Meta pulls Instagram Muse Image in days
Meta launched Instagram Muse Image on July 8 and removed it by July 10 after the Creative Artists Agency raised consent concerns over celebrity likenesses.
Meta launched an Instagram feature on July 8 letting users reference any public profile in AI image generation by @-mentioning accounts. The feature did not notify the referenced accounts or ask for consent. Users could opt their own profiles out via privacy settings, but the default was opt-in.
The Creative Artists Agency raised concerns within days, naming Tom Hanks and Meryl Streep as examples of public figures whose likenesses could be used without their knowledge. Meta removed the feature on July 10, two days after launch, and acknowledged it “missed the mark.”
The Muse Image model itself, announced alongside the Instagram feature and released as a standalone API, remains available through other surfaces including the Meta AI app and meta.ai.
What it means for you
Meta launched a default-opt-in feature using real people’s likenesses without notification, got pushback from a major talent agency within 48 hours, and pulled it. The speed of that sequence is the data point here. Consent norms around AI-generated likenesses are moving faster than most product teams are accounting for.
For operators building AI features that touch profiles, user-generated content, or any data where someone has a reasonable expectation of control, “users can opt out” is not the same as “users consented.” The Muse Image reversal makes that concrete. Any AI feature that defaults to using someone’s likeness without explicit consent is now a higher-risk launch, regardless of technical legality. For a practical framework on building with this in mind, see when AI features launch and vanish overnight.