Sources
- TechCrunch AI
Make it yours
Inspired by this story? Turn the idea into your own AI art in seconds — free to start, no card required.
Start creating free
Inspired by this story? Turn the idea into your own AI art in seconds — free to start, no card required.
Start creating freeMeta's Muse Image model, now live across Instagram and the Meta AI app, lets any user tag a public Instagram account and pull that account's photos into AI-generated creations — without the account owner's knowledge or consent.
Muse Image, built by Meta's Superintelligence Labs, was designed to generate high-quality images from text prompts and reference photos. The reference-photo mechanic is where the privacy issue sits: when a user tags a public Instagram account in a prompt, Muse Image treats that account's public posts as fair-game source material. According to TechCrunch, as long as a profile is public, another user can use those images as part of an AI-generated creation — no additional permission required.
That's a meaningful distinction from earlier AI training debates, which centered on whether companies used public data to train models. This is different: Muse Image is using your photos at inference time, in real-time generation, at another user's request. The output could be a stylized portrait, a character render, or a composite scene — built from your face, your art, or your branded imagery.
Meta has not added a standalone Muse Image opt-out toggle in its AI settings. The protection is account-level: switching Instagram to a private profile prevents Muse Image from accessing your posts when another user tags you.
To do that on iOS or Android, go to Settings and privacy → Account privacy → Private account. On desktop, the path is Settings → Privacy → Account privacy. Once private, only approved followers can see your posts, which removes them from Muse Image's reach.
There are also narrower controls worth checking. Under Settings → Privacy → Photos of you, you can limit who can tag you in posts — though this controls tagging by other users, not necessarily how Muse Image processes a prompt that references your handle. Meta's AI data controls, found under Settings → Meta AI → Data and privacy, let you submit an objection to your information being used for AI, but Meta's own documentation describes this as a request, not a guaranteed block.
For AI-art creators, the tension here is real and immediate. A public Instagram account is standard practice for building an audience, attracting clients, and getting discovered. Going private protects your images from Muse Image but cuts off organic reach entirely — follower requests replace open discovery.
Creators who use Instagram primarily as a portfolio showcase face a binary choice that didn't exist six months ago: visibility or control. Watermarking images doesn't block Muse Image from using them; it only marks the output if the watermark survives generation, which is not guaranteed.
If keeping a public portfolio is non-negotiable, the practical middle ground is to audit what you post. High-resolution, unobstructed face photos and clean reference-quality images are the most useful inputs for a model like Muse Image. Cropped, stylized, or compositionally complex images are harder to repurpose cleanly. That's not a solution — it's a mitigation — but it's a real one.
Meta has not announced any timeline for adding a dedicated Muse Image opt-out. Until that changes, the privacy setting is the only lever creators actually control.