AI cheating on exams: how countries are trying to stop it
Smart glasses and tiny earpieces are turning exam halls into signal battlegrounds. China jams radio waves, the US studies lens glare, and South Korea lacks the law to jam.

Cheating with AI glasses is real enough that governments are now hunting for hidden Bluetooth signals in exam halls. Smart glasses, micro cameras and wireless earbuds can photograph a question, send it to an AI and whisper back the answer.
So a new family of anti-AI-cheating tools is emerging. The approaches differ a lot, and none is airtight.
How do you detect AI cheating in an exam room?
South Korea's KCA (Korea Communications Agency) is developing a mobile radio detector. It picks up Bluetooth signals to confirm a device is present and analyses Wi-Fi traffic patterns for signs of AI use. It can also be used against combinations of micro cameras, phones and wireless earbuds. With the right equipment, it can spot LTE signals too.
The catch is noise. Many wireless devices transmit in and around an exam site at once, so telling a cheating device from an innocent one is hard. And a device with the camera and AI built in, or one wired to another gadget, sends no radio signal to catch. KCA says it has expanded simulations and testing beyond its original plan. The work is still at software and prototype stage.
Can exam halls just block the signal?
Jamming does exactly that: it floods certain frequencies so phones and earbuds can't connect. It stops a question reaching an outside AI server, but not an on-device model. Interference can also leak beyond the room. One radio expert said blocking a single venue is possible today, but covering a nationwide exam like Korea's suneung would cost a lot.
In Korea it would also be illegal. Article 29 of the Radio Waves Act bans interfering with others' radio use, with exceptions for things like presidential security, military operations and counterterrorism. College entrance and national certification exams are not on the list.
What China, the US, the UK and Taiwan are doing
China already runs jamming at the gaokao, its university entrance exam. Some areas briefly shut off or turned down nearby mobile base stations. This year it deployed 30,705 radio security staff, 7,306 mobile monitoring vehicles and 15,624 sets of monitoring gear. AI firms joined in: Tencent's Yuanbao and Moonshot AI's Kimi limited image input during exams, Alibaba's Quark stopped photo-based problem solving, and ByteDance's Doubao temporarily refused exam-question uploads.
In the US, a Duke University team led by Maria Gorlatova proposed telling smart glasses from ordinary ones by the odd reflections and light leakage AR lenses produce at certain angles. The College Board bans any glasses that can reach the internet or AI. The UK's JCQ bans wearables, and covers AI built into operating systems such as Microsoft Copilot and Apple Intelligence. Taiwan will treat wearing or carrying smart glasses as a serious violation from the 2027 academic year.
Apps worth a look
Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.
Talk to an AI conversation partner; every sentence comes back corrected and explained in your language. 9 languages, A1 to C2.
Open EDIFACT and X12 messages, see every segment explained, convert to JSON/XML, export PDF and Excel. Runs locally, no upload.
Pick who it is for, the occasion and their personality. The search engine reads product descriptions from local shops and ranks what fits. Free, no sign-up.
More from this day
See all →
Comet 220P/McNaught suddenly got 1,000 times brighter. Here's why

Coral Hart used AI to write 370 books in 20 months

WhatsApp Zaps: self-deleting photos and videos for any chat

Green SM taxis: Karnataka halts new registrations, eyes mandatory meters
More in: Security
See all →
NoisyTCP Linux bug can crash a server remotely, no login needed

Artex AI hacking tool: FBI says it breaches networks in under three hours

Estonia's digital exile government: running the state from abroad
