I woke to a headline that felt like a live wire: a deepfake of a public figure had already seeded doubt across feeds. I checked my face on five sites and realized I had no paperwork saying who could use it. You should care, because if you don’t set the terms, someone else will write them for you.
I’m telling you this because Cate Blanchett — the actor who plays other people for a living — just pushed a practical tool into the center of a messy public problem. On Tuesday Blanchett announced the Human Consent Registry, a free service from her non‑profit, RSL Media, that lets people record how their name, image, voice and other signature traits may be used by AI. “Your identity is your IP in the age of AI,” Blanchett said, handing ordinary people a way to state limits on their likeness.
At film festivals and Twitter threads people notice deepfakes within minutes — How the registry works
You create an account, give a few identifiers (name, profession, links to your website or social profiles), and select an AI‑use consent level. The choices are set up like a stoplight: Prohibited (red), Permitted with Terms (yellow), or Permitted (green). When you finish, you get a Human Consent ID that AI developers and platforms can check before using any part of your likeness in training data.
How does the Human Consent Registry work?
I tested the flow so you don’t have to: it asks for basic public identifiers and the scope of permission you want to grant. RSL Media says the registry covers name, image, likeness, voice, movement and other signature attributes. The system ties those choices to an open protocol, Really Simple Licensing (RSL), which has been adopted by publishers and platforms as a standard for AI usage rights.
When a viral clip erases context people demand answers — Why this matters
You already see the problem: celebrities, journalists and ordinary people have had AI‑generated likenesses surface without consent. The registry gives everyone a public record to declare permissions and lets developers check that record before scraping or training models on identifiable data.
That record matters because it creates friction for bad actors and a clear signal for responsible platforms — think of it as a public ledger that a company like OpenAI, Meta or Google could reference. RSL has been gaining traction among publishers; now RSL Media wants that same approach applied to human likeness.
Can AI companies ignore the Human Consent Registry?
Short answer: yes—today. There’s no universal enforcement mechanism baked into the registry. RSL Media is publishing the tool and the protocol; compliance depends on platforms and model builders adopting the check. Still, a public registry raises the reputational and legal cost of ignoring consent, and may accelerate policy or contractual obligations across the industry.
On social feeds people trade screenshots like proof — Limits and next steps
You should know the tradeoffs: you’re putting personal data in a third‑party service, and the registry doesn’t itself stop scraping. RSL Media acknowledges that and plans follow‑ups: registries tailored for Work, Characters and Marks that map related rights into the same protocol.
RSL’s approach builds on an open standard—Really Simple Licensing—that publishers have used to label allowed uses of content. TechCrunch and other outlets have covered RSL’s growth among digital publishers, and this is the first major push to treat likeness the same way as text, images and code.
How do I register my likeness?
Go to the Human Consent Registry, fill in the profile fields, attach the links that prove it’s you, pick a consent level, and save. The service is free. If you want a theatrical example: Verify yourself to Lydia Tár.
I’ll be watching whether major platforms adopt the Human Consent ID check or whether lawmakers decide to harden the promise with legal teeth. For now, this is a practical tool with social leverage — not a legal shield. Will you set the terms for your likeness before someone else does?