AI likeness protection starts with ownership, consent, and proof.

A likeness is a licensable asset, not a texture to borrow.

Generative models can reproduce a creator's face, voice, and style with enough fidelity to pass as authentic. That makes identity a commercial surface, not just a personal characteristic. Every time a model renders a recognizable likeness without the creator's knowledge, a licensing event has occurred with no record, no consent, and no payment route attached to it. Detection tools and takedown workflows address that harm after the fact. They identify a violation, not an authorization. They can remove content from one platform and leave it on a dozen others. They give creators a grievance and a process, but they do not give creators a record that was established before the use happened, and they pay nothing.

The registry makes consent the starting point, not the remedy.

Suede treats likeness as registered identity: a creator's face, voice, and persona are entered into the registry with consent parameters and license terms attached before any use can occur. When an authorized party generates content using that likeness, the event references the registry entry. The provenance record is created at the point of generation, not reverse-engineered from a takedown claim. That record carries the creator's terms, routes payment to the creator's designated address, and gives both parties a verifiable chain of authorization. The contrast with detection-only approaches is not marginal. Detection arrives after the harm. The registry prevents the unauthorized event from being untracked in the first place, and it ensures that every authorized use is payable rather than just permitted.

Related ownership concepts

• AI Voice Protection
• Content Provenance
• Talk to Suede