Key facts
- Meta is updating its AI glasses to cease recording if the safety indicator light is covered.
- The update aims to prevent users from covertly recording individuals.
Meta is rolling out an update to its AI glasses to stop recordings if the safety indicator light is covered. This addresses concerns about non-consensual recording, though privacy advocates argue further risks remain, including data annotation by humans and potential misuse.

Meta's AI glasses are facing scrutiny over privacy concerns, with regulatory bodies and lawsuits questioning the company's data handling and user consent practices. The company's efforts to address these issues through software updates and marketing campaigns are being closely watched.
Meta is implementing a software update for its AI glasses to address concerns about non-consensual recordings. The update will ensure that the device stops recording if a user covers the LED indicator light, which is intended to alert bystanders that recording is in progress.
Alex Himel, Meta's executive vice president of wearables, explained that the camera will now cease functioning if the light is covered during a recording. This measure aims to make the capture LED a more reliable indicator for people around the wearer. Himel stated that while a small number of users attempt to bypass these protections, the company will continue to enhance its approach.
Despite this fix, privacy advocates suggest that risks persist. It remains unclear if the new safeguard will block recordings when using semi-transparent stickers to make the light less noticeable, and bad actors might still find ways to tamper with the device. Data regulators in the European Union are currently examining the privacy implications of Meta's glasses, with Germany considering a ban due to alleged breaches of strict privacy laws.
Furthermore, a proposed nationwide class-action lawsuit in the United States accuses Meta of misrepresenting the privacy risks associated with its AI glasses. The lawsuit claims that users cannot control their data or the data captured of others, and alleges that human data annotators in Kenya review footage, including deeply personal content, to train Meta's AI. Plaintiffs are seeking damages and a change in Meta's marketing messaging to accurately reflect the privacy risks.