Meta said Friday it is shutting down an AI image-generation feature it launched only days earlier, after the tool drew sharp criticism over privacy concerns, including a formal objection from a Hollywood union.
“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,” Meta said in a statement. “We’ve heard the feedback that this feature missed the mark, so it’s no longer available.”
A Feature Built on Public Photos
Meta, which owns Facebook and Instagram, introduced the tool, called Muse Image, on Tuesday. It marked the company’s first image-generation model to come out of Meta Superintelligence Labs. The feature was built into the Meta AI chatbot and let users generate images using photos as input, then edit the results directly through sketches drawn on top of the generated image.
The trouble was where those input photos could come from. Muse Image let users generate images using content pulled from public Instagram accounts, meaning a person’s public photos could be fed into the tool without that person taking any action to allow it. The feature also came switched on for users by default, rather than requiring anyone to opt in before their content could be used this way.
That combination, automatic activation plus access to public account content, is what triggered the backlash.
Backlash Builds Fast
Hannah Einbinder, the Emmy-winning actor known for her role on “Hacks,” posted criticism of the feature on Instagram, telling followers it had been turned on automatically and urging them to switch it off.
SAG-AFTRA, the union representing actors and other media professionals, went further. On Thursday, the union publicly urged its members and other Instagram users to opt out of the feature entirely.
“Anything other than a clear and conspicuous opt-in for these types of uses of Instagram users’ images is unacceptable, and an utter miscalculation of public sentiment regarding the obvious dangers and harms inherent in such use,” SAG-AFTRA said in its statement.
The union’s objection centered on a concern that has followed AI image tools broadly: the risk that photos of real people can be used to create digital replicas or likenesses without their direct consent, even when the source material was already public. For an organization representing performers, whose faces and images carry professional and financial value, an opt-out model rather than an opt-in one raised immediate red flags.
Meta Reverses Course
Meta’s decision to pull the feature came just three days after its launch, a fast reversal for a company that had positioned Muse Image as a flagship product from its newly formed superintelligence research group. The company did not detail how many users had engaged with the tool before shutting it down, nor did it specify whether it plans to relaunch a modified version with different consent controls in the future.
SAG-AFTRA welcomed the reversal once it was announced.
“With the dangers of nonconsensual digital replicas well known to all, a feature that encouraged that behavior is unwise. We appreciate its discontinuance. It is the responsible thing to do,” a union spokesperson said.
A Pattern Tech Companies Keep Repeating
The episode fits a pattern that has played out repeatedly as tech companies race to ship generative AI features. A company builds a tool that draws on user content by default, faces immediate criticism once users realize their material is included without a clear opt-in step, and then pulls or modifies the feature within days.
What makes this case notable is the source of the pushback. SAG-AFTRA’s involvement signals that concerns about AI and digital likeness now extend well beyond Hollywood contract negotiations and into how ordinary Instagram users’ public photos get treated by AI systems. The union spent much of 2023 and 2024 fighting for contract protections against AI-generated performances; this dispute shows those same consent principles being applied to a mainstream consumer product with no direct connection to film or television production.
For Meta, the incident lands at a sensitive moment. The company has been investing heavily in its Superintelligence Labs division and pushing to establish itself as a serious competitor in AI image and video generation, an area where it has trailed competitors like OpenAI and Google. Muse Image was meant to show that investment paying off. Instead, its rollout became a case study in what happens when a company treats public content as fair game for AI training or generation without building in a clear, upfront choice for the people whose photos are involved.
Meta has not said whether Muse Image will return in a different form, or whether the company plans to build opt-in consent mechanisms into future features drawing on user or public content. The company’s statement framed the shutdown as a response to feedback rather than an admission that the underlying design was flawed, language that leaves open the possibility of a revised version down the line.
For now, the reversal leaves Meta without an image-generation feature it had promoted as a signature launch just days earlier, and it hands SAG-AFTRA a concrete example to point to in ongoing arguments about how tech companies should handle consent when AI tools touch real people’s images.
