Meta disables image generation from public Instagram profiles in Muse Image — privacy turned out to be more important
Meta has abruptly withdrawn one of the key features of its new image generation model, Muse Image. The feature in question allowed users to create images by referencing public Instagram accounts in the prompt. This option, introduced just three days ago, has already been deemed not meeting user expectations in terms of privacy.
Promise and Failure
On July 7, the company announced Muse Image — the first image generation model developed by Meta Superintelligence Labs and integrated into Meta AI. Among the stated capabilities was a feature that allowed adding the name of a public Instagram profile to a text prompt. The service was supposed to use public photos from that account to generate a new image. However, by July 10, Meta published an update to the announcement, stating that this feature had been immediately disabled.
The reason given was "not meeting user expectations" regarding privacy. In practice, this meant the feature was enabled by default, sparking a wave of criticism. Actress Hannah Einbinder publicly urged users to disable this option, and the SAG-AFTRA union (Screen Actors Guild-American Federation of Television and Radio Artists) issued a strong statement.
Community and Union Reaction
SAG-AFTRA called this approach a "complete miscalculation of public opinion." The official statement emphasized: "Anything other than explicit and conspicuous consent for such uses of Instagram users' images is unacceptable and is a complete miscalculation of public opinion regarding the obvious dangers and harms inherent in such use." After the feature was disabled, the union supported Meta's decision, acknowledging it as the right move.
This incident demonstrates how fine the line is between innovation and ethics in the era of generative AI. Meta appears to have underestimated the level of user concern about the use of their public data for training and operating models.
What Remains?
Despite the withdrawal of the controversial feature, the Muse Image model itself continues to operate as part of Meta AI. Users can still generate and edit images based on text prompts and uploaded photos. Additionally, the announcement mentions that Muse Image is already being used in over 30 new AI effects for Instagram Stories, as well as in WhatsApp chats with Meta AI in a limited number of countries.
Recall that earlier, on June 9, Meta also introduced the multimodal model Muse Spark 1.1, focused on agent tasks and coding.
Analyst's Opinion: This case is not just a technical glitch, but an important signal for the entire industry. Users no longer want to be "raw material" for AI models without their explicit consent. Meta, realizing its mistake, acted quickly, but reputational damage has already been done. Companies will have to reconsider their approaches to collecting and using public data, otherwise regulators and the public will do it for them.