Meta shuts down Instagram AI deepfake feature 3 days after launch

Signals Inbox·July 11, 2026·AI Creative Tools

Meta has disabled the Muse Image feature that let people generate AI images by mentioning almost any public Instagram account. It lasted roughly three days. The model worked, but Meta discovered that turning millions of people into opt-out AI characters was much harder to defend.

Q1What officially happened?

Meta updated its official Muse Image announcement on July 10 and said it was disabling the ability to reference public Instagram accounts. Meta admitted that the feature had missed the mark. It had been announced only three days earlier, on July 7.

Q2What could people actually do?

A user could mention a public Instagram account inside a Meta AI prompt. Muse Image could then use that account’s public photos as references for a new AI-generated image. The account owner did not have to approve each generation or even know that it had happened.

Q3Why did the feature cause such a backlash?

Because Meta treated public visibility like permission to become an AI character. Public accounts were eligible by default unless their owners changed a sharing setting or made their profiles private. That created obvious risks around fake endorsements, sexualized images, harassment, scams and reputational damage.

Q4What is the bigger lesson?

Consent is becoming part of the product, not a legal checkbox hidden in settings. Social platforms have the data to build extremely personal AI tools, but that advantage can turn into a liability when users feel that their identity has been taken rather than invited. Meta proved that distribution cannot rescue a feature people fundamentally do not trust.