Meta kills Instagram's instant deepfake generator

Meta just pulled the plug on a controversial Instagram AI feature just days after launching it, following swift and intense backlash.

Earlier this week, Meta's dedicated AI division, Meta Superintelligence Labs, rolled out Muse Image, a new AI image generator. Alongside it came a feature that let anyone generate AI images based on public Instagram accounts simply by @-mentioning them. The catch? It was enabled by default, required no permission from the account owner, and didn't even notify people when their photos were referenced.

While Meta claimed the tool was meant to help creativity and pointed out that users could dig through their settings to opt out, critics immediately sounded the alarm. The National Center on Sexual Exploitation warned the feature was an open door for sextortion and scams, slamming Meta for putting the burden on individuals to protect their own likeness. Hollywood talent agencies like CAA raised concerns, and the Screen Actors Guild (SAG-AFTRA) even issued a step-by-step guide urging members to turn it off immediately.

By Friday, Meta surrendered. In an update to its blog post, the company admitted the feature "missed the mark" and removed it entirely.

Why it matters

For anyone with a public profile, this was an alarming wake-up call about how tech giants view your personal data. It proved that without loud public outrage, platforms are more than willing to treat your public photos as raw fodder for generative AI tools.

We think Meta's rollout was remarkably reckless. Opt-out privacy designs are already frustrating, but applying them to personal likenesses—especially amid rampant issues with deepfakes and non-consensual imagery—is a massive unforced error. The quick reversal shows that Hollywood talent unions and exploitation watchdogs still have enough leverage to push back, but it also signals that tech giants will continue testing how much user data they can grab until someone yells stop.