Anyone can generate your AI image now: Meta’s Muse Image sparks ‘security warning’

Meta has just decided that anyone can generate AI images of another person from their public Instagram account, simply by typing that person’s handle into a prompt. The feature comes from Muse Image, an AI image generation model Meta launched earlier this week that integrates with public Instagram profiles.

Under Meta’s own policy, the person whose likeness is used is not notified when this happens, making it difficult to know if or how an account has been used to generate images. Meta describes the tool as a way to make AI image generation more personal by letting people reference public Instagram accounts in their prompts.

Cybersecurity firm Malwarebytes flagged the risk in a blog post, warning the tool could be used for impersonation, scams or other abuse. It noted that public Instagram photos were already being harvested by attackers to build deepfakes for identity verification fraud, and that Muse Image lowers the barrier to creating convincing synthetic images of real people.

An opt-out setting does exist: on Instagram, go to Profile, then Menu, then Sharing and reuse, and turn off the Posts and Reels toggles under ‘Allow people to reuse your content on Instagram and with AI features at Meta.’ The feature is still rolling out, starting in the US, so the setting may not yet be visible to every user.

Opting out only stops future image generation — any AI images already created using an account’s likeness remain in circulation and are not retroactively removed. Switching an account to private is currently the closest thing to comprehensive protection against the feature.

Meta’s AI tools have seen other security issues surface this year, including a ‘confused deputy’ flaw in its AI support chatbot that allowed unauthorized account changes such as password resets, an issue researchers said multi-factor authentication helped mitigate.