Meta Launches Muse Image, Generating Photos of Other People's Instagram Accounts Without Consent

Meta launched Muse Image on July 7, its first in-house image generation model built by Meta Superintelligence Labs. The tool lets any Meta AI app user tag a public Instagram account's username in a prompt and receive a generated image of that person, without their knowledge and without separate consent to use their photos.

The mechanism is simple. Just type a prompt in the Meta AI app with an @ symbol and the username of a public Instagram account, and the system pulls that person's photos as a visual reference to generate a new image based on them. Meta's promotional materials give examples like placing yourself in front of a landmark or removing a stray person from a shot, but the mechanism works identically on other people's accounts, as long as they're public.

Opted in by default

The core problem lies in the default settings. Every public Instagram account is automatically included in the pool of sources Muse Image can draw from. To opt out, you have to go into your profile, open the three-line menu, find the section on sharing and reuse of content, and then separately disable posts and stories under the option that allows use of content on Instagram and in Meta's AI features.

Meta's own terms explicitly state that a person whose photos were used for generation will not receive any notification of that fact. There's also no prior consent involved, the feature works on an opt-out basis rather than opt-in. That reverses the standard order, in which the user decides in advance whether their likeness can be processed by AI systems.

You won't be notified about content created using Meta AI features - from Meta AI's terms of service

Another problem is the lack of retroactive effect. Turning off consent in settings only blocks future generations, images already created from someone's likeness remain in the system and can keep circulating regardless of a decision made later by the person photographed.

End of the Midjourney partnership

The launch of Muse Image formally ends Meta's years-long partnership with Midjourney, whose license powered the company's earlier image tools. Meta Superintelligence Labs, the unit responsible for developing the model, was created as part of a broader reorganization of the company's AI research, and Muse Image is its first standalone product to reach billions of users at once.

The company plans to open the tool to advertisers and marketing agencies in the coming weeks, meaning users' likenesses could start appearing not just in private generations made by friends but also in commercial material. In parallel, work is underway on Muse Video, a version of the tool for generating video clips, expected to launch in the coming months.

Market pressure and criticism

For Meta, Muse Image is also a response to pressure from investors who expect concrete proof that massive spending on AI infrastructure is translating into real products and revenue. Analysts note that the sheer scale of investment has stopped being enough to satisfy the market on its own.

Meta needs to deliver more proof of both adoption and monetization - Ralph Schackart, analyst at William Blair

Privacy advocates warn that pulling real users into generated images without explicit consent is a legal and reputational problem that could explode at any moment, especially if someone uses another person's likeness in a compromising or false context.

What it means for Europe and Poland

In the European Union, the opt-out model will run up against GDPR (the EU's General Data Protection Regulation), which requires explicit, informed consent to process someone's likeness rather than silent acceptance buried in settings. If Meta wants to launch Muse Image in its current form in Poland and other EU countries, it could become the subject of intervention by national data protection authorities, similar to earlier disputes over training AI models on European users' data.

For Polish companies and creators who use Instagram commercially, the problem is already practical today. A brand's or influencer's public account automatically becomes raw material for images generated by third parties, which could lead to likeness violations, fake advertisements, or manipulated content attributed to recognizable faces without their knowledge.

Sources: TechCrunch (techcrunch.com), CNBC (cnbc.com), Digital Trends (digitaltrends.com), PYMNTS (pymnts.com)