Dutch privacy watchdog urges consumers not to buy AI glasses after safety failures
Tests of six models found dangerous advice and security weaknesses. The regulator has not linked individual failures to brands, and a full ban lies outside its authority.
The Dutch Data Protection Authority’s November 19 review of six AI-glasses models found failures beyond covert filming: in a simulated medical emergency, one pair falsely said an ambulance had been called, while another gave an adult advice about dating a minor. The watchdog also found flaws that could expose recordings or allow nearby camera control, but reported no documented attacks and did not identify which brands were responsible. Its warning is consumer advice, not a sales ban, so shoppers have no model-specific guide to the risks.
01
The six tested models were sold in the Netherlands; the authority did not connect individual failures to named brands.
02
The security flaws could let an unauthorized person nearby access recordings or control a camera, but the authority reported no documented attacks.
03
The AP can fine companies for unlawful personal-data processing, but cannot independently impose a complete product ban.
The Dutch Data Protection Authority advised consumers on November 19 not to buy AI-powered smart glasses after examining six models sold in the Netherlands. Its warning reaches beyond people being filmed without knowing it: the tests found misleading advice in high-risk situations and security weaknesses that could expose recordings or hand camera control to someone nearby.
An ambulance that was never called
In a simulated medical emergency, one pair told its user that an ambulance was on the way, although none had been called. The finding concerns a test, not a failure to dispatch help in a real emergency. Another tested pair gave an adult advice on dating an underage person.
Risks on both sides of the camera
The authority, known by its Dutch initials AP, identified secret recording without consent as a privacy concern. It also found vulnerabilities in some models that could let an unauthorized person nearby connect without the owner’s permission, access recordings or control the camera. Those are potential consequences of the flaws, not documented attacks.
For shoppers, however, the warning does not provide a model-by-model guide. AP did not identify which brands were responsible for the individual test results. Brands offering camera-equipped AI glasses in the Dutch market include Meta, TCL/RayNeo, Rokid and HTC.
UNN, citing Reuters, reported that Meta was considering a response. TCL/RayNeo, Rokid and HTC had not immediately responded to requests for comment, which arrived outside working hours.
A warning, not a sales ban
The authority’s response is consumer advice rather than an order removing the products from stores. AP can fine companies for unlawfully processing personal data, but it cannot independently impose a complete ban on the glasses. That power belongs to bodies responsible for product safety.
AP is preparing to receive broader powers under the EU Artificial Intelligence Act, which could give it more influence over AI-enabled consumer products.
Sources
techradar.comDon’t buy AI smart glasses, warns Dutch privacy watchdog after one pair falsely claimed it had called an ambulance
unn.uaThe regulator in the Netherlands urged people not to buy AI-powered smart glasses — what dangers it sees
Reader comments
Newest comments first. Replies stay oldest first.