Meta has switched off a newly introduced Muse Image feature that allowed people to use public Instagram profiles as visual references in AI-generated pictures. The reversal came only a few days after launch, following criticism that the default setup placed the burden on account owners to protect their own likenesses.
The distinction matters: Meta has not withdrawn Muse Image itself. The company removed the option to reference a public Instagram account by mentioning its username in a prompt. Muse Image’s other generation and editing tools remain part of Meta AI, based on the information Meta has published so far.
What Meta turned off
Muse Image arrived on July 7 as the first image-generation model announced by Meta Superintelligence Labs. It powers image features in the Meta AI app, Instagram and WhatsApp, with availability in other Meta services planned. In addition to producing pictures from text, it can edit an image from a sketch and transform supplied photos.
One launch feature went further. A user could mention a public Instagram account in a prompt, and Meta AI could reference photos from that account when building a new image. The person whose profile supplied the visual material did not have to approve each request. Meta provided a setting to opt out, but the capability was available by default for public profiles, according to reports about the rollout.
On July 10, Meta said that approach had “missed the mark” and that the feature was no longer available. The company described its original aim as giving people a useful creative tool while allowing them to control whether their public content could be referenced.
The launch reports concerned public profiles; they did not indicate that the feature exposed private posts. The dispute was about something different: whether making a profile visible should also count as permission for an AI tool to use its photos in a new way. Those are separate choices, even though the original setup effectively tied them together.
Why an opt-out was not enough
The objection was not simply that AI could imitate someone’s appearance. Image generators have been able to do that from uploaded photos for years. Muse Image reduced the friction: a public username could become the starting point for a synthetic scene, without the account owner initiating the request or explicitly agreeing to that particular use.
That design made an ordinary Instagram privacy decision carry a new consequence. Someone may keep a profile public so clients can find their work, relatives can follow them or an audience can see their posts. It does not necessarily follow that they expect strangers to place their likeness into generated situations. Moving the control into settings also meant that people had to learn about the feature before they could refuse it.
SAG-AFTRA, the union representing actors and other media professionals, urged Instagram users to opt out and argued that uses involving a person’s image should require clear, conspicuous consent. After Meta disabled the function, the union welcomed the decision and pointed to the risks of nonconsensual digital replicas. Other critics raised concerns about impersonation, harassment and sexualized fakes.
The speed of the reversal is notable too. Only a few days separated the introduction from the shutdown, and Meta did not identify a single incident that prompted the decision. Its public explanation and the available reporting point instead to the feature’s consent model and the response to it. That is different from evidence of a widespread abuse campaign, which the cited reports do not establish.
Muse Image has not been cancelled
Some headlines have described Meta as discontinuing an AI image feature, which can easily be read as the end of Muse Image. Meta’s statement is narrower. The public-profile reference function has gone; the wider model and its ordinary prompt, editing and transformation capabilities have not been announced as discontinued.
This also leaves several practical questions unanswered. Meta has not publicly said whether images already created with public-profile references will be deleted, retained or treated differently. It has not announced that the function will return with an opt-in system, nor has it provided a timetable for a redesigned version. For now, users should treat the removal as indefinite rather than temporary, but not assume that every policy question surrounding earlier outputs has been settled.
A lesson for consumer AI products
The episode shows how quickly a technically small feature can change the meaning of data that is already public. Instagram photos were visible before Muse Image launched. What changed was the ease with which another person could turn a profile into raw material for a generated composition.
For developers, the safer pattern is increasingly clear: sensitive uses of identity should begin with informed participation, not a setting that people discover after launch. An opt-in would not eliminate abuse, but it would establish that the account owner chose to make their profile available for this specific purpose. Additional limits on prompts, age, reporting and traceability would still be necessary.
For Instagram users, no action appears necessary to stop this particular form of profile referencing now that Meta says it is unavailable. Normal precautions still apply: a public account remains visible, its photos can be copied, and third-party tools are outside Meta’s direct control. The company’s reversal closes one unusually convenient route inside Meta AI; it does not make public images impossible to misuse elsewhere.
Join the conversation
Stay on topic and respect other readers. Your first comment may appear after editorial review.