., & Eyssel, F. (2026). Knowledge-based design requirements for generative social robots in higher education. arXiv:2602.12873.