# Meta’s new AI image generator is using your public Instagram photos > Source: > Published: 2026-07-08 15:54:41.423865+00:00 Meta launched an [AI image generator](/lumo/image-generation) called Muse Image this week, built directly into [Instagram](/blog/instagram-anonymous-story-viewer), [WhatsApp](/blog/is-whatsapp-safe), and the [Meta AI](/blog/turn-off-meta-ai-facebook) app — with Facebook and Messenger expected to [soon follow(new window)](https://about.fb.com/news/2026/07/introducing-muse-image-meta-ai/). But there’s a major privacy concern: If your Instagram account is public, Meta has already opted your photos into features that can be used for generative AI remixes. ## How the Meta AI image generator works Meta’s Muse Image tool allows anyone to tag your public profile in a prompt and generate AI images using your photos. This can include your face, your likeness, and other people who appear in your photos, including children with no power of consent. You don’t get asked first. And, according to Meta’s [own policy(new window)](https://help.instagram.com/652943337035777), you don’t get told after either: “You will not be notified about content created using AI features at Meta.” ### Anyone can use Muse Image with your public Instagram photos [Reporters testing this new feature(new window)](https://gizmodo.com/if-you-have-a-public-instagram-account-you-might-be-surprised-what-ai-users-can-now-do-with-your-face-2000782694) successfully generated images of people they’d never followed or interacted with, without their consent. Some on social media have called the feature “a privacy landmine waiting to detonate” (via [X(new window)](https://x.com/mindinpanic/status/2074603011389657563)), pointing to a pattern of harm from non-consensual AI image tools on other platforms. The rollout of Meta’s Muse Image tool follows a familiar pattern from major tech companies: Data sharing is turned on by default, the opt-out is buried deep in settings, and public backlash becomes the main way users find out what happened to their content. Google did this with AI training on Search image uploads, Grok did it with image generation on X, and even [Meta did this with scanning camera roll photos on Facebook](/blog/meta-ai-facebook-camera-roll) and [end-to-end encryption on Instagram](/blog/instagram-end-to-end-encryption). ## How to opt out of Meta AI image generator features on Instagram If you don’t feel comfortable with Meta having broad access to your Instagram public content, especially since it may be used for AI features or [AI training](/blog/family-photos-ai-risks), here’s what you can do: - Open Instagram and go to your profile. - Tap the three lines in the top right corner, then select **Sharing and reuse**. - Under **Allow people to use your content on Instagram and with AI features on Meta**, toggle off both** Posts**and** Reels**. **Watch the toggles closely**. The on and off states look nearly identical at a glance, and it’s easy to leave one active by mistake. To confirm a toggle is off, the black button needs to sit on the right side of the switch. Check Posts and Reels separately, as they don’t move together. This only stops new AI remixes made with the Meta AI image generator — images already made from your account stay up, and this can’t be undone. A safer way to prevent mishaps is by setting your Instagram to private (under **Settings** → **Account privacy**), since a private account is visible only to approved followers. Instagram, like all Meta platforms, is not designed to be a fully private space. [Meta is one of the](/blog/meta-ai-ads) [biggest advertising machines](/blog/meta-ai-ads), and its business model depends heavily on collecting, analyzing, and monetizing user data. If you’re looking for stronger privacy by default, consider exploring private [European social media apps](/learn/european-alternatives/european-social-media-apps). ## Try a private AI image generator If you want to use an AI image generator without worrying that your personal photos, prompts, or creative ideas may be mishandled, Lumo is built around privacy from the start. Your images and prompts are never used to train models or influence outputs for other users. It uses [zero-access encryption](/learn/encryption/types-of-encryption/zero-access), meaning your uploaded images and generated visuals stay private, encrypted, and invisible to prying eyes — including us.