Meta has removed an Instagram-linked artificial intelligence feature just three days after its launch, following growing criticism over privacy, consent and the use of public photos in AI-generated images.
The feature allowed people to reference public Instagram accounts while creating images through Meta AI. By entering a username, users could ask the system to produce a new visual based on photos available on that account.
Meta introduced the function on July 7 as part of Muse Image, its new image-generation model. It was withdrawn on July 10 after users, performers and privacy experts questioned why public accounts had been included automatically instead of being asked for permission.
“We’ve heard the feedback that this feature missed the mark, so it’s no longer available,” Meta said in an update announcing the change.
The company has not removed Muse Image itself. Other tools, including text-to-image generation, photo editing, image combinations and visual effects, remain available. The reversal applies only to the option that allowed one person to reference another person’s public Instagram profile.
Public Photos Enter AI
Meta initially promoted the feature as a way to make AI-generated content more personal and social.
People could mention a public Instagram account in Meta AI and ask it to create invitations, collaborative graphics or other images inspired by photographs posted on that profile. The company suggested the tool could help friends, businesses and creators produce content connected to people they already followed.
The problem was how the system handled consent. Public Instagram accounts were reportedly included by default. Account holders who did not want their photos used as references had to find the relevant controls and disable them manually.
The setting appeared inside Instagram’s “Sharing and reuse” controls, where users could turn off access for posts and Reels. Many account holders were unaware that the option had been activated or that their public content could be used in this way.
Critics also raised concerns that people might not be notified when their accounts were referenced in an AI-generated image.
That distinction quickly became central to the backlash. Posting an image publicly allows other users to view it, share it or engage with it within Instagram. Critics argued that using the same image to guide the creation of a new synthetic picture was a different form of reuse and should require separate permission.
Consent Model Draws Criticism
The controversy spread rapidly across Instagram and other social platforms, where users began circulating instructions showing others how to disable the setting.
Privacy advocates said the feature placed too much responsibility on users. Instead of asking people whether they wanted to participate, Meta included them and expected them to opt out later.
That approach became especially sensitive because Instagram profiles often contain clear photographs of faces, family members, workplaces, homes and everyday routines.
Public figures were an obvious concern, but the feature also affected ordinary users, small businesses, photographers, journalists and creators who may not have expected their images to become source material for AI generations. Security experts warned that the tool could make it easier to create misleading or harmful images involving real people. A feature designed for playful or creative use could also be used for impersonation, harassment, scams or false representation.
The risk did not depend on the output being fully realistic. Even an image that appeared edited or artificial could spread without context and create confusion about where it came from or whether the subject had approved it. Concerns also emerged over images produced before an account holder disabled the setting. Turning off access could prevent future use, but it did not necessarily remove images that had already been generated.
Performers Join Backlash
The entertainment industry quickly became one of the strongest critics of the feature.
SAG-AFTRA urged members and other Instagram users to disable the setting, arguing that people should have the right to decide whether their names and likenesses are used in AI-generated content.
The union has spent years pushing for stronger safeguards around digital replicas, particularly as studios and technology companies develop tools capable of recreating performers’ faces and voices.
After Meta removed the feature, SAG-AFTRA welcomed the decision. “With the dangers of nonconsensual digital replicas well known to all, a feature that encouraged that behavior is unwise,” the union said.
Creative Artists Agency also criticised the rollout, saying that names, images, likenesses, voices and creative work should not be used by AI systems without clear and documented consent.
Actor Hannah Einbinder was among those who publicly drew attention to the setting and encouraged users to switch it off. Posts from performers and creators helped bring the issue to a wider audience beyond privacy and technology circles.
The reaction showed how quickly a consumer AI tool could become part of a much larger dispute over identity rights, creative ownership and control over personal data.
Muse Image Remains Available
Meta’s broader image-generation plans have not changed. Muse Image can still create pictures from text prompts, transform existing photos, combine multiple visual references and respond to sketch-based edits. Meta has also connected the model to AI effects inside Instagram Stories and image-generation features in WhatsApp in selected markets.
The company plans to expand Muse Image across Facebook, Messenger, Instagram and WhatsApp. It is also expected to bring the technology into advertising products, including creative tools used by businesses and agencies. That makes the removal of the Instagram account feature a targeted retreat rather than a reversal of Meta’s wider AI strategy.
The company continues to treat image generation as an important part of its consumer, creator and advertising platforms. However, the controversy has exposed the risks of connecting AI tools directly to the large amount of personal content already stored across social networks.
Public Does Not Mean Permission
The dispute has highlighted a growing tension at the centre of social AI products.
Technology companies often treat publicly available content as material that can be searched, analysed or referenced. Users may see the same content differently, particularly when it includes their face, identity or personal life.
A photograph being visible online does not necessarily mean the person in it has agreed to have their likeness reconstructed, altered or placed in a fictional scene.
Meta said the feature was designed to give users control through account settings. The backlash suggested that control offered after automatic inclusion was not enough. For features involving personal identity, users increasingly expect clear permission requests, visible notifications and simple ways to understand how their content may be used.
The speed of Meta’s reversal is notable. Muse Image was introduced as an important product from the company’s new AI division, yet one of its most distinctive functions lasted only a few days. The episode is likely to influence future launches across the industry. Social platforms developing similar tools may face pressure to adopt opt-in systems, restrict the use of real people’s likenesses and alert account holders whenever their content is referenced.
Meta has kept its image model in place, but the message from the backlash was clear: access to public photographs does not automatically provide permission to turn them into AI-generated content.
Comments