AI-proof exams are becoming a growing focus as universities rethink how students should be assessed in the age of generative AI. A recent Nature report highlights how academics are experimenting with new approaches as AI tools become increasingly capable of completing conventional assignments and examinations. Some educators are designing assessments that emphasise reasoning, practical work and skills that are harder for AI systems to reproduce convincingly. Others are examining ways to identify inappropriate AI assistance. The discussion reflects a wider shift in higher education from simply restricting technology to redesigning assessment itself. For students, the trend could mean greater emphasis on oral examinations, project-based learning, classroom problem-solving, practical demonstrations and personalised assignments. Universities are also facing questions about fairness, accessibility and academic integrity while introducing AI-related assessment methods.
AI-Proof Exams: How Universities Are Rethinking Assessments
