Meta’s Muse Image Lets Anyone Use Your Instagram Photos in AI

Meta launched Muse Image on Tuesday, a new AI image generator built by Meta Superintelligence Labs. The model lets any user pull a public Instagram account into an AI-generated picture by tagging the handle, and the feature is on by default. The launch is Meta’s second major foundation model this year, following the April debut of the Muse Spark large language model.

What Muse Image Does and Where It Lives

Muse Image is the first in-house image generation model from Meta Superintelligence Labs, internally code-named Mango before launch. The unit is led by Alexandr Wang. The launch is part of Meta’s effort to ship its own foundation models after years of leaning on outside labs.

The model is free to use through the Meta AI app, meta.ai, Instagram Stories in the US, and WhatsApp in limited countries, with Facebook and Messenger set to follow later this year, per the official launch post on Muse Image and Muse Video. Power users who exceed a free limit can subscribe to one of Meta’s subscription plans, the company said. A user can ask the model to mock up a portrait in front of a historical landmark, erase a photobomber from a shot, or generate a working QR code from a prompt, Meta said. Meta also said Muse Video, a video version of the model, is “already in development.”

The Photo-Tagging Feature Drawing the Backlash

The single feature drawing the most attention is also the simplest. A user can open the Meta AI app, type the @ handle of a public Instagram account, and ask Muse Image to generate a new picture using that account’s photos. The feature is opt-out by default, and Meta’s policy says users “will not be notified about content created using AI features at Meta.” Updated help pages disclose that “other Instagram users may be able to create new reels, posts or stories that reuse part or all of your published photos” with AI features at Meta, and that the result “may be discoverable in search engine results.” Meta says users “have control” through a setting that has to be toggled off.

After The Verge first reported the design, the concern spread. A user on X, Volodymyr Pavlenko, put it bluntly.

Pulling real users into generated photos without explicit consent is a privacy landmine waiting to detonate.

That post calling the feature a privacy landmine was quoted across the coverage of the launch. Gizmodo’s own test went a step further: a reporter was able to generate an AI image of a real-world friend they had never followed on Instagram, “essentially a digital stranger,” without asking that person’s permission. Meta did not dispute the test, and pointed to “built-in protections” against violent, sexual, or defamatory AI imagery of real people.

A Privacy Pattern That Already Cost Meta $5 Billion

The Muse Image default is not Meta’s first time shipping a photo-related product on an opt-out basis. In 2019, the company paid a then-record $5 billion fine to the Federal Trade Commission over Cambridge Analytica, after regulators found that the political consulting firm had improperly harvested data from tens of millions of Facebook users without their knowledge to build voter-targeting profiles ahead of the 2016 U.S. election. Facebook had known about the data misuse for years before it became public. The FTC called it “the largest ever imposed on any company for violating consumers’ privacy” in a press release announcing the $5 billion Facebook privacy fine, and the order imposed a 20-year privacy regime on the company.

In 2021, Meta shut down Facebook’s facial-recognition system. That tool had automatically recognized people in photos and videos, and powered the platform’s “tag suggestions” feature. The move came amid lawsuits and regulatory pressure over the company’s collection of biometric data.

Meta’s privacy record at a glance:

• $5 billion: FTC fine paid in 2019 over Cambridge Analytica
• 20 years: Length of the FTC’s 2019 privacy order against the company
• 2021: Year Meta shut down Facebook’s facial-recognition system
• Tens of millions: Facebook users whose data Cambridge Analytica improperly harvested

That history sits a click away from any new Meta product built on user photos. Critics and regulators have long flagged the same pattern Muse Image now extends: broad use of people’s data unless they actively turn it off. Meta has said its product has “built-in protections” against violent, sexual, or defamatory AI imagery, and that users have control over how their content is reused. The model is a faster, more capable version of the kind of tool the company has shipped before, the same default-on, opt-out design Meta has applied to the Model Capability Initiative employee tracking program.

How the Opt-Out Actually Works

Meta frames the feature as user-controlled. The setting, when found, does shut it off.

1. In the Instagram smartphone app, tap your profile, then the hamburger button in the top right corner.
2. Find the tab labeled “Sharing and reuse.”
3. Toggle “Allow people to use your content on Instagram and with AI features at Meta” off for both Posts and Reels.
4. On a desktop browser, the fastest path is to set the account to private, per Gizmodo’s test.

The desktop browser path is the most direct: set the account to private, and the photos are no longer reachable for Muse generations. In the smartphone app, per Wired’s reporting cited by Gizmodo, the toggle lives under Profile, hamburger menu, “Sharing and reuse,” with separate controls for Posts and Reels. The setting is not surfaced in the same screen as the more familiar privacy controls. The default remains on for any public account that has not gone through the menu.

The opt-out governs whether a user’s own photos can be reused as raw material for AI generations, not whether they can be tagged in images other people generate. The model is also drawing on Instagram for “social context” beyond a single tagged handle, Meta’s launch blog said, including using posted images as visual references when a user enters a public @ username in a prompt. Public accounts that have not opted out are the surface the feature runs on.

The Other Use Cases Meta Is Pitching

Meta is selling Muse Image to two customer bases at once. For consumers, the company is highlighting a set of light, creative tools: erasing a photobomber, dropping a portrait in front of a historical landmark, generating a working QR code from a prompt, and redecorating a room with Facebook Marketplace finds. The Marketplace integration was the centerpiece of a launch demo, in which a user applies Muse to see what a secondhand couch might look like in their garage. Use of the model is free for “everyday creation”, Meta said.

For advertisers, Muse Image will power image-generation tools inside the company’s AI-driven Advantage Plus service. The pitch is “high-quality, on-brand ad variations with fewer iterations,” per Meta’s advertiser blog, and brands will start to see Muse-powered variants in their campaigns in the coming weeks.

What Muse Image Signals for Meta’s AI Business

Muse Image is the second major model from Meta Superintelligence Labs this year. The launch positions Meta behind OpenAI’s image generation but ahead of Google’s Nano Banana 2 in Meta’s own benchmarks. Meta has also said it plans to use Muse Image to reduce reliance on third-party image and video models like Midjourney and Black Forest Labs.

Wang, who also oversaw the April unveiling of the Muse Spark large language model, runs the unit behind the launch. The launch comes as Meta has been criticized for having a “nebulous AI strategy,” even as the company has committed to large AI infrastructure spend in 2026. The model is one of the products the company is leaning on to convert that spend into subscription and ad revenue, and the U.S. government is pressing Meta to open its AI models to safety review. The privacy debate around Muse Image will run in parallel as the model rolls out.

Muse Video, an AI video generator, is “already in development” per the same launch announcement. Regulators have not yet commented on Muse Image specifically.

Frequently Asked Questions

What is Meta Muse Image?

Muse Image is the first in-house AI image generation model built by Meta Superintelligence Labs (MSL), the AI unit run by Wang. The model is free through the Meta AI app, meta.ai, Instagram Stories in the US, and WhatsApp in limited countries. It is the second major model MSL has released, after the April debut of the Muse Spark large language model.

Who can use your Instagram photos in Muse Image AI images?

A user can pull any public Instagram account into an AI image just by typing the account’s @ handle into a prompt inside the Meta AI app. The setting is on by default, and neither the account holder nor the people in the photos are notified when their images are used.

How do you opt out of Meta Muse Image using your Instagram photos?

The fastest path is to set the Instagram account to private on a desktop browser, per Gizmodo’s test. In the smartphone app, users can go to Profile, hamburger menu, “Sharing and reuse,” and toggle off both Posts and Reels. The setting is not surfaced in the same menu as the more familiar privacy controls.

When did Meta launch Muse Image?

Meta launched Muse Image on Tuesday, July 7, 2026, with the model rolling out across the Meta AI app, meta.ai, Instagram Stories in the US, and WhatsApp in limited countries, plus the video counterpart Muse Video in development. Facebook and Messenger are set to follow later this year.

What is Meta Superintelligence Labs?

Meta Superintelligence Labs (MSL) is Meta’s dedicated AI division, led by Wang. Muse Image is MSL’s second major release, after the April launch of the Muse Spark large language model. The unit sits at the center of Meta’s effort to build its own foundation models in-house.