Instagram Pulls Its Controversial AI Feature: What It Means for Brands

What if your next photo posted on Instagram was used to generate someone else's image, without you knowing? That's exactly what happened on July 7, 2026. Meta launched Muse Image, an AI image generator that let any user tag a public Instagram account in a prompt to recreate its photos. The option was turned on by default, with no notification sent to the accounts involved (MacRumors, 2026). Three days later, under pressure from users and two professional organizations from the creative industry, Meta pulled the feature. Here's what this episode actually changes for brands and agencies that post content on Instagram.

What happened: Muse Image, the AI feature that tagged your public accounts

On July 7, 2026, Meta launched Muse Image, its first image-generation model built by Meta Superintelligence Labs, rolled out across Meta AI, Instagram, and WhatsApp (Meta Newsroom, 2026). A built-in option let any user tag, with a simple mention, a public Instagram account in a prompt to generate an image based on its photos or Reels.

The problem: this option was turned on by default for every public account, with no opt-in and no notification to the owner when their content was used (MacRumors, 2026). Anyone could therefore generate an image reusing the visual identity of a brand, a creator, or a business, with no approval requested beforehand.

By July 9, trade press was already publishing guides on how to turn off the option, proof that the general public was discovering the setting at the same time the controversy broke (TechCrunch, 2026). The setting existed, but no one had been invited to use it before launch.

Why did Meta pull the feature after just three days?

Meta pulled the feature on July 10, 2026, just three days after launch, under combined pressure from users, the SAG-AFTRA actors' union, and talent agency CAA (TechCrunch, 2026). The main trigger: the lack of explicit consent from the people whose content was being reused.

Meta justified the withdrawal in blunt terms: "Our intention was to provide a useful creative tool and let people control whether their public content was reused in this way. We heard that this feature missed the mark, so it's no longer available" (Social Media Today, 2026).

Only the ability to tag third-party public accounts disappeared. The Muse Image model is still active for generating images from your own photos or classic text prompts. It wasn't the AI tool itself that caused the problem, but the lack of consent in how it was applied by default to other people's content.

What does this episode reveal about how AI uses your public content?

Muse Image isn't an isolated case. Since May 27, 2025, Meta has been training its AI models, including Meta AI and Llama, on European users' public posts, under a "legitimate interest" legal basis that works on an opt-out rather than opt-in basis (The Hacker News, 2025). The noyb association, founded by Max Schrems, has filed complaints against Meta in 11 European countries over this same mechanism.

This opt-out setup, where public content feeds training or generation by default unless the owner explicitly opts out, structures much of Meta's AI ecosystem. It's exactly the pattern that failed with Muse Image, applied this time to a case that's more visible and easier for the general public to grasp: directly reusing your face or your product in a generated image.

User trust is taking a hit as a result. 69% of Americans know social media platforms use their data to train AI, 91% say they're worried about it, and 34% have already stopped using a platform for that reason (ExpressVPN, 2025). For a brand, every public post is potentially raw material for AI systems you don't control, well beyond the Muse Image case alone.

How can you protect your brand's content against Instagram's AI tools?

Three concrete settings limit your brand content's exposure. None of them is perfect, but combined, they meaningfully reduce the risk of your content being reused without consent by Meta's AI tools.

First, turn off sharing and AI reuse from your profile: Menu, Sharing and reuse, then turn off the "Posts" and "Reels" toggles under the option allowing reuse by Meta's AI features (Malwarebytes, 2026). Important limitation: this only affects future generations, content already used before you turn it off stays in circulation.

Next, for users and businesses based in the European Union, an official objection form lets you opt out of Meta's AI training on your public content, a recourse strengthened by GDPR that users outside the EU don't have under the same conditions.

Finally, for high-value brand visuals, digital watermarking that hides a signature within an image's pixels is gaining ground in 2026 as a complementary protection measure, on top of the platform's native settings. This practice doesn't replace Instagram's settings, it supplements them for your most sensitive assets.

What agencies and brands should watch in the coming months

Marketing teams' adoption of generative AI keeps growing despite the controversy: 87% now use it in at least one recurring workflow, up from around 51% two years ago (Salesforce, 2026). The gap between professional adoption and public distrust keeps widening, and that's precisely the gap Muse Image exposed.

On the audience side, the signal is clear. 50% of Gen Z has already blocked, muted, or unfollowed an account because its content looked like "AI slop," and 88% of users say AI video tools have eroded their trust in the information they see on social media (Sprout Social, 2026).

For an agency managing several client accounts, the focus falls on two fronts: never publish an AI-generated visual without clearly labeling it, and document the privacy settings applied to each client account. An incident like Muse Image, on a poorly configured account, quickly becomes something you have to explain in a status meeting.

More episodes like this are likely. Platforms regularly test AI features in live conditions, with a consent framework that's still unstable. Staying informed about these setting changes is now part of the everyday work of a social media agency, on par with algorithm monitoring.

Frequently asked questions about the Muse Image withdrawal

Does Muse Image still exist today?

Yes, partly. Only the ability to tag a third-party public Instagram account in a prompt was removed on July 10, 2026. The model is still available for generating images from your own photos or classic text prompts.

How can I find out if my content has already been used by Muse Image?

There's no way to check. Meta doesn't notify accounts whose content has been tagged, and no public history is accessible. That lack of traceability is exactly what fueled the controversy.

Does turning off sharing and reuse protect my content retroactively?

No. This setting only blocks future reuse. Any image already generated before you turned it off remains in circulation, with no way to have it removed after the fact.

Should brands stop using generative AI after this episode?

No, but they should be selective. 87% of marketers already use generative AI in a recurring workflow (Salesforce, 2026). The risk doesn't come from AI itself, but from the murky consent mechanisms applied to third-party content.

Does GDPR protect European businesses against this kind of use?

Partially. Businesses based in the European Union have a formal right to object to Meta's AI training on their public content, a recourse that users outside the EU don't have under the same conditions.

Muse Image was only online for three days, but the episode's implications go well beyond that short window. It exposes a broader mechanism: the public content you post on Instagram can, by default, feed AI systems without you being informed. Check your sharing and reuse settings, document them for your client accounts if you run an agency, and keep an eye on the next AI features Meta tests. The next Muse Image won't give you a heads-up before launch either.

How Propulse helps you put this into practice

Being wary of AI tools that reuse content without consent only makes sense if your own Instagram growth rests on healthy foundations. An account that grows through fake followers or artificially generated content faces the same trust risks as those exposed by Muse Image.

Propulse builds Instagram growth on real accounts, with no generative AI and no fake followers:

• Organic growth automation, with interactions coming from real accounts
• Connection via residential 4G IP to simulate human browsing and keep the account secure
• No artificial content generation, no reuse of third-party content without consent
• Precise targeting by industry, without relying on unstable AI algorithms
• Transparent performance reports to track your audience's real growth

Our clients gain 3x qualified followers on average within 15 days (see our case studies), never relying on murky consent mechanisms.

Discover our Instagram growth services. Then check out our pricing to get started today.