Discussion

AI cheating cases expose a growing challenge for South Korea’s qualification exams

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#14353

South Korea’s state qualification exams have recorded at least eight cases of AI-assisted cheating since 2024, and the government is now increasing funding for prevention. The cases show how quickly a smartphone and a chatbot can turn an ordinary exam into a test of whether invigilation rules still work.

Watch Desk analysis

What happened

The Korea JoongAng Daily reports that agency data obtained by lawmaker Kim Wi-sang showed at least eight cases had been detected by the Human Resources Development Service of Korea as of last month. In one case, from August 2024, a candidate photographed a computer-based exam with a smartphone and used an AI application to answer questions. Prosecutors later suspended the indictment.

The paper also describes cases involving ChatGPT and Gemini in certification exams, and reports that a candidate caught using Gemini during a practical exam in Gwangju was fined 3 million won ($2,250). After the top scorer in a tax-accountant qualification exam admitted to using AI earlier this month, lawmakers urged the government to increase anti-cheating funding by 18.2 billion won. Those details and figures are reported by the Korea JoongAng Daily, which says the government had proposed cutting the exams’ budget for next year before the increase.

Why it matters

These are not just rules for one exam room. Qualification tests help determine who can practise skilled work, so confidence in the result matters to candidates and the public. The reported cases also make the practical problem plain: when candidates can consult AI on a phone, preventing cheating is not simply a matter of telling people to leave their devices in their bags.

The report gives a timeline and examples, but it does not establish how widespread undetected cheating may be. Nor does the existence of cases prove that AI alone caused the government’s initial budget proposal or the delayed response described by the paper.

Our read

The useful signal is the combination of repeated incidents and a funding change, not the novelty of someone asking a chatbot for help. Exam administrators need rules and supervision that address phones and AI tools in practice, while making sure legitimate candidates are not left guessing what is permitted. More enforcement is not automatically better enforcement; it has to be clear, workable and fair.

What to watch

  • How the increased anti-cheating funding will be used.
  • Whether exam administrators publish updated rules for phones and AI tools.
  • Whether agencies release further figures on detected cases and prevention measures.

Discussion spark: Should qualification exams respond to AI-assisted cheating mainly with tighter phone controls, or redesign assessments so a chatbot cannot provide the decisive advantage?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.