South Korea threatens criminal charges over AI glasses in exams: schools must redesign assessment

South Korea's education ministry announced on 29 October that students caught cheating with AI-powered glasses in the College Scholastic Ability Test (CSAT), the national university entrance exam, will face criminal charges for obstructing official duties by fraudulent means. The ministry also wants to extend the ban for cheaters from one year to up to three, through amendments to the Higher Education Act, as reported by The Korea Times, The Korea JoongAng Daily and SBS News. This year's exam is on 19 November.

The ministry will station one device specialist at each of the 1,312 test sites, set up emergency teams to handle suspected cases, and review radio-frequency detectors that might be ready for use as early as next year. It will also recommend that universities apply the same rules to their own entrance exams. Two applicants were reportedly disqualified at university essay exams earlier this month for wearing AI glasses, one at the University of Seoul on 3 October and one at Yonsei University on 10 October.

This matters well beyond Korea. Smart glasses look like ordinary glasses, and SBS reported that metal detectors cannot tell the two apart. A student's whole future can ride on a single exam sitting, so the pressure to cheat is high, and the penalty now moves from the exam hall to the courtroom. Any country that still decides places, grades or qualifications on one closed-book paper faces the same question, and so does any parent whose child is about to sit one.

Dan Fitzpatrick's view is that the ministry is right to protect a fair exam, and also that this is the start of a race it cannot win. Every new device that can see, listen and answer will need a new rule, a new inspector and a new detector. Banning tools protects one sitting. It does nothing to show what a young person can actually do, and criminal charges against teenagers are a heavy price for a system that stays easy to game.

For school and college leaders, the practical steps are these. Ask which of your assessments only work if no one has AI, and add a live element to them: a short conversation, a drawn-out reasoning task, or work explained aloud. Say plainly to pupils and parents which tools are allowed for which tasks, in writing, before the exam season. Do not rely on detection software or device bans alone. Keep coursework where AI use is permitted and the student's own thinking is visible in the process.

Dan Fitzpatrick works with school leaders and teachers on AI in classroom practice, assessment and policy, and can speak to what schools should change this term.

Dan Fitzpatrick is available for interview and comment on this story. Available today for live or pre-recorded radio, TV, podcast and print interviews, with a broadcast-quality remote setup for live or pre-recorded down-the-line interviews. Contact: dan@theaieducator.io