Meta Shuts Down Instagram AI Feature After Public Backlash

Meta pulled an Instagram feature this week that let users create AI-generated images by tagging other people's accounts as visual inspiration. The feature, called Muse Image, launched earlier in the week and was gone by Friday, July 10, 2026 TechCrunch.

Here's how it worked: any Instagram user could tag someone else's public account, and the AI would generate a new image based on photos from that account. The person being tagged was never notified that their photos were being used this way TechCrunch.

Meta announced the removal in a brief post on its Instagram website, saying: "We've heard the feedback that this feature missed the mark, so it's no longer available" TechCrunch. The company didn't explain exactly what went wrong, and didn't say how many images were created while the feature was live.

The pushback came from entertainment agencies and users who spotted the tool quickly. Dylan Byers of Puck News first reported the removal, noting that talent agencies like Creative Artists Agency (CAA) raised concerns TechCrunch. This matters because agencies represent celebrities and creators who have a financial stake in how their images and likenesses are used.

Even before Meta officially pulled the feature, the company had quietly included an opt-out setting buried in Instagram's profile menu. On July 9 — one day before the removal — TechCrunch published instructions on how to use it TechCrunch. The existence of an opt-out shows Meta anticipated some concern about consent, but most people would never have found it, since it was hidden deep in settings and switched on by default.

The core issue here is how Meta set things up. Rather than ask permission first, the company turned the feature on automatically for everyone, then made people dig through menus to turn it off. It's like a company mailing everyone a service without asking — you have to actively opt out, and most people won't know they need to.

This is a familiar pattern in tech. When companies are building new features that use other people's content — whether photos, voices, or likenesses — there's often a tension between moving fast and respecting consent. Meta appears to have chosen speed and convenience for the user creating the image, placing the burden of discovery and action on the people whose photos were being borrowed.

The speed of the reversal tells us something about how Meta assessed the risk once organized opposition showed up. The company pulled the feature within days, not weeks, which suggests the involvement of talent agencies and their legal teams shifted internal decision-making faster than ordinary user complaints alone would have. In my view, this reflects how much murkier the legal and reputational territory still is when a generative AI tool touches someone else's likeness, compared to earlier applications of AI that mostly worked with data a company already owned.

Meta has not said whether the Muse Image tool as a whole is still available on Instagram for other uses, or whether the company will bring back the @-mention reference feature in a revised form with clearer consent rules. The company's framing — that it listened to feedback — leaves open the door for a version 2.0.

At a broader level, this episode shows how unevenly the rules around AI and other people's content are being written, even now. Instagram users have shared photos on the platform for years, and there are established norms around what you can do with someone else's picture — you can reshare it, tag it, comment on it. Using that person's photos as raw material to train an AI engine to generate new images is a different kind of use, and the fact that Meta defaulted it to "on" rather than "ask first" reveals how new this territory still is, both for platforms and for the people using them.