Meta allowed Instagram photos to be used for AI generation without explicit user consent

Meta in July 2026 launched the Muse Image image generation model and by default opened access to public Instagram photos for creating AI content — users must manually disable this option through privacy settings, otherwise their photos remain accessible to anyone.

How the new policy works

With the launch of Muse Image, Meta changed the rules for using public content on Instagram. Photos from open accounts can now be used to generate AI images — not only by the author themselves, but by any other user on the platform.

• Muse Image — Meta's proprietary generative model for image creation
• All Instagram accounts with public privacy settings are affected
• Consent scheme: opt-out — the feature is enabled by default
• To opt out, users must manually change their privacy settings

Users who don't keep up with policy updates automatically allow their photos to be used — without even realizing it.

Why opt-out is more dangerous than opt-in

At first glance — a technical detail. In practice, it's a choice that determines the fate of data for tens of millions of people.

With opt-in, a user must explicitly grant permission before their content is used. With opt-out, it's enough not to respond to the policy change — and the photos are already available to the system. Research on user behavior shows that the vast majority of people don't read agreement updates and don't change default settings. As a result, Meta gets a massive visual dataset based not on the informed consent of its audience, but on their passive silence.

Professional photographers and content creators are especially vulnerable: their public portfolios risk becoming training material for AI systems that potentially compete with them in the visual content market. Publishing work in public to attract clients is fundamentally different from agreeing to use it in generative models.

Context: Meta and the fight for data

Muse Image fits into Meta's broader strategy in generative AI. The company is developing open models of the Llama family, embedding AI assistants in WhatsApp, Facebook, and Instagram. Access to a multi-billion library of public Instagram photos — an obvious competitive advantage when creating your own image model.

In Europe, this approach is systematically challenged using GDPR, which requires informed and explicit consent for personal data processing. In the US, the legal framework is softer, but pressure from regulators and human rights organizations is increasing with each such precedent.

What it means

Meta turns Instagram into a de facto dataset for Muse Image, betting on the passive silence of millions of users with public accounts. Those who don't want to participate need to find and disable the appropriate setting themselves — the company doesn't send active notifications about this change.

*Meta is recognized as an extremist organization and banned in Russia.