US senators asked Meta to explain how it is handling serious privacy risks with its AI glasses

Two U.S. senators have asked Meta to explain how it is handling what they call serious privacy and safety risks with its AI smart glasses. Ron Wyden and Edward Markey sent a letter to CEO Mark Zuckerberg saying the glasses have increasingly been implicated in disturbing invasions of privacy for users and bystanders, Engadget reports.

The letter cites reports from 404 Media and the BBC about men secretly filming women with the glasses and posting the recordings online. It also points to reports and a lawsuit in which Meta's AI contractors said they reviewed footage showing people undressing and engaging in sexual activity. Digital Trends, summarizing the senators' press release, says Meta's own human rights assessment flagged risks in 2021 and that more than 70 plaintiffs have sued.

The senators posed specific questions. They ask how Meta prevents accidental camera activation and tampering with the recording indicator, and whether the glasses blur faces and license plates in certain AI interactions, as Meta does on Facebook and Instagram. They want to know when employees or contractors can access imagery, when it is used to train AI models, whether Meta ran detailed human rights due diligence and safety testing before launch, and whether both owners and people recorded can request deletion of data.

Meta has taken some steps. Reporting says it ended its contractor relationship after the earlier investigation, added an option for wearers to stop visual data being retained for AI training, and updated the glasses to prevent recording when the indicator light is covered. The senators argue that people captured without consent have no corresponding control. Responses are due in late November, with outlets giving November 25 or Thanksgiving, November 26.

For developers building on camera wearables or on-device vision, the letter shows where scrutiny is heading: bystander consent, who reviews captured media, and whether AI training use is clear and optional. Those are design questions as much as legal ones.

A reply from Meta is not public yet. The letter is an inquiry, not a regulation, and the reports in it include allegations that Meta has not been quoted as confirming in the sources reviewed.

Building with AI? See this week's hot AI repos and how to use each one.