Two-year law course experiment finds the no-AI group finished last both years

A two-year classroom experiment at Vrije Universiteit Amsterdam found that students barred from using AI performed worst, in both years, on a legal drafting task. Law professor Thibault Schrepel randomly split students in his 'Law of AI' course into three groups. Everyone had the same assignment: in teams of four or five, spend 20 minutes improving a provision of the EU AI Act, graded on substance, clarity, proportionality, and innovation. The first group could not use ChatGPT. The second received ChatGPT-generated revision suggestions embedded directly in the text and could keep using the tool, but got no guidance on how. The third group received hands-on training in legal prompt engineering and in checking AI suggestions for consistency and accuracy. All students then sat the same exams, a multiple-choice test and a take-home exam that required revising another AI Act provision. The experiment ran in 2024 with 66 students and was repeated in 2025 with 164. The no-AI group ran out of ideas quickly and mostly made minor edits. Schrepel writes that he had assumed AI without guidance would do more harm than good, and concludes, 'I was wrong.' The sample is limited to one course at one university, but the results bear on universities weighing outright bans against structured instruction in AI use.