Meta launched Muse Image and then shut down its most controversial Instagram feature within days. The problem wasn't the model. It was the decision to treat people's public photos as available material unless they knew where to object.
Three days. That's how long Meta's Muse Image controversy needed to become a product retreat. The tool arrived in early July as part of Meta's push to put generative AI directly inside Instagram and its other apps. The feature that caused the trouble was plain enough: you could tag a public Instagram account with an @-mention and ask Meta's image model to generate a new picture using that account's public posts as a reference.
Users didn't opt in. Public accounts were covered by default, and account owners had to find the setting to stop other people from reusing their posts with AI features, according to guides published by Malwarebytes and Business Insider. Private accounts were treated differently, but that still left a huge share of Instagram exposed to a system most people didn't know existed until the warnings started circulating.
That's the part Meta should have understood before launch. A public post is not an open invitation to build a synthetic version of someone's face, body, photography style or personal archive. If you run a creator account, a small business account or just a normal public profile, you know the difference between being seen and being repurposed. Meta blurred that line and then asked users to clean it up in settings.
The backlash came fast. Axios reported that Public Citizen called the feature an egregious privacy invasion, while CAA pushed Meta to require opt-in consent for uses involving a person's name, image, likeness, voice or creative work. The Verge reported that Meta disabled the Instagram capability after criticism over identity misuse and sextortion risks - and warnings that user protection was too weak. That's a rare retreat for a company that has been trying to show investors it can ship consumer AI at speed.
The risk was obvious
The complaints were not abstract. Coverage from The Verge and other outlets described fears that the feature could be used for deepfakes and impersonation, or to generate sexualized images. Meta has said it built safeguards into Muse Image, but safeguards are not the same thing as consent. That distinction matters. If a stranger can start from your Instagram handle and make a new image of you, the burden has already been put in the wrong place.
This also lands on top of Meta's existing AI safety record. Reuters reported in 2025 on internal Meta AI standards that had allowed chatbots to engage in romantic or sensual conversations with minors, generate false medical information and produce racially biased responses. Meta later said the examples were inconsistent with its policies and had been removed. Fine. But when the next controversy is an image tool that again depends on users discovering the danger after launch, you don't get to call it a surprise.
The regulatory backdrop is just as uncomfortable. On July 10, the European Commission said in preliminary findings that Facebook and Instagram may breach the EU's Digital Services Act through features including infinite scroll, autoplay and personalized recommendation systems. The Commission said those designs can promote compulsive use and harm minors and vulnerable users. If the findings become final, Meta can face fines of up to 6% of global annual turnover.
So the timing could hardly be worse. Regulators are already asking whether Meta designs products around user welfare or engagement pressure. Then Meta ships an AI image feature that lets one user summon another user's public Instagram material into a synthetic picture without prior permission. You don't need a law degree to see the problem.
Consent has to come first
Frankly, the fix was obvious from the start. A tool that uses someone's likeness should be opt-in. Not opt-out. Not hidden under Sharing and reuse. Not announced through a wave of third-party explainers telling people how to protect themselves after the fact.
Meta will probably bring some version of this feature back. It has too much riding on consumer AI, and Instagram is too valuable a testing ground to abandon image generation after one messy week. OpenAI, Google, Adobe and xAI are all fighting for the same user habit: type what you want, get an image or video instantly. Meta wants that habit inside its own apps - where the photos and identities live, and the social graph too.
That is exactly why consent cannot be treated as an afterthought. The more personal the data, the less acceptable it is to rely on buried controls. If Meta wants Muse Image to be a creative tool rather than another deepfake panic, it has to start with the person whose image is being used, not the person typing the prompt.
The three-day shutdown is not the whole story. The real test is whether Meta treats it as a design failure or just a communications problem. Users will forgive a rough product more easily than a product that assumes their face is available until they object.
Also read: OpenAI's safety chief exits as the company folds oversight into research • Meta pulled its AI image tool from Instagram three days after launch • Zuckerberg Breaks His Three Year Silence On X To Launch Muse Spark 1.1
