“Take-home exams are dead.”
A professor used ElevenLabs voice AI to conduct oral exams for 36 students at 42 cents each. Three LLMs graded transcripts independently, then revised after seeing each other’s work. Agreement within one point jumped from 0% to 62%. Students cheat with LLMs on take-home exams. Oral exams force real-time reasoning but don’t scale. Now they do. Fighting AI cheating with AI grading is either brilliant or dystopian.