TITLE:
The Impact of Generative Artificial Intelligence on College Students’ Computer Thinking in the Task of Complex Computer Programming: Based on Social Cognitive Theory
AUTHORS:
Wenhui Yin
KEYWORDS:
Generative Artificial Intelligence (GAI), Computational Thinking (CT), Social Cognitive Theory, Complex Computer Programming
JOURNAL NAME:
Creative Education,
Vol.17 No.1,
January
22,
2026
ABSTRACT: Generative Artificial Intelligence (GAI) is rapidly reshaping programming education, yet little is known about how college students cognitively, emotionally, and behaviorally engage with AI during complex programming tasks. Grounded in Social Cognitive Theory, this qualitative study explores how the use of GAI influences computational thinking (CT) among 16 undergraduates enrolled in a 15-week programming course supported by AI tools. Through semi-structured interviews and thematic analysis, five core themes emerged: (1) Cognitive Processing, in which AI scaffolds task decomposition, logical reasoning, multi-solution comparison, and iterative debugging while also introducing risks of cognitive substitution; (2) Emotional Experience, characterized by fluctuating self-efficacy, reduced frustration, heightened motivation, and anxiety triggered by AI errors or unpredictability; (3) Behavioral Strategies, where students adopt hybrid human-AI problem-solving pathways involving independent attempts, structured prompting, verification loops, and comparison-based reasoning; (4) Human-AI Collaboration Models, reflecting dynamic role negotiation in which students act as planners and evaluators while AI functions as solver, debugger, and explainer; and (5) Limitations and Risks, including error accumulation, misleading explanations, lack of contextual awareness, and emerging dependency concerns. These findings demonstrate that GAI operates as a cognitive partner that reshapes students’ CT development, self-regulatory behaviors, and learning identities. The study extends Social Cognitive Theory into human-AI interaction contexts by illustrating triadic reciprocity among personal factors, environmental AI feedback, and adaptive behavioral strategies.