Meta adds anti-tamper LED to AI glasses, but privacy worries remain

Meta is introducing a new safety mechanism for its AI glasses: if the LED indicator that signals recording is tampered with — say, covered by tape — the camera will automatically disable. The company claims this is an industry-first and a response to users who used “sophisticated efforts to modify or destroy the capture LED.” In its blog post, Meta acknowledges that some users have hidden agendas, such as recording people without consent. Yet even as it touts this privacy feature, the company is moving in the opposite direction with its broader AI strategy. On the same day, Meta announced that it will use public Instagram photos to train its AI models unless users manually opt out. It is also testing a prototype of its AI glasses that would “continuously collect audio while taking photos every few seconds,” according to the Financial Times.

This duality underscores a deeper pattern. Meta’s privacy policy already allows any image shared with Meta AI to be used for training, and the company faces multiple lawsuits over privacy violations involving its AI glasses — including a case where Kenyan workers were exposed to graphic content while training AI on users’ videos. Meta’s history compounds the distrust: the Cambridge Analytica scandal, employee keystroke recording for AI training, and plans to use data from AI chats for targeted ads. The LED safeguard, while necessary, appears to be a band-aid on a much larger privacy wound. As Meta pushes ahead with AI features that demand more personal data — often with opt-out defaults — consumers remain wary that the company’s vision of the future depends on eroding privacy, not protecting it.