Preserving Cognitive Struggle in the Classroom

Mary Lynn Reed, School of Mathematics and Statistics
Mary Lynn Reed, Professor, School of Mathematics

When I received thirty homework submissions that were nearly identical, I realized that trying to outsmart generative AI is a losing battle. This realization forced me to restructure my courses to protect the cognitive struggle that is essential to genuine learning. I eliminated online quizzes, stopped grading take-home work entirely, and moved labs and assessments in-person where I can observe the learning process directly. I also use AI’s subtle mistakes as learning opportunities, presenting incorrect generative AI outputs in class and asking students to identify exactly where the reasoning failed. Ultimately, my goal is to teach students to admit uncertainty and explain their thinking, ensuring they build the foundational knowledge necessary to critically evaluate these tools rather than blindly copying answers.

To implement Mary Lynn’s strategy for preserving cognitive struggle in the classroom, you can follow these five steps:

Stop grading standard take-home assignments and eliminate online quizzes, as these formats are highly vulnerable to automated shortcuts that bypass student engagement.

Move quizzes, coding labs, and key exercises into the physical classroom so you can observe the learning process directly and ensure authentic, independent work.

Counteract student disengagement and empty correct answers by requiring students to verbally explain a problem, logic, or programming solution to you or their peers.

Integrate screen-free, handwritten practice into your curriculum as a deliberate visual tool to support critical thinking and step-by-step logic construction.

Find examples where AI generates incorrect reasoning in a subtle way, display them in class, and challenge students to locate where the logic fails to build their analytical credibility and evaluative skills.

  Click here for another example of how RIT faculty have adapted their teaching methods in response to AI.