From Trojan horses to AI-proof exams: how professors are tackling students’ AI use

TL;DR - Nature reports that academics worldwide are redesigning assessments in response to student AI use. Approaches range from creating exams that AI tools cannot pass to embedding hidden prompts intended to expose AI-assisted cheating.

• Assessment design is shifting to account for widespread access to AI tools.
• Some professors are developing “AI-proof” exams that current systems struggle to complete.
• Others are inserting concealed AI-directed prompts into assignments as cheating detectors.
• The brief does not provide evidence about these methods’ accuracy, fairness, or effectiveness.