A new default setting means your public photos can fuel AI images made by total strangers, and you won’t even be told.

Meta rolled out its first AI image generator this week, and the launch came with a detail that’s already drawing criticism. The tool, called Muse Image, was built by Meta Superintelligence Labs to compete with OpenAI’s GPT Images 2.0 and Google’s Nano Banana 2. But its integration with Instagram sets it apart from the usual text to image tools flooding the market. Anyone can tag a public Instagram account inside the Meta AI app, and the person’s photos become raw material for a generated image, no permission required.

Meta describes this as a way to personalize content, useful for a mocked up event invitation or a quick creative concept. The company’s messaging frames tagging a username as a shortcut. What it does not frame clearly is that this applies by default to every public account, and the person being tagged has no say in the matter unless they’ve already changed a setting most people don’t know exists.

Opting out takes some digging

Turning this off is not as simple as flipping a switch on the main privacy screen. Users have to open Instagram, tap their profile, hit the three lines in the corner, then find Settings and activity, then Sharing and reuse. From there, a toggle labeled ‘Allow people to create with and reuse your content’ needs to be switched off for both posts and reels. It’s a multistep process buried several layers deep, and when reporters checked their own accounts this week, the updated language describing AI use hadn’t even appeared yet in the settings menu for some users.

Switching to a private account stops future use of your content, and Meta says that if an account goes private for more than 24 hours, existing AI remixes made from it get deleted too. But nothing forces the deletion of AI images already generated from public photos before a person opts out or locks their account.

No notifications, no easy answers

Perhaps the most jarring detail is that Instagram will not tell you when someone uses your photos this way. A support document confirms that people are not notified about AI content created using their images, even though notifications still go out for more ordinary content reuse like stickers or templates. For anyone hoping to track how their likeness ends up circulating, that silence removes one of the only ways to find out.

Privacy advocates have not held back. Foxglove’s Donald Campbell called the setup a recipe for disaster, and Privacy International framed it as another example of AI companies treating personal images as material to mine rather than protect. Regulators are already circling similar territory. Ofcom is investigating X over Grok generated images made without consent, a sign that Meta’s rollout is landing in an environment already sensitive to this exact kind of harm.

A pattern beyond Instagram

Muse Image fits into a broader shift where AI features arrive switched on by default, leaving users to opt out rather than in. Google recently began storing search related media by default to train its own AI systems, a similar move that quietly expanded how personal content gets used. Meta says Muse Image is heading next to Facebook and Messenger, and a video generation version is reportedly already in the works. For now, anyone who wants to keep their Instagram photos out of someone else’s AI project needs to go looking for that setting themselves, because Meta isn’t going to point it out.