Meta pulled the Instagram feature that let users generate AI images of any public account by @-mentioning it, four days after launch. The tool, powered by Meta's new Muse Image model, went live Tuesday and was disabled Friday, July 10, after backlash from advocacy groups, the Screen Actors Guild, and Instagram users who said the design turned every public profile into raw material for someone else's AI generations without permission.
The mechanic was straightforward and, to critics, indefensible. Any Instagram user could open Meta AI, tag a public account, and produce an image referencing that person's likeness or content. Consent was opt-out, not opt-in, and the opt-out sat several menus deep in account settings. For the days the feature was live, the default state for every public Instagram account on the platform was 'available as an AI reference'.
Meta acknowledged the reversal in an update appended to its original Muse Image blog post. "Earlier this week, we announced that one way for people to generate images in Meta AI is by @-mentioning public Instagram accounts that they want to reference," the company wrote. "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. We've heard the feedback that this feature missed the mark, so it's no longer available."
Key facts
- 01Meta shut down the Instagram @-mention AI image feature on Friday, July 10, four days after announcing it on Tuesday.
- 02The feature let any user generate AI images from public Instagram accounts using Meta's new Muse Image model without account-owner permission.
- 03Opt-out was possible but required digging through settings, which critics called an unacceptable design burden.
- 04The Screen Actors Guild formally advised its members to opt out and published instructions before the feature was pulled.
- 05The National Center on Sexual Exploitation warned the feature was a ready-made tool for sextortion and impersonation scams.
The sharpest criticism came from Haley McNamara, executive director and chief strategy officer of the National Center on Sexual Exploitation, who flagged the feature on Friday as a distribution mechanism for non-consensual imagery.
The Screen Actors Guild moved in parallel, telling its members to opt out and publishing step-by-step instructions on how to do so. That is a notable posture from a union that spent much of the past two years negotiating AI-likeness protections into its film and television contracts. A tool that turned a public Instagram grid into training and reference material for AI image generation cuts directly across those provisions, regardless of what Meta's terms of service technically permitted.
Muse Image itself, the model underneath the feature, is not being pulled. It remains the image-generation engine inside Meta AI, and users can still generate images from text prompts or from their own content. What Meta removed is the specific interaction pattern that allowed one user to invoke another user's public account as a reference without that account holder's affirmative consent.
The episode fits a broader pattern for Meta in 2026. The company has been pushing AI features into Instagram and Facebook aggressively, with Adam Mosseri recently arguing that AI-generated content should live in its own feed rather than be filtered out of the main one. The Muse Image tagging feature was on the more permissive end of that strategy — and it lasted four days.
Opt-out defaults are the crux of the fight. For a feature that lets strangers generate images of you, the burden of protecting your likeness should not sit with you having to find and toggle a setting. The economics of scale make this obvious: a platform with hundreds of millions of public accounts cannot reasonably expect each account holder to preemptively defend against a feature they did not know was launching. Consent frameworks that work at the scale of AI generation have to be opt-in.
Meta has not said whether the @-mention pathway will return in a modified form, whether public accounts will be asked to opt in before any similar feature ships again, or whether Muse Image will get new guardrails against likeness generation more broadly. The company's statement addressed only the removal, not the roadmap.
The reversal is a data point on how fast AI product launches can now be forced into retreat when the design missteps on consent. Four days is not a review cycle — it is a real-time correction driven by advocacy groups, a major union, and users making the cost of the feature visible before Meta had finished rolling it out. For every AI company shipping generative features on top of existing social graphs, the working assumption should now be that opt-out likeness tools will not survive contact with the public. The next iteration of this feature, whenever it ships, will have to start from consent — or it will not ship at all.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
