Meta's New AI Image Generator Uses Your Instagram Face by Default — Opt Out Before Someone Else Does

Meta launched Muse Image on July 7, calling it “the most advanced image generation model yet.” By July 9, privacy advocates were telling every public Instagram account holder to check their settings — because Meta had already opted your face into the system.

🔍 THE BOTTOM LINE

Meta’s new AI image generator lets anyone tag your public Instagram profile in a prompt and generate images using your face, without notifying you. The opt-out is buried four taps deep in Instagram’s settings. Privacy experts from the EFF to Proton are calling it a consent crisis. If your Instagram is public, you need to act now — or accept that your face is now raw material for a stranger’s AI prompt.

What Muse Image Actually Does

Meta’s announcement describes Muse Image as an AI image generator integrated directly into Instagram, WhatsApp, and the Meta AI app. The feature lets users tag any public Instagram account in a text prompt and generate images that pull from faces and photos featured in that account’s posts.

The key word is “public.” If your Instagram account is set to public — and millions are — Meta has automatically enabled your photos as source material. You were not asked. You were not notified. And according to Meta’s own policy page, you will not be told when someone uses your likeness: “You will not be notified about content created using AI features at Meta.”

Meta says private accounts and users under 18 are automatically excluded. But as The Guardian reported, the company has not clarified whether children depicted in photos on adults’ public accounts could be pulled into prompts. Proton’s privacy team flagged this gap directly: children in public photos risk having their faces appropriated by a stranger’s AI generation.

The Opt-Out Maze

Finding the setting to disable Muse Image’s access to your photos is, in Malwarebytes writer Danny Bradbury’s words, “its own adventure.”

Here’s the path: Instagram → Profile → Menu (three lines, top right) → Sharing and reuse → under “Allow people to reuse your content on Instagram and with AI features at Meta” → toggle off Posts and Reels.

Proton’s analysis is more blunt: “Data sharing is turned on by default, the opt-out is buried deep in settings, and public backlash becomes the main way users find out what happened to their content.” They also warn that the on/off toggles look nearly identical at a glance — easy to leave one active by mistake.

Opting out only prevents future image generation. Any AI images someone already created of you before you found the setting remain in circulation. There is no retroactive removal.

The only comprehensive solution is switching your account to private. It is, as Malwarebytes notes, a blunt instrument — but it works.

Why This Matters Beyond Instagram

The pattern is familiar. Google did it with AI training on Search image uploads. Grok did it with image generation on X. Meta itself previously did it with scanning camera roll photos on Facebook. Each time: opt-out by default, buried settings, public outrage as the discovery mechanism.

What makes Muse Image different is the directness of the harm vector. Previous AI training controversies involved abstract data pipelines — your photos improved a model somewhere. Muse Image lets a specific stranger type your handle and get a specific image of your face back. The gap between “your data trained a model” and “someone just generated a picture of you” is the gap between a background risk and a foreground threat.

Thorin Klosowski, a senior privacy activist at the Electronic Frontier Foundation, told The Guardian: “This is the sort of setting that should absolutely be opt-in for Instagram users.” The EFF does not say things like that lightly.

Meta’s own Oversight Board has already said the company needs stronger detection tools and better labeling of AI-generated content. When your own governance body says your defenses aren’t enough, consumers should take notice.

The Deepfake Pipeline Problem

Public Instagram photos were already being harvested by attackers for identity verification fraud, creating deepfakes for phishing and impersonation scams. Muse Image lowers the barrier from “attacker scrapes photos and runs a custom model” to “anyone types a handle and clicks generate.”

This connects to a pattern this site has tracked: AI deepfakes driving violence against women and deepfake scam calls surging. Each new tool that makes synthetic image generation easier expands the attack surface. Meta just built a front door into it and left it open by default.

NZ Angle

New Zealand has no specific AI-generated content law. The Privacy Act 2020 covers the collection and use of personal information, but it was written before AI image generators could produce a convincing portrait from a single tagged Instagram handle. The Privacy Commissioner has flagged AI as a priority area, but enforcement depends on complaints — and most New Zealanders won’t know their face was used until someone tells them.

For Kiwi Instagram users, the practical advice is the same as everywhere: check your settings. If you run a business or public figure account and need it public, toggle off the AI reuse setting. If you’re a personal user who doesn’t need public visibility, consider going private. And if you have children in your public photos, the calculus is simpler: turn it off.

❓ FAQ

Can someone generate AI images of my face right now? If your Instagram account is public and you haven’t toggled off the reuse setting, yes. Muse Image is rolling out in the US first, so you may not see the setting yet — but the feature may already be accessible to others.

Does opting out remove AI images someone already made of me? No. Opting out only prevents future generation. Any images created before you changed the setting remain in circulation.

Are children protected? Meta excludes accounts belonging to users under 18 from being tagged. But the company has not clarified whether children who appear in photos on adults’ public accounts are also protected. Privacy advocates say this gap is a serious risk.

What’s the difference between this and Meta training AI on my photos? AI training uses your data to improve a model’s capabilities over time. Muse Image lets a specific person generate a specific image of your face on demand. The harm is immediate and targeted, not abstract and aggregate.

Is going private the only safe option? It’s the only comprehensive one. Opting out of the reuse setting prevents AI image generation, but your photos remain publicly visible. If you want zero risk of your face being used as AI source material, a private account is the only guarantee.

🔍 THE BOTTOM LINE

Meta has turned every public Instagram account into an involuntary face database for its AI image generator. The opt-out exists but is designed to be found only by people who already know to look for it. The EFF says it should be opt-in. Meta’s own Oversight Board says the company’s defenses are insufficient. Until the default flips, the responsibility falls on you — four taps into Instagram settings, or one switch to private.