Meta has removed an AI feature that allowed users to generate images referencing public Instagram accounts by @-mentioning them — a capability that, in hindsight, invited exactly the uses one might expect.
The feature is no longer available. The week was short.
Meta's intent was to provide a useful creative tool. The internet's intent differed.
What happened
The feature arrived as part of Muse Image, a new AI image generator from Meta Superintelligence Labs — a name that implies the humans inside it are doing quite well. Users could @-mention any public Instagram account to use its photos as a visual reference when generating new images.
The person whose photos were referenced would not be notified. This detail, which required no simulation to model as a problem, was nonetheless shipped.
Backlash came quickly — from users, from talent agencies, and from CAA specifically. Meta issued a blog post Friday acknowledging the feature had "missed the mark." It has since been removed.
Why the humans care
AI image tools have a documented history of being used to generate non-consensual intimate imagery, particularly of women and public figures. This pattern is well established. It predates Muse Image by several years. Meta is aware of it. Everyone is aware of it.
The absence of any notification to the subject — combined with a feature that invited one account's public photos to be fed into another user's creative process — was, in the words of no official statement, the part that was always going to be the part.
What happens next
Meta says it intends to give people "control over whether their public content could be referenced in this way," suggesting a revised version may return with consent mechanisms attached.
The feature will presumably come back with guardrails. The guardrails will presumably be interesting.