FOLLOW US

By Samuel López | USA Herald

SAN FRANCISCO - A growing federal privacy battle over Meta’s artificial-intelligence glasses is putting one of the most difficult questions surrounding wearable AI technology before the courts: What happens when an AI device sees something its owner—or the person standing in front of it—never intended to share?

More than 70 people who purchased, used or were allegedly recorded through Meta smart glasses have sued Meta, claiming intimate and sensitive footage was transmitted for human review as part of the company’s AI-development process. The proposed class-action litigation is pending in the U.S. District Court for the Northern District of California.

The allegations are striking.

According to the lawsuit, footage captured through the glasses allegedly included people undressing, using bathrooms, bathing, engaging in sexual activity and entering passwords or other sensitive information. Plaintiffs contend that some recordings were never deliberately created in the first place.

One California plaintiff identified as “PL18” alleges that after leaving his glasses on a bathroom counter, he discovered images of family members bathing and using the bathroom appearing in his gallery—footage the complaint says nobody in the household intended to create.

Another plaintiff, Steven Beltran, alleges that his glasses falsely activated approximately twice per day.

But the lawsuit goes substantially further than accidental recording.

Plaintiffs allege that when certain Meta AI functions were activated, video and audio could be transmitted to Meta’s systems and ultimately reviewed by human data annotators working overseas, including contractors in Kenya, for purposes connected to improving Meta's artificial-intelligence systems.

An earlier version of the complaint alleged that California users were unaware that data was being routed to human annotators and therefore could not consent to interception, transmission, disclosure or use in that manner.

The case has since expanded into an even broader privacy dispute involving bystanders—people who may never have purchased Meta glasses, agreed to Meta's terms of service or knowingly interacted with Meta AI.

The amended allegations contend that even someone who realizes the glasses are recording cannot necessarily consent to subsequent collection, storage, AI use or human inspection when those downstream practices have not been disclosed directly to that person.

That distinction could become one of the most consequential issues in the case.

Traditional disputes over cameras often focus on whether someone had a reasonable expectation of privacy or whether recording was permitted under applicable state law. AI-enabled wearables introduce another layer: recording something and subsequently processing, retaining, analyzing or using that recording as AI-training material are not necessarily the same act.

Meta disputes the plaintiffs' allegations.

“If you use Meta AI, we may review that data to help improve our products and people’s experiences,” a Meta spokesperson told the Los Angeles Times, adding that the company takes steps to filter data to remove identifying information and protect privacy.

Meta has also publicly emphasized privacy safeguards built into its glasses.

The company says a white capture LED alerts nearby people when photographs or video are being captured for a user's gallery. Meta also says that, beginning with its second-generation glasses, blocking the LED automatically disables the camera, and newer protections are intended to detect physical tampering with the indicator.

Meta further states that photos and videos intentionally captured for a user's gallery are privately stored on the glasses and that only the user can see them unless the user imports or shares them.

Those assurances are significant because plaintiffs are attacking what they contend is a gap between Meta's privacy marketing and what allegedly happens when AI functionality is activated.

The litigation includes allegations involving fraud, false advertising, consumer-protection statutes and electronic privacy or wiretapping laws. One related proposed class action, Canady v. Meta Platforms, Inc., alleges that Meta and Luxottica secretly disclosed smart-glasses videos to third-party contractors in violation of federal and state wiretapping laws.

The dispute also arrives as regulators are paying closer attention.

In May, the Texas Attorney General's Office opened an investigation into Meta Glasses, citing concerns about private recordings, facial data and Meta's representations concerning privacy. The investigation does not establish wrongdoing, but it demonstrates that questions surrounding the technology have moved beyond private litigation and into regulatory scrutiny.

The legal stakes extend far beyond Meta.

AI glasses potentially transform cameras from devices people visibly hold into equipment worn throughout everyday life. A bystander may recognize a smartphone pointed toward them as a camera. Ordinary-looking eyeglasses provide considerably less obvious notice.

And the controversy is emerging just as wearable AI becomes more powerful.

Meta says millions of people already use its AI glasses and continues expanding the technology. The company is expected to unveil its latest wearable lineup at its September 23 developer conference.

For courts, lawmakers and technology companies, the developing question may therefore become much larger than whether a camera was technically recording.

It is whether consent to being seen by another person can ever automatically mean consent to being captured, transmitted, analyzed by artificial intelligence and potentially reviewed by another human being thousands of miles away.

The Northern California litigation remains ongoing, and the allegations against Meta have not been proven in court.