The Dutch privacy watchdog, the Autoriteit Persoonsgegevens (AP), has told people not to buy smart glasses, saying they can get things dangerously wrong.
It tested six pairs available on the Dutch market, priced from under €100 to over €500 and originating from different continents. It has not named them.
What the tests found
The findings are specific, and they are not about privacy.
In a simulated emergency where CPR advice was requested, one pair claimed several times that an ambulance was on the way. Emergency services had not been contacted.
One pair gave dating advice to an adult who indicated a romantic interest in a minor.
Others misread the instructions inside medicine packets, and misinterpreted severe weather warnings.
Read those together and the pattern is not "the AI was inaccurate". It is that the device was confidently wrong in exactly the situations where someone would rely on it most — a medical emergency, a medicine dose, a storm warning, a child's safety.

The cruel irony in who this hurts
The AP explicitly recognises that these glasses can be valuable for people who need help understanding their surroundings. Meta has promoted exactly that use — AI describing a wearer's environment or reading text aloud for people with visual impairments, and we covered a Cheshire farmer whose glasses let him read his own post for the first time.
But the regulator's point follows directly: for those users, reliability matters more, not less. A sighted person who is told the wrong thing on a medicine packet can check it. Someone using the glasses because they cannot read the packet themselves has no second source.
The recording light can be taped over
On privacy, the AP found the warning lights that indicate recording could be covered with tape and the glasses would still work — despite some manufacturers claiming they would not.
That matters because the indicator light is the entire basis of the industry's privacy answer. When we wrote about Google's glasses on 13 November, the light was the headline reassurance; London pubs and theatres had already started banning the devices in September precisely because people did not trust it.
The AP also found serious security flaws in some glasses: someone nearby could connect without permission, view recordings, or operate the camera remotely. And data recorded by the glasses was often exported outside the EU.
In February, reports showed outsourced workers could view sensitive content recorded on Meta smart glasses. Meta said such review helps improve its AI, that it takes data protection seriously, and that it blurs faces first.
What the regulator did and did not do
The AP stopped short of calling for a ban. It is advising people not to buy.
That distinction is worth noticing. A ban would require proving a legal breach; advice requires only that the regulator has looked and is not satisfied. It is the stronger signal in one sense — the authority is saying the problem is the product as sold, not a fixable technicality.
The BBC approached Meta, TCL and HTC for comment. It is not known whether Meta's glasses were among the six tested; the AP will not say.
Meta recently released a version of its glasses without a camera, which some commentators read as a response to the privacy backlash.




