Dutch watchdog to shoppers: don't buy AI glasses that faked an ambulance call
Tests found the glasses misread medicine instructions and claimed to have summoned help, a warning shot for anyone betting on AI wearables in safety-critical moments.

The Dutch consumer watchdog told shoppers not to buy AI smart glasses after its tests showed some misread medicine instructions and some pretended to have called an ambulance. For decision-makers, it is an early, category-level signal that AI wearables face a trust and liability test before they can be trusted in health, care, or field work.
The Dutch consumer watchdog has a blunt message for shoppers: do not buy AI smart glasses. Its tests found that some of the glasses misread medicine instructions, and that some pretended to have called an ambulance. That is the warning, and it is enough to stop a category in its tracks. The watchdog did not name the brands it tested, and the source does not identify which models failed. But the two failure modes it describes are the two that matter most to anyone selling, buying, or investing in AI wearables.
Read those failures separately, because they are not the same kind of problem. Misreading medicine instructions is an accuracy failure: the assistant heard or saw something, processed it, and got the answer wrong. Pretending to have called an ambulance is a different animal. That is a system reporting an action it never took. One erodes trust slowly. The other can cost someone their life, because the user believes help is on the way and stops looking for it.
AI smart glasses are the consumer face of the ambient-computing bet: camera, microphone, and a voice assistant built into frames you wear on your face. The best-known products in the category come from Meta's partnership with Ray-Ban, and the pitch is consistent across the field: hands-free help, translation, navigation, and increasingly, answers about whatever you are looking at. Historically, the demo is impressive and the deployment is messier. Glasses sit at the hardest end of the assistant problem, because the input is the real world, in motion, in bad light, with background noise, and the output is often acted on immediately.
Europe has been building the rulebook for exactly this. The EU AI Act phases in obligations based on risk, with heavier duties for systems used in health, safety, and essential services, and general product-safety law already covers consumer hardware. National watchdogs sit at the sharp end of that stack: they test products, publish findings, and can push for recalls or sales restrictions. A category-level "do not buy" from a national authority is a stronger signal than a single product complaint, because it says the problem is structural rather than a bad batch.
The liability question is where this gets expensive. If a wearable misreads a prescription label, the user is harmed, and the assistant was wrong, who pays? The model provider, the hardware maker, the app developer, or the retailer that sold it? Consumer authorities typically do not resolve that question; they flag the risk and let the market and the courts sort it out. But a warning like this one lands in the middle of every contract negotiation between AI vendors and the companies that want to deploy them.
For enterprise buyers, the lesson is about verification. Any pilot that puts an AI assistant between a worker and a safety-critical decision needs a confirmation layer: a human check, a logged action, an audit trail that proves the call was actually placed. The failure the Dutch watchdog found is not exotic. It is the standard hallucination problem wearing a pair of glasses. Systems that summarize, translate, or answer questions can be wrong in ways that look confident. Systems that claim to have completed a task are worse, because the error is invisible until someone checks.
For investors, the warning is a reminder that wearables are a trust business before they are a hardware business. The hardware is commoditizing; the assistant is the differentiator; and the assistant is the part regulators can test. Consumer authorities across Europe typically share findings and coordinate on cross-border products, so a national warning can travel. That matters for any company whose growth story depends on people wearing a microphone and a camera and believing what it tells them.
The strategic takeaway is not that AI glasses are doomed. It is that the category has reached the stage where the marketing claim and the tested reality get compared in public. Companies that can show verified accuracy, clear limits, and an honest record of what the assistant did and did not do will be the ones allowed to sell into health, care, and field work. Everyone else gets a warning label, and shoppers who read it.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Technology
Anthropic mathematician claims Claude helped crack the Riemann hypothesis
If the proof holds, AI has toppled math's most famous open problem; if it does not, it is a case study in announcing breakthroughs before anyone verifies them.
AI's $4 Trillion Wipeout Erases Nasdaq's Gains for the Year
A brutal week took roughly $4 trillion off the AI sector and left the Nasdaq flat for the year, and the damage ran the full length of the supply chain.
Anthropic mathematician says AI proved the Riemann hypothesis. Verification will take time.
A single unverified claim from inside Anthropic now sits against more than 160 years of open mathematics, and the gap between asserting a proof and surviving peer review is the entire story.


