Meta's New AI Tool Quietly Turns Your Instagram Photos Into Fodder for Strangers' Images

Key points

• Meta launched a new AI image tool called Muse Image inside its Meta AI app, and it is turned on by default for public Instagram accounts.
• Any user can type @yourhandle into the app to pull your public photos and reels into AI-generated images.
• Meta says it will not train its models on the images produced this way, but the feature raises fresh questions about consent and impersonation.
• Users who want out have to change their Instagram privacy settings themselves; there is no separate opt-out inside Meta AI.
• Security researchers and privacy advocates warn the feature is a gift to scammers building fake personas and romance-scam profiles.

Meta has switched on a new feature that lets strangers pull your public Instagram photos into AI-generated images, and you did not have to agree to it first.

The tool is called Muse Image. It lives inside the Meta AI app, the company's rival to ChatGPT. In a post announcing the launch, first reported by The Hacker News, Meta said users can "@-mention Instagram accounts in the Meta AI app to bring specific Instagram profiles right into your images."

In plain English: if your Instagram is public, anyone with the app can type your handle and generate pictures using your face, your outfits, your holiday snaps.

Meta pitches this as helping people design party invites or personalised cards. In practice, the failure mode here is obvious the second you say it out loud.

What can someone actually do with your photos?

Anything they can prompt an AI to draw. That is the point of these tools.

A stranger can pull your public selfies into a fake beach photo, a fake gym photo, a fake photo of you somewhere you have never been. Meta has guardrails to block the worst stuff, nudity, violence, recognisable public figures in compromising scenes, but guardrails on generative AI, the kind of software that invents new images from a text description, are notoriously easy to talk around.

The people who spend all day writing romance scams and fake-influencer accounts are going to love this.

One thing the post-mortem will say, eventually, is that the default should have been off. Meta chose the opposite. If your Instagram is public, which is the default when you sign up, Muse Image treats your photos as raw material for anyone else's prompt.

How do you turn it off?

There is no dedicated switch inside Meta AI. You have to change your Instagram account itself.

Open Instagram. Go to Settings, then Account privacy. Flip your account to private. That is it. Once your profile is private, Muse Image cannot reach your photos through the @-mention trick.

You can also block the Meta AI account directly, though a private profile is the cleaner fix.

For businesses and creators who need a public profile, the choice is uglier. You either accept that your professional photos are now a public asset for AI remixing, or you go private and lose reach. Meta has not offered a middle path.

Should people be worried?

If you use Instagram casually and your account is public, yes, a little. Not because Meta is going to do something dramatic with your face, but because everyone else on the app now can.

The practical risks are the same ones we already see with scraped social media photos: impersonation, catfishing, fake dating profiles, harassment campaigns that put a target's face in embarrassing fake scenes. Muse Image just makes the workflow one tap easier.

Meta says it will not use images generated through Muse Image to train future models. That is a narrow promise. It says nothing about what other users do with the pictures they generate of you, which is the actual problem.

Operational takeaway: if you would not hand a stranger your photo album, set your Instagram to private today.