Proposed class action accuses Meta of extracting biometric faceprints from Facebook/Instagram photos to train AI and build facial recognition

What happened

On September 4, 2026, a proposed class action was filed in the U.S. District Court for the Northern District of Illinois — Alvarez et al. v. Meta Platforms, Inc. — accusing Meta of harvesting facial images and biometric data from Facebook and Instagram users without consent, and using that data to train generative AI models and to develop an unreleased facial recognition system called NameTag. The plaintiffs — Illinois resident Francisco Alvarez and his minor son, California resident Jeremy Wahl and his 10-year-old daughter — allege Meta extracted biometric identifiers (digital "faceprints") from faces appearing in social media images without the notice or consent required by law, violating the Illinois Biometric Information Privacy Act (BIPA) and California privacy and publicity statutes. The proposed class could include millions of people whose images were uploaded to Facebook or Instagram, or submitted to Meta's AI products, dating back to September 4, 2021 — including non-users whose faces appeared in others' photos.

The complaint's generative-AI allegations target models including Emu and Muse Image: Meta has said it trained Emu on large quantities of Facebook and Instagram images and text, with chief product officer Chris Cox calling those platforms a "data advantage"; Muse Image, released in summer 2026, briefly allowed users to generate images based on other people's public Instagram accounts before Meta removed the feature within days. The facial-recognition allegations build on a June 2026 WIRED investigation that found inactive NameTag code inside the Meta AI companion app (designed to convert faces captured by Meta smart glasses into biometric signatures and compare them against stored faceprints); Meta removed the code on June 5, the day after the report, and cites a May 2026 patent application describing a possible architecture. A Meta spokesperson called the lawsuit "without merit" and said NameTag has not shipped to consumers, no final decision has been made on it, and the company is "not building a universal face database."

Why it matters to creators

This is one of the first cases to connect two creator-relevant practices: training commercial generative image models on ordinary people's photographs posted to social platforms, and converting those same photographs into biometric templates for identification systems. It tests whether platform photos — including photos of people who never agreed to anything — can be lawfully converted into faceprints and training data, and whether a generative model that can reproduce or clone a real person's appearance creates liability for the platform that built it.

What is not yet known

Which images, if any, were actually used to generate biometric data — the complaint acknowledges that information is solely in Meta's possession; whether the case will be certified as a class action; and how the court applies BIPA and California privacy law to photos used for AI training and facial recognition.