Meta's latest AI experiment had a short life. A new feature that let people generate AI images based on public Instagram posts launched on Tuesday and was shut down by Friday, following a wave of user backlash.
The feature arrived alongside Muse Image, the first image-generation model built by Meta Superintelligence Labs, which is accessible through Meta AI. While Muse Image itself remains available, including tools for editing photos directly, the specific capability that let the model reference public Instagram content was pulled.
In a statement on Instagram's blog, Meta acknowledged the feature "missed the mark" and said its goal had been to build a useful creative tool while giving users control over whether their public posts could be used this way.
Swift Backlash
Users reacted quickly once the tool rolled out, raising concerns about privacy and the potential for deepfake content. Guidance on how to opt out began circulating widely on platforms like X and Reddit within days.
Privacy advocates weighed in as well. Apar Gupta, founding director of the Internet Freedom Foundation, argued that Meta's dominance in social media doesn't give it license to repeatedly test the boundaries of user consent.
Entertainment Industry Pushes Back
The backlash extended well beyond individual users. SAG-AFTRA, the union representing roughly 160,000 entertainment and media professionals, had urged members to opt out of the feature and welcomed its removal, calling anything short of clear opt-in consent for this kind of use "unacceptable."
Creative Artists Agency also spoke out, stressing that no one's name, image, likeness, voice, or creative work should be used by AI models or any third party without documented consent.
The episode reflects a broader tension between AI developers and the entertainment industry over likeness rights. Some celebrities, including Matthew McConaughey and Jeremy Clarkson, have taken the step of trademarking aspects of their public identity to guard against unauthorized AI use.
Not the First AI Misstep
Meta isn't alone in facing this kind of reaction. OpenAI encountered similar criticism after launching its Sora 2 video-generation platform in 2025, when it initially allowed users to create content featuring trademarked characters — drawing objections from studios like Studio Ghibli. OpenAI later struck a licensing partnership with Disney, but ultimately discontinued the Sora 2 app in March.
Source: Business Insider