The AI as Cognitive Mentor: Scaffolding Student Scientific Inquiry Through Dialogue

• In Event: AI-Mediated Learning Environments: Scaffolding Inquiry, Metacognition, and Engagement

Abstract

This study investigates how an AI chatbot, designed as a cognitive mentor, can scaffold 8th-grade students through the foundational yet difficult process of developing scientific research questions (RQs) and hypotheses. Using a qualitative comparative case study of two groups, we analyzed student artifacts and AI-student dialogues. Findings reveal the AI's effectiveness is not uniform; one case achieved substantial improvement and conceptual coherence, while the other demonstrated superficial learning and profound conceptual misalignment. This suggests successful scaffolding depends on the quality of student engagement. We conclude that AI is a powerful supplement but requires a collaborative human-AI model to address deep conceptual gaps.

Authors

• Selcuk Kilinc, Texas A&M University

• Tugce Aldemir, Texas A&M University

• Vivek Sabanwar, Texas A&M University

• Ali Bicer, Texas A&M University

• Donggil Song, Texas A&M University