Does structured training or institutional guidance change how AI dependency affects students' critical thinking?

Why the same AI tool can help one student think and let another stop thinking

The difference seems to lie less in the tool than in what surrounds its use. A scoping review of generative AI in education reports that tools such as ChatGPT support critical thinking when they are embedded in structured learning activities, rather than simply handed over as answer machines [3]. That framing matters: the same chatbot can be a thinking partner in a guided task or a shortcut that ends thought before it starts.

AI dependency itself does not produce a single, uniform effect. One study examining the link between AI dependency and students' critical thinking disposition found that dependency is associated with different patterns of that disposition, and it points to cognitive fatigue and engagement as factors in the relationship [1]. In plain terms, heavy reliance may drain the mental energy that critical thinking requires, or it may leave students disengaged from the work — but the study's abstract frames this as an association, not a proven cause.

What institutional guidance actually changes

A study of AI-saturated classrooms argues that cognitive dependency is a pedagogical, neurocognitive, ethical, and institutional problem at once, and that responsibility for AI dependency cannot be placed only on students [2]. The implication is that guidance works by changing the classroom conditions — how tasks are set, what access students have, and what habits of prompt use are taught — rather than by warning students to try harder.

Teacher guidance appears in a related line of work as a specific lever. Research on AI-assisted learning notes that the spread of AI has outpaced many institutions' ability to keep up, and it highlights teacher guidance and students' ability to verify AI-generated content as central concerns [5]. Verification is the practical skill at stake: a student who checks an AI answer is doing something different from one who copies it, and guidance is what makes that checking routine.

Dependency is only one thread in a larger picture

Other factors shape critical thinking alongside dependency. A study of undergraduates in Indonesian higher education examined AI dependency together with ethical awareness and awareness of algorithmic bias, and it notes that institutions worldwide have official guidelines for integrating AI [4]. That combination suggests guidance may matter not only for how much students rely on AI but for whether they recognize its biases and ethical limits — though the abstract does not establish which factor drives the outcome.

Put together, these studies are consistent with a conditional answer: structured training and institutional guidance look like the conditions that tilt AI use toward critical thinking, while unstructured dependency tilts it toward fatigue, disengagement, or uncritical copying [2][3][5]. The strength of that conclusion is limited, because the evidence comes from abstracts and reviews rather than a direct experiment comparing trained and untrained students.

Sources used in this answer

Related research questions