One of the criticisms instructors often level at the use of generative AI in the classroom is that it subverts students’ ability to “think critically.” This presentation introduces Four Phases of Thinking with AI—Passive, Parallel, Deep, and Generative Thinking—which models the relationship between human cognition and artificial intelligence. Each phase represents a balance between cognitive offloading and students’ intellectual agency.

In Passive Thinking, AI functions as a worker, performing tasks that require little human reasoning. Parallel Thinking invites co-exploration, where humans and AI think alongside each other. Deep Thinking transforms AI into a co-reasoner, supporting synthesis, reflection, and ethical evaluation. Finally, Generative Thinking advances into meta-cognition, as learners design or customize AI for responsible decision-making.

Through this progression, instructors can scaffold assignments that develop both AI literacy and higher-order thinking. The model aligns with Bloom’s Taxonomy and Webb’s Depth of Knowledge, highlighting how cognitive complexity increases as reliance on AI decreases. Ultimately, this framework repositions AI as an intellectual collaborator, one that challenges students not to think less but to think more deliberately and creatively.