Rethinking Student Critical Thinking: How AI Chatbot Interactions Shape Prompting and Critical Thinking (Stage 1, 5:05 PM)

• In Event: e-Lightning Ed-Talk Session 9 - Stage 1

Abstract

In the era of generative AI, learners must not only craft effective prompts but also engage in real-time critical thinking (CT). This study explores how students’ CT performance levels shape their prompting strategies and CT behaviors during AI-assisted instructional design. Across a six-week task, 416 prompts from nine students were analyzed using the CLEAR framework and Facione’s CT model. High-CT students employed more adaptive and reflective prompting, while low-CT students favored concise and explicit approaches. CT behaviors also diverged: high performers engaged more in evaluation and inference, whereas low performers emphasized analysis and interpretation. Association analysis revealed a strong co-absence of inference, explanation, and logic. Findings inform differentiated scaffolding to support equitable AI literacy and future-oriented CT development.

Authors

• Chen Yang, The Education University of Hong Kong

• Shurui Bai, The Education University of Hong Kong