Meta Removes Instagram AI Feature That Let Users Generate Images of Other People’s Photos
Meta has pulled one of the features from its newly launched Muse Image AI generator after it triggered immediate and intense backlash from Instagram users, celebrities, and major talent agencies. The removed feature allowed anyone to generate AI images by @-mentioning other people’s public Instagram accounts — without the account owner being notified. The incident is the latest example of a major tech company misjudging how AI image generation tools will be perceived and used in practice, according to reporting by TechCrunch.
The Meta Instagram AI feature removed controversy unfolded quickly: the tool went live earlier in the week, public criticism followed almost immediately, and by July 10, 2026, Meta had announced it was taking the feature offline. The reversal is notable in its speed and in the directness of Meta’s admission that the feature was a mistake.
What Muse Image Is — and What the Controversial Feature Did
Muse Image is a new AI image generator built by Meta Superintelligence Labs, the company’s dedicated AI research unit. Meta announced the tool on July 7, 2026, as part of a broader set of AI features being rolled out across its platforms.
Most of Muse Image’s features are straightforward AI generation tools — text-to-image capabilities similar to what competitors like Midjourney and Adobe Firefly offer. But one feature stood out immediately as problematic: the ability to generate images by @-mentioning public Instagram accounts as references.
In practice, this meant that any Instagram user could type “@[someone’s handle]” into the Muse Image tool and use that person’s publicly posted photos as a source for AI-generated imagery. The feature was designed to work with public accounts — accounts whose photos are already visible to anyone on the internet — but it came with a critical omission: the person whose account was referenced was not notified when their photos were used this way.
Why the Backlash Was So Fast
The problem with this feature is not difficult to understand. AI image generators can produce realistic-looking images of people, including images that portray subjects in situations, poses, or contexts they would never have agreed to. When a tool makes it easy to feed a specific person’s photos into that process by simply @-mentioning their account, the potential for misuse is obvious.
Social media and AI have already created serious problems in this area. AI has been misused to generate non-consensual intimate images of female celebrities and ordinary people — a pattern that has caused genuine harm and prompted legislation in several countries. The scale of that abuse has made platforms and the public acutely sensitive to any feature that makes it easier to target a specific individual’s likeness with AI generation tools.
Meta’s feature did not require any special access or approval. Any user, on any public account, could be referenced. That openness, combined with the absence of any notification to the referenced person, made the feature feel less like a creative tool and more like a surveillance-adjacent capability — one that could easily be pointed at people without their knowledge or consent.
Talent Agencies and Users Push Back
Backlash came quickly from multiple directions. Ordinary users pushed back on social media, and TechCrunch published a guide explaining how users could opt out of having their photos referenced by the tool. More significantly, major talent agencies also raised objections.
According to Puck News founding partner Dylan Byers — who was the first to report Meta’s decision to remove the feature — the reversal came “amid scrutiny from users and talent agencies, including CAA”. CAA (Creative Artists Agency) is one of the largest and most powerful talent agencies in the entertainment industry, representing thousands of actors, musicians, athletes, and other public figures. Its involvement signals that the concern went well beyond individual user complaints — it reached the institutional level.
Meta’s Statement: “Missed the Mark”
In a blog post published on July 10, 2026, Meta announced the feature’s removal with language that was unusually direct in its admission of error:
“Our intent was to provide a useful creative tool and to give people control over whether their public content could be referenced in this way,” the company wrote. “We’ve heard the feedback that this feature missed the mark, so it’s no longer available.”
The phrase “missed the mark” is notable for a company that typically frames product decisions in neutral or aspirational language. It is also somewhat incomplete as an explanation: the feature did not merely miss a mark, it raised a fundamental question about consent that Meta apparently did not adequately think through before launch.
A Pattern of AI Missteps on Social Platforms
Meta is not alone in struggling with the line between creative AI features and consent. The broader pattern across social platforms has been one of companies launching AI-adjacent features, facing backlash when users discover how those features can be misused, and then rolling them back or modifying them under pressure.
What makes the Muse Image situation particularly instructive is how predictable the problem was. Any feature that lets one user generate AI imagery of another user by referencing their public account — without notification or opt-in — is one that is virtually guaranteed to produce harmful uses. The question is not whether it would be misused, but at what scale and how quickly.
For talent agencies, creators, and ordinary users with large public followings, the stakes are especially high. Their public presence is part of their livelihood and identity. A tool that allows anyone to generate imagery of them — realistically, at scale, without their knowledge — represents a meaningful threat to their control over their own image.
What This Means for AI Image Generators Going Forward
Meta’s rapid reversal may actually set a useful precedent. By pulling the feature within days rather than defending it, Meta acknowledged that the threshold for rolling out AI image features involving other people’s identities needs to be higher than it was in this case.
For AI image generators broadly, the Muse Image incident is a reminder that technical capability and responsible deployment are not the same thing. Building a feature is not the same as establishing that the feature should exist. Any tool that makes it easier to generate realistic imagery of specific individuals — especially without notification or consent mechanisms — faces a burden of justification that goes beyond whether the images themselves look good.
Meta has not said whether it plans to bring back a modified version of the @-mention feature with different consent controls. For now, Muse Image continues to operate without it. How Meta approaches this feature in the future — and whether it can build the consent infrastructure that would make such a tool genuinely safe — will be worth watching.
Read more tech related articles here.
