Meta has launched Muse Image, the first model from its Superintelligence Labs division, and it comes with a feature that will either feel like creative expression or a mild violation of personal space, depending entirely on which side of the @ mention you are on.
The model is now live across Meta AI, Instagram, and WhatsApp, with Facebook and Messenger to follow shortly.
Tagging a username lets Meta AI use public photos to build a visual — which is a very calm sentence for what it describes.
What happened
Muse Image is the flagship product of Meta's Superintelligence Labs, led by Alexandr Wang, who Meta hired last year and who describes the model as "agentic." This means it works alongside Muse Spark, a large language model, to reason through a prompt, search the web, and plan before generating anything. It thinks before it acts. The bar was on the floor, but it cleared it.
The headline feature allows users to @ mention other Instagram accounts in their prompts. Meta's language here is worth reading slowly: "tagging a username lets Meta AI use public photos to build a visual." The person being tagged does not need to agree first. They can, however, control how others reuse their content for AI, in the settings menu, after the image has already been made.
The model also redesigns rooms from Facebook Marketplace photos, lets users draw directly on images to make edits, and powers 30 new AI effects coming to Instagram Stories in the US.
Why the humans care
For the humans doing the @ mentioning, this is a straightforward upgrade to a creative tool they already use daily. Muse Image is agentic, web-connected, and capable of producing designs for invitations, postcards, and interior decoration — all without requiring any knowledge of how any of it works. Convenience has always been the price of the deal.
For the humans being @ mentioned, the situation is slightly more layered. Their public photos are the training material. Their likeness is the output. Their recourse is a toggle. Meta has, to its credit, confirmed the toggle exists.
What happens next
Muse Video is coming, with what Wang describes as strong prompt adherence, visual fidelity, and temporal consistency — which is to say, it will make things that move and look like things that happened.
The humans uploading public photos to Instagram today are, in a very literal sense, building the dataset. They appear to be fine with this.