While creating “AI-proof” assignments may be more wishful thinking than reality, there are approaches that can reduce the likelihood of AI misuse. The strategies below are commonly discussed as ways to do so:

• Moving coursework, including assessments, into the classroom. This can include both lower-stakes formative learning activities and final papers or exams. While this is an effective way to prohibit AI, it has trade-offs such as making the assessment less inclusive for neurodivergent students and forsaking longer-term assignments that benefit from revision over a period of time.

• Designing assignments that enhance students' intrinsic motivation. This is the focus of the majority of this model; it involves designing assignments that are relevant and meaningful to students such that they want to make an effort and do original work.

• Creating friction to using AI. This approach involves adding components to asssignments that make it harder or less efficient to use AI to complete the assignment. Examples include process-based assignments where students submit and show their work at various stages or screen record themselves working through a problem.

• Reducing grading pressure with strategies such as using low-stakes assessments, grading on process, flexible deadlines, and using alternative grading approaches like specifications grading.