Proceedings of the International Group for the Psychology of Mathematics Education · Volume 1
Busra Caylan-Ergene, Ozkan Ergene, Igor’ Kontorovich
Sakarya University
⬇ Download PDFThis paper examines how pre-service teachers interact with AI-generated responses, with attention to perceived credibility of AI outputs and their impact on teachers’ solutions. Twelve participants completed a task on representing eπ on the number line and subsequently interacted with a specially designed GPT that deliberately produced erroneous responses. GPT responses included hallucinations primarily involving fabricated concepts, principles, and academic sources. Only one participant noted and questioned the validity of AI-generated responses, whereas others endorsed them as shared. Delayed recognition, identified in the post-interaction written reflections when participants revisited their earlier GPT dialogues, was linked to limited subject-matter confidence, the professional tone of AI explanations, and trust in AI.
Artifical IntelligenceHallucinationsPre-service teachers