Class action alleges Meta used Facebook and Instagram photos to train its AI image models

What happened

On 4 September 2026 a 66-page class-action complaint was filed against Meta Platforms, Inc. in the Northern District of Illinois (Alvarez v. Meta Platforms, Inc., 1:26-cv-10773). The complaint alleges Meta collected people's face photographs uploaded to Facebook and Instagram and used them to train and test its generative AI image models — naming the Emu and Muse Image models — as well as to power its facial-recognition "NameTag" feature. The complaint treats the face-geometry embeddings stored in the trained models as biometric identifiers.

The suit is pleaded as privacy and commercial-appropriation claims (Illinois Biometric Information Privacy Act, California right of publicity and privacy), not copyright, and it names Illinois, national and California classes covering users whose uploaded images were allegedly fed into Meta's models.

Why it may matter to you

This is one of the first lawsuits to name specific Meta models and allege that ordinary users' photographs were repurposed as training data for text-to-image generators. Even pleaded as privacy claims, the underlying training-data allegation is exactly the kind of reuse this site tracks.

What is not yet known

Meta has not yet answered the complaint. The filing is the plaintiffs' allegations only; nothing has been proven or ruled on.