Meta disables Instagram AI feature that generated deepfakes without creator consent

Meta has disabled an Instagram feature that enabled users to generate AI-powered images sourced from public accounts, backing away from the controversial capability just days after its rollout.

What was the feature?#

The company had rolled out the ability to tag public Instagram accounts directly when creating images through Meta's new Muse Image AI model, allowing the AI to draw on content from those profiles without explicit permission from account owners. While Meta provided an opt-out mechanism buried in settings, the feature's default design gave creators minimal control over how their publicly shared photos could be repurposed in synthetic images.

Why it faced backlash#

Critics quickly flagged serious concerns about the approach. Safety advocates pointed out that the system opened doors to misuse—particularly for creating intimate deepfakes without consent. As Haley McNamara, executive director of the National Center on Sexual Exploitation, stated earlier this week, the tool represented "an obvious instrument for #sextortion and other scammers." The industry swiftly reacted: the Screen Actors Guild instructed members to opt out and shared instructions for doing so.

Meta's reversal#

In an update to its announcement blog post, Meta acknowledged the feature "missed the mark," saying the company "heard the feedback" and removed it entirely. The company had framed the opt-out as putting "control" in users' hands, but the approach of forcing people to navigate settings to protect themselves from having their likenesses repurposed drew particular criticism for placing the burden on individuals rather than building privacy protections into the system by default.