CHICAGO — A proposed federal class action accuses Meta Platforms of using faces appearing in Facebook and Instagram photographs to develop artificial-intelligence and facial-recognition systems without adequate notice or consent.
The lawsuit, filed September 4 in the U.S. District Court for the Northern District of Illinois, alleges that Meta extracted or generated biometric information from photographs connected to its social platforms. The plaintiffs point to the company’s unreleased “NameTag” facial-recognition system and image-generation models including Emu and Muse Image.
These are allegations, not court findings. No class has been certified, Meta has not been found liable, and the case remains at an early stage.
A biometric case, not primarily a copyright case
The plaintiffs assert claims under the Illinois Biometric Information Privacy Act and California privacy, publicity-right and commercial-misappropriation laws. Their central theory is that numerical facial information allegedly created or retained during AI training and recognition processes qualifies as protected biometric data.
That makes this dispute different from lawsuits alleging that copyrighted articles, books, music or images were copied to train AI models. The question here is whether identifiable faces inside social-media photographs were converted into regulated identity data.
The complaint also acknowledges that Meta has not publicly disclosed which images, if any, supplied the facial information at issue. That information would have to be developed through the litigation.
Meta told WIRED that the lawsuit is without merit and misrepresents its work. The company said NameTag has not shipped to consumers, no final decision has been made about releasing it, and Meta is not building a universal facial database.
Why the case matters to radio and media
Radio stations, television outlets and digital publishers have spent years uploading photographs of performers, listeners, contest winners, employees and community events to Facebook and Instagram. The case asks whether a platform may transform faces appearing in those archives into AI-related identity information without separate notice and consent.
The filing does not presently impose new liability on stations merely for posting authorized photographs or testing an AI product. However, media companies should be especially careful when uploading photographs of minors or submitting identifiable people’s images directly to generative-AI services.
If the plaintiffs’ biometric theory survives dismissal, it could expand the consent, disclosure and data-retention obligations facing AI companies that use photographs containing identifiable people. Survival at the dismissal stage would allow the claims to proceed; it would not establish that Meta violated the law.
The RNN angle: Radio and media companies supplied social platforms with years of audience, artist and employee photographs. This case asks whether those platforms can convert the faces in those archives into AI training data or recognition tools—and what consent is required before they do.
Read the filed complaint and follow the federal docket.
This report provides general information and is not legal advice.