Meta has rolled back a new Instagram AI feature just days after introducing it in the US, following widespread criticism over privacy and consent. The feature, part of the company’s new Muse Image AI model, allowed users to generate customised AI images by referencing public Instagram accounts, inviting concerns that people’s photos could be used without their permission.
The company acknowledged the criticism, saying it had “missed the mark” and confirmed that the feature is “no longer available.” Meta has not said when or if the feature will return.
Why was Meta’s Muse Image AI feature criticised?
Muse Image is Meta’s first dedicated AI image-generation model. Like competing tools from Google, OpenAI and Adobe, it can generate images from text prompts, edit photos and create new visuals. Its distinguishing feature was its deep integration with Instagram, which could’ve been a boon for the company but proved to be a bane.
Users could simply mention a public Instagram account while interacting with Meta AI, allowing the chatbot to pick up publicly available photos from that profile to generate customised AI images featuring that person. The feature was enabled by default for public accounts in the US, meaning users had to manually opt out if they did not want their content to be used.
Privacy advocates and creators argued that this approach lacked consent. Hollywood performers’ union SAG-AFTRA welcomed Meta’s decision to withdraw the feature, calling it a “win”, while Privacy International criticised the rollout as another example of AI companies treating users’ images as data for AI training and generation.
The AI image feature was available only in the US, but Meta had indicated that it planned to expand it to additional markets over time. Users in supported regions could disable content reuse by going to Instagram Settings and turning off “Allow people to reuse your content on Instagram and with AI features at Meta.”
However, there’s a trade-off. The same setting also controls existing collaboration features such as Remixes, templates, stickers and certain content-sharing options. Turning it off prevents Meta AI from using public content for AI image generation, but it may also reduce opportunities for creators to increase the reach of their posts through Instagram’s collaborative features.
Companies are racing to introduce image-generation capabilities that are tightly integrated into existing apps, but user expectations around privacy and consent are rising just as quickly. Rivals including Google, OpenAI and Adobe have generally required users to actively upload or provide images they want AI to edit, making Meta’s default-enabled approach particularly controversial.
For users, this is a reminder to review privacy settings whenever new AI features are introduced. Those who are uncomfortable with public content being repurposed for AI-generated images may prefer to opt out. Meta’s swift retreat underscores a critical lesson: in the age of AI, user trust and explicit consent are non-negotiable, even for public data.
