Meta launched Muse Image on July 7, its first AI image generator from Meta Superintelligence Labs, the company’s dedicated AI unit led by Alexandr Wang. Internally code-named Mango, the tool is free for everyday use through the Meta AI app, Instagram Stories and WhatsApp, with paid tiers kicking in past a usage limit. It offers the standard capabilities of a modern image generator, including text-to-image creation, prompt-based editing, preset prompts and style transfers, plus Meta-specific touches like generating custom ads and previewing how a secondhand couch might look in a room through Facebook Marketplace. Meta says the model is competitive with leading generators on early benchmarks and that a video version, Muse Video, is already in development.
The launch was immediately overshadowed by one feature. Muse Image lets a user tag any public Instagram account and use that person’s publicly visible photos as the basis for a new AI-generated image, and the person is not notified.
Critically, the setting is opt-out by default, meaning anyone with a public profile is included unless they actively disable it. Meta’s own policy states that people may create content with your Instagram content using its AI features and that “you will not be notified” when they do. The design drew fast criticism after The Verge flagged it, with one widely shared post calling it “a privacy landmine waiting to detonate.” Meta counters that the feature only applies to content users chose to make public and that controls exist to turn it off.
The mechanism is a meaningful shift in how AI models use personal images. Rather than the familiar debate over scraping photos for training data, this makes a person’s likeness a live input at the moment of generation, which raises sharper questions about consent, impersonation and misuse. It is easy to imagine the feature being used to place someone in embarrassing, misleading or defamatory scenes without their knowledge, and the harm is amplified by Instagram’s scale and the realism of current image models.
Why It’s Drawing Scrutiny
Meta’s history is a large part of why the reaction has been so sharp. The company paid a then-record $5 billion FTC fine in 2019 after the Cambridge Analytica scandal, in which tens of millions of users’ data was harvested without their knowledge, and it shut down Facebook’s facial-recognition system in 2021 amid lawsuits over biometric data collection.
Against that record, an opt-out default that quietly enrolls users fits a pattern regulators and advocates have repeatedly flagged: broad use of people’s data unless they take action to prevent it. Placing the likeness feature inside Instagram Stories, one of Meta’s most-used surfaces, only widens the exposure. Privacy experts argue such a capability should be opt-in, requiring explicit consent before anyone’s face can be fed into a generator.
The Bigger Picture
The controversy partly obscures a genuine strategic push. Muse Image is another entry in a steady stream of Meta AI releases, following the Creator assistant and the game-making app Pocket, and it advances CEO Mark Zuckerberg’s effort to make AI creation mainstream and monetizable through ads and subscriptions, helping justify the tens of billions of dollars Meta is spending on AI infrastructure this year.
But the episode illustrates the central tension in that strategy: Meta’s biggest advantage in AI is its vast trove of user-generated content and social context, and its biggest liability is public distrust of how it uses exactly that data. Whether the likeness feature draws regulatory attention, particularly in Europe, may shape how aggressively Meta can lean on its social graph to power the next wave of AI products.
