Meta is closing a loophole that let wearers of its AI glasses cover the recording indicator light and keep filming, as the company faces regulatory probes in the European Union, a potential ban in Germany, and a proposed US class action over its privacy marketing. Under the update, the camera will shut off the moment the capture LED is obscured during a recording, ending a bypass that surfaced when users learned they could start filming first and place a sticker over the light afterward.
Alex Himel, Meta's executive vice president of wearables, announced the change on Threads and framed it as a reliability fix for the only bystander-facing safeguard the glasses carry. Meta had already blocked recording when a user tried to start capture with the LED already covered, but the reversed sequence defeated that check. The new behavior monitors the light throughout a session rather than only at the start.
“the camera will now stop working if the light is covered during a recording.”— Alex Himel, Meta executive vice president of wearables
The fix does not resolve every workaround. Bad actors can still mod the hardware or tamper with the light in ways that may not trigger the shutoff, and it is unclear whether semi-transparent stickers, a popular tactic for dimming the LED without fully hiding it, will be caught by the new safeguard. Meta paired the update with a marketing push, including a Los Angeles billboard reading, "The camera lets you capture the moment. The light lets everyone around you know when you do."
Key facts
- 01Meta's update forces the camera to stop recording if the capture LED is covered mid-session, closing a bypass users exploited by reversing the sticker-then-record order.
- 02Germany is considering a ban after a nonprofit filed a criminal complaint alleging the glasses breach German privacy law.
- 03The European Data Protection Board is set to release a report later this summer on EU-wide actions.
- 04A proposed US nationwide class action alleges Meta's 'designed for privacy' marketing is false and that human annotators in Kenya review captured footage to train Meta AI.
- 05Meta has yet to respond to the class action; the suit seeks damages and an order barring Meta from marketing the glasses as privacy-protecting.
Himel argued the bypasses reflect a small share of usage rather than a systemic issue.
“only a tiny minority of people try to get around these protections, and the fact they have to go to these lengths shows our efforts are working. We'll keep improving and evolving our approach.”— Alex Himel, Meta executive vice president of wearables
Regulators in the EU are less convinced. Data authorities are already probing whether Meta adequately field-tested the LED to ensure the public understands what it signals, and Germany is weighing a ban after a nonprofit filed a criminal complaint alleging the glasses violate the country's privacy laws. The European Data Protection Board is expected to release a report later this summer laying out what actions member states should take to prevent nonconsensual recording across the bloc.
In the US, a proposed nationwide class action is advancing on a separate but related front: the accuracy of Meta's privacy marketing. Meta has not yet filed a response. The complaint challenges advertising claims that Meta AI Glasses are "Built for your privacy and others' too," alleging that "in reality, users of Meta AI Glasses cannot control what happens to their data, or the data they capture of the people around them."
The core allegation is that Meta does not disclose a human review pipeline in which data annotators in Kenya watch footage captured by the glasses to train Meta AI. According to the complaint, that footage has included sexual activity, bathroom visits, credit card information, and private text messages. One cited example describes a user setting his glasses on a nightstand, capturing video of his wife undressing that was then routed to overseas reviewers.
The suit argues Meta knew about the review pipeline and marketed the product as privacy-protecting anyway: "Despite their actual knowledge of the human review pipeline and the inherent safety and privacy risks associated with such sharing of data, Defendants have continued to market the Glasses using the same 'designed for privacy' representations." Plaintiffs are seeking damages, potentially including punitive damages, for every US buyer and want the court to bar Meta from continuing to advertise the glasses as privacy-protecting.
The complaint also frames the alleged data collection as a business strategy tied to the economics of AI training. "New data input is central to AI production, but as AI labs exhaust publicly available data, proprietary sources of new data can cost hundreds of millions of dollars," the filing states, arguing that Meta is instead harvesting first-person video from unaware consumers. Meta disputes the framing through its ongoing marketing, which continues to lean on privacy-by-design language.
Meta's response so far, the LED-shutoff update and a public awareness campaign, addresses the narrower question of whether bystanders can tell when they are being filmed. It does not touch the class action's broader claim that wearers themselves lack meaningful control over where captured footage ends up. Whether semi-transparent stickers defeat the new check, and whether modified hardware slips past it entirely, will decide how much the fix actually changes on the ground.
For Meta, the harder problem is that the glasses sit at the intersection of a consumer wearable business and a training-data pipeline, and the regulatory pressure is now aimed at both. An LED patch is cheap; disclosing a Kenya-based review pipeline in marketing, or losing the ability to route captured video into model training, would reshape the product's economics. The EU report and the US class action will decide which of those Meta ends up doing, and the AI-glasses category, which competitors are lining up to enter, will inherit whichever precedent gets set first.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
