If your Instagram account is public, Meta has already opted you into letting strangers generate AI images of you. The company launched Muse Image on July 7, its first image-generation model out of Meta Superintelligence Labs, and buried inside the rollout is a default setting that lets anyone in the Meta AI app @-mention your Instagram username and pull your public photos into an AI-generated image. You won’t get a notification when it happens.
Meta’s own description of the feature is blunt about what it does: tag a username in the Meta AI app and it uses that account’s public photos to build a visual that’s ready to post. An Instagram help page confirms that content on public accounts can be used to create AI images, and that account owners aren’t told when their content gets used this way. The setting is on by default for every public Instagram account, and turning it off requires actually finding the toggle.
Here’s how: open Instagram, go to settings, scroll to Sharing and Reuse, and switch off both “Posts” and “Reels” under the option labeled “Allow people to create with and reuse your content on Instagram and with AI features at Meta.” Two catches worth knowing:
- Turning the setting off does not delete any AI content already generated from your photos before you flipped the switch.
- Muse Image and its opt-out toggle are still rolling out, so the setting may not appear in your app yet even if you go looking for it.
Private Instagram accounts are excluded from Muse Image entirely, which limits the exposure somewhat. But “public by default, opt-out if you notice” is a pattern Meta has run before, and it puts the burden on users to discover a setting most of them don’t know exists.
The @-mention usage is where this gets uncomfortable. Wang’s pitch treats tagging a friend as a fun shortcut, but the same mechanism lets anyone pull a stranger’s public photos into an AI image without asking. That’s a consent gap the industry has already been burned by once this year, when X’s Grok chatbot let users generate non-consensual “undressing” images of real people before X shut the feature off for non-subscribers. Meta hasn’t built anything as explicitly abusable as Grok’s tool, and private accounts are walled off, but an opt-out-by-default system for public photos is still asking users to trust that nobody misuses a feature Meta itself is marketing as easy and shareable.
Apple’s approach to Apple Intelligence has been the opposite of this all along, gated behind explicit permissions, on-device processing where possible, and features that mostly ask before they act. Meta is racing to ship AI everywhere, on every surface, before Apple even ships its rumored smart glasses, and Muse Image is a reminder that the price of that speed keeps landing on users who never asked to be part of the training set or the content pipeline.
