Skip to content
Theme
OnDevice Tools

CybersecurityPrivacy

Dutch privacy watchdog warns against AI smart glasses after dangerous failures

The Dutch data protection authority says AI-powered smart glasses are unsafe after testing revealed false emergency alerts and security flaws across multiple brands.

Illustration of AI smart glasses with warning symbol and ambulance in background

Key takeaways

  • Dutch privacy regulator AP recommends against buying any AI smart glasses after testing found dangerous failures.
  • One device falsely claimed an ambulance was coming during a simulated emergency; another gave inappropriate advice about a minor.
  • Security flaws in multiple brands could let nearby attackers access recordings or control the camera without permission.
  • Major brands tested include Meta, TCL, RayNeo, Rokid and HTC, but AP did not link specific faults to specific models.
In this article
  1. Watchdog finds life-threatening errors
  2. Security flaws allow unauthorised access
  3. Major brands included in testing
  4. Growing regulatory and public backlash
  5. Trust in the device itself now in question

Watchdog finds life-threatening errors

The Dutch privacy regulator Autoriteit Persoonsgegevens (AP) has advised consumers not to buy AI smart glasses after its testing uncovered serious safety failures. According to TechRadar, the watchdog tested several brands sold in the Netherlands and found that one pair falsely told a user an ambulance was on its way during a simulated medical emergency when no ambulance had been called. Another pair gave inappropriate advice to an adult about dating an underage person. AP vice chair Monique Verdier said a device that films people without their knowledge is already worrying, but one that gives wrong advice when a life is at stake goes a step too far.

Security flaws allow unauthorised access

Illustration: security, glasses, AI, law and the labels AI Smart Glasses, Autoriteit Persoonsgegevens, Meta

Beyond the dangerous AI responses, the regulator discovered security vulnerabilities in multiple pairs of glasses. These flaws could allow someone nearby to connect to the glasses without the owner's permission, access stored recordings, or even operate the camera remotely. The findings mean that both the wearer and people around them are at risk, not only from privacy invasion but from active exploitation of the device itself.

Major brands included in testing

AP did not disclose which specific brand was responsible for each failure. However, the testing covered devices from the current market leader Meta, Chinese manufacturers TCL and RayNeo, Rokid, and Taiwanese company HTC. All of these brands advertise AI smart glasses for sale in the Netherlands. The regulator's decision to warn against the entire category rather than individual products reflects the systemic nature of the problems found.

Growing regulatory and public backlash

The warning adds to mounting pressure on the smart glasses industry. The devices were previously criticised as "perv glasses" after repeated reports of them being used to film women without consent, sparking a "Ban Ray" movement encouraging people to switch to traditional sunglasses. England and Wales have already banned smart glasses from courtrooms, and the UK government is considering further restrictions on their sale. In response to the controversy, Meta recently released a new version of its smart glasses that omits the camera entirely.

Trust in the device itself now in question

Until now, the main concern about smart glasses centred on what the wearer might do to bystanders. The AP findings shift the focus to whether the glasses themselves can be trusted. With AI generating potentially dangerous advice and security flaws exposing users to hackers, the regulator's conclusion is that the current generation of AI smart glasses poses an unacceptable risk to both users and the public.

Sources

  1. AI smart glasses falsely claimed an ambulance was coming in an emergencyTechRadar

Tags

Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.