Denmark is turning to one of the oldest tools in education to deal with one of the newest problems: talking.
According to a report from mezha.net, Denmark is requiring students to orally defend their written work as a way to counter cheating with AI. The idea is straightforward. If a chatbot can produce a polished essay in seconds, the essay itself stops being reliable evidence that a student learned anything. A live conversation about that essay is much harder to fake — you either understand the argument you handed in, or you don't.
The story landed on the front page of Hacker News, where it drew 227 points and 107 comments, a sign of how much the question is preoccupying people who work in and around technology.
Beyond the headline policy, the available reporting is thin on specifics, and details such as which levels of education are covered and how the requirement will be implemented are not established by the source material here.
The broader context is easy to state, though. Schools and universities worldwide have spent the past few years trying to detect AI-written text, mostly with disappointing results: detection tools produce false accusations, and the underlying models keep improving. Denmark's approach sidesteps detection entirely. Instead of asking whether a machine wrote the words, it asks whether the student can account for them.
That shift has costs. Oral examinations take staff time, they are harder to scale to large cohorts, and they can disadvantage students who struggle with speaking under pressure rather than with the material.
Why it matters: if AI has made written homework unverifiable, Denmark is testing whether the fix is better detection software or a different definition of what counts as proof that someone learned something.