What happens to a request that breaches it?

It is refused before any model is called, with the rule name and the matched term quoted back. The refusal is recorded in the creator's audit log. It is never rewritten into something permitted.

Does the API enforce this too?

Yes — the licensed-likeness API and the MCP server call the same evaluation the website calls. There is one implementation of the consent gate, so an automated integration cannot get a weaker check than a browser.

Boundary type

Criminal & extremist context

Depictions placing the likeness in the commission of a crime, in terrorist or extremist iconography, or in hate propaganda.

Why this line exists

Placing a real person in the commission of a crime, or in extremist iconography, is defamatory by construction and unlawful in most markets this operates in. It is not a creator-level choice.

How a request is matched to it

Terms naming criminal acts, extremist symbols and hate propaganda.

Matching is deterministic and anchored to word boundaries, so a term inside a longer innocent word does not fire. The matched term is returned in the refusal, which means a creator can see why their rule caught something and a requester can see what to change.

Whether creators permit it

Never. This category cannot be permitted by any boundary set.

Questions

Can a creator permit criminal & extremist context?: Never. This category cannot be permitted by any boundary set.
What happens to a request that breaches it?: It is refused before any model is called, with the rule name and the matched term quoted back. The refusal is recorded in the creator's audit log. It is never rewritten into something permitted.
Does the API enforce this too?: Yes — the licensed-likeness API and the MCP server call the same evaluation the website calls. There is one implementation of the consent gate, so an automated integration cannot get a weaker check than a browser.