Smart Glasses You Can't Tell From the Real Thing
Vivian, a university student in China's Hebei province, scans her exam paper and asks an AI for the answers. Her smart glasses look exactly like ordinary eyewear, and proctors can't tell the difference. "If I think I'm going to fail, I'll use them in any subject," she says — and across Chinese campuses, AI smart glasses are becoming an everyday tool.
An experiment by researchers at the Hong Kong University of Science and Technology offers striking evidence. Students who took an exam wearing smart glasses connected to ChatGPT averaged 92.5 points — more than 20 points above the overall average of 72. Most wearers scored in the top 5%.
Rising demand has even spawned a rental market. Kechangsi, a smart-glasses rental company, says it has lent devices to more than 1,000 people in the past four months. This is no longer a matter of individual rule-breaking; it's a structural phenomenon with a market behind it.
Korea Is No Exception
In Japan, a 2024 case at Waseda University's entrance exam involved a test-taker who photographed the exam with smart glasses and transmitted it outside the room. Proctors caught nothing during the exam itself — the cheating only came to light through a tip afterward. It's a vivid illustration of surveillance systems failing to keep pace with technology.
South Korean campuses are no different. Last year, numerous students at top universities including Yonsei and Korea University were caught using AI to cheat, sharing answers from ChatGPT and Gemini in real time through anonymous group chat rooms. As AI becomes effortless to access, the methods of cheating are evolving with it.
Korea's Ministry of Education released ethical guidelines for AI use this past February, but critics say policy still lags far behind the technology. Broad principles like integrity, transparency, and fairness offer little help in concrete situations.
The Augmented-Human Era: Time to Rethink How We Test
Should we really dismiss all of this as simple cheating? In an age when AI extends human cognitive abilities, the traditional exam — built to measure memory and raw computation — may itself be the anachronism. Just as almost no one memorizes phone numbers in the smartphone era, we have to ask whether doing complex calculations without AI is still a skill worth measuring.
The first question is what we're actually trying to assess: the ability to memorize information, or the ability to solve problems. If the capacity to achieve more by working alongside AI is what tomorrow's society demands, then today's exam system may be the thing that's truly out of date.
