Students' Epistemological Beliefs and their Chatbot Preferences in AI-mediated Physics Learning

Synthesis: Sirnoorkar & Mamidpalliwar (2026) investigate the association between introductory physics students' preferences for chatbot behavior and their epistemological beliefs, using a custom online waves module with simulations integrated with a chatbot. Preferences were captured through three options (guided-inquiry, direct answer, and a combination); beliefs via the EBAPS survey. Students who preferred chatbots that initially engage in guided-inquiry but provide answers when explicitly sought ('Combination') demonstrated more sophisticated epistemological beliefs than those preferring answer-providing chatbots, though no association was found between EBAPS total scores for guided-inquiry vs answer-oriented preferences, and differences did not survive Bonferroni adjustment. The work connects to Physics Education, Pedagogical Agent, and Socratic Method research.

Chatbot Preference and Epistemological Beliefs

Generative AI promises to transform higher education. This study asks whether introductory students' preferences for chatbot behavior are associated with their epistemological beliefs surrounding physics, in a custom-built online module on waves containing simulations integrated with a chatbot.

Design

Chatbot preferences were captured through three provided options — guided-inquiry, direct answer, and a combination of inquiry and answer. Epistemological beliefs were captured through the standardized Epistemological Beliefs Assessment for Physical Sciences (EBAPS) survey.

Findings

Students who preferred chatbots that initially engage them in guided-inquiry but provide answers when explicitly sought ('Combination') demonstrated more sophisticated epistemological beliefs than those who preferred answer-providing chatbots. No association was observed between EBAPS total scores among guided-inquiry vs answer-oriented preference groups, and differences did not remain statistically significant after Bonferroni adjustment. Proceedings of the 2026 Physics Education Research Conference.