Meta disables the feature of generating images from public Instagram photos — privacy turned out to be more important than AI ambitions
Meta has taken an unprecedented step by disabling the ability of its new AI model, Muse Image, to generate images based on mentions of public Instagram accounts. The company acknowledged that the feature "did not meet user expectations" regarding privacy and removed it from the official announcement just three days after launch.
On July 7, Meta introduced Muse Image as a flagship image generation model from Meta Superintelligence Labs, integrated into Meta AI. The key "feature" was the option to add a public Instagram profile to the prompt — the neural network automatically used public photos to create new content. However, by July 10, the company updated the documentation, announcing the immediate disabling of this function.
Community and Union Reaction
Meta's decision appears to be a forced concession under public pressure. Actress Hannah Einbinder publicly urged users to disable the option, and the SAG-AFTRA union issued a strong statement. The organization called "unacceptable" any mechanism that uses images without explicit and conspicuous consent, emphasizing that such actions represent a "complete miscalculation of public opinion regarding the obvious dangers and harms."
The key issue was the automatic opt-in feature by default. This created direct privacy risks even for users whose accounts were public. Meta effectively admitted that it failed to distinguish between "public access" and "consent for use in AI generation."
What Remains?
Despite the incident, the Muse Image model itself has not disappeared. It continues to operate in Meta AI for standard tasks — generating and editing images based on text prompts and uploaded photos. Additionally, Meta is already using Muse Image in over 30 new AI effects for Instagram Stories and in WhatsApp chats with Meta AI in a limited number of markets.
This case is a vivid example of how quickly AI giants can roll back features when faced with real resistance. For me, this is a signal: in the race for multimodality, companies like Meta are beginning to realize that even "public data" requires a much more nuanced approach to ethics than just the technical ability to process it. I expect that in the coming months, we will see a tightening of consent policies for content use on social networks.