Project Leader: Dr Michele Pizzochero, Department of Physics

The rapid advancement of artificial intelligence (AI) is transforming various sectors across our digital society, including education. Traditionally, AI's role has been regarded as a tool for automating tasks and enhancing efficiency. However, the interaction between AI and human learning is increasingly bidirectional: not only do humans teach machines, but they also learn from them. Despite the growing integration of AI into educational contexts, there is a notable gap concerning the characterization of the learning experience delivered by AI. This project seeks to develop a general framework to assess analogies and differences between learning experiences delivered by a lecturer compared to a large-language model impersonating one, using physics as a case study.

This study will employ a series of Turing-like tests adapted for educational contexts. The research will involve undergraduate students in Physics and Natural Sciences at the University of Bath. A diverse sample will be selected to ensure variability in prior knowledge and learning preferences.

Report and finding due Spring 2026