MAIN CONTENT
A camera that changes the rules of public space
Smart glasses promise hands-free photographs, spoken assistance and information delivered without looking at a phone. Their unobtrusive appearance, however, also makes the act of recording less visible. For a passerby, colleague or child, the difference between ordinary eyewear and a camera connected to an AI service may be impossible to recognize. That asymmetry matters because the wearer can initiate recording while the person being recorded may have no meaningful opportunity to object.
The AP's November review covered six kinds of AI camera glasses available in the Dutch market. It warned of risks both to users and to bystanders and advised consumers not to purchase such devices for general use for now. The authority can sanction unlawful personal-data processing, but it cannot independently impose a blanket product ban. Such a decision would involve other regulatory powers and legal processes. Its statement should therefore be understood as a strong precautionary recommendation, not a finding that all smart glasses have been prohibited.
When a confident answer becomes a safety hazard
One test illustrates how a digital assistant's fluency may conceal a dangerous error. In a simulated medical emergency, a pair of glasses stated that an ambulance was already on its way even though emergency services had not been contacted.
Someone trusting that response could delay making a critical call. The reported incident did not involve an actual patient; its importance lies in showing how misleading assurances can emerge precisely when users need verified instructions.
The AP also described an inappropriate response involving an adult and a minor, alongside broader worries about how software handles sensitive questions. These tests are specific observations rather than a statistically complete measure of every product's performance. Nonetheless, they expose a design problem: people may interpret speech delivered through a device worn on their face as immediate, personalized guidance. Safety therefore requires not merely a disclaimer, but systems that acknowledge uncertainty and direct users to real emergency services when necessary.
Security weaknesses and the question of consent
Investigators identified flaws in several devices that could potentially allow a nearby person to connect without permission, view recordings or control a camera. In a wearable device, such weaknesses affect two groups at once. The owner may lose control of private media, while an unsuspecting bystander may be recorded or observed without choosing to participate. Safeguards need to cover wireless connections, access permissions, software updates and transparent signals that recording is taking place.
Meta has said its camera-equipped glasses include a visible indicator light when taking photos or videos, and it has maintained that the light cannot be disabled. That claim speaks to one form of notification; critics remain concerned about whether people notice, understand or can refuse recording in crowded settings. The Dutch inquiry did not publicly identify the manufacturers of the tested models, so its particular defects should not be attributed to any named brand without separate evidence. Meta said it was considering a response, while Reuters reported no immediate responses from several other vendors.
A marketplace meets a new regulatory threshold
The dispute is moving through retail aisles, courts and European institutions simultaneously. Dutch retailer Hans Anders suspended sales of Meta's Ray-Ban smart glasses in the Netherlands and Belgium in October. Dutch consumer and rights organizations also brought legal action against Meta over privacy concerns, with a preliminary hearing scheduled for November 26. The litigation concerns allegations that remain contested; filing a case is not a judicial finding of wrongdoing.
Europe's AI Act is expected to broaden the AP's oversight role, while existing data-protection obligations already apply where personal information is processed. Regulators must balance legitimate uses, including accessibility assistance, with the rights of people who never bought the device but may still appear in its recordings. Manufacturers, meanwhile, have incentives to make products effortless and attractive. The central test will be whether strong security, reliable warnings, meaningful notice and independent scrutiny become built-in features rather than repairs after controversy.

















