Host: Your face isn't just a style, and your voice isn't just some preset on a computer—but in the world of generative AI, that’s exactly how the technology sees them. To an AI model, your face is just a distribution of data and your voice is a waveform that anyone can 'wear' with the right prompt.

Listener: That’s actually pretty terrifying. It feels like we’ve gone from someone just stealing a photo to someone being able to completely inhabit your identity. Is there actually a way to stop that, or are we just stuck with 'report' buttons that don't do much?

Host: Exactly. Most platforms just give you a report button and hope for the best, which is basically an apology process after the damage is done. But what we’re looking at today is a move toward actual infrastructure for your likeness—treating your face and voice like property that needs a real system to protect it.

Listener: So how do you even start 'owning' your own face in a machine-readable way? Do I have to register myself like a trademark?

Listener: Okay, so I’ve registered. But the internet is huge. I can't spend my whole day hunting down deepfakes of myself. Does the system actually look for me?

Host: That’s the second layer: detection. The document argues that the platform should be on the hook for looking, not the victim. The system monitors for unauthorized avatars or voices across the platform. It treats impersonation like a background rate—checking new accounts and uploads constantly so you don't have to stumble onto your own double by accident.

Listener: And if it finds a match? Is it just an instant, silent ban? Because I can see that getting messy if someone just happens to look a bit like a celebrity or something.

Host: That’s a great point. The approach here is 'enforcement with warning.' Instead of silent takedowns where content just vanishes, the system issues a notice. If it's a plausible mistake, there's a grace period to comply or contest it. But if it's clear bad-faith impersonation, that courtesy goes away. The key is that the authority stays with you—it’s your likeness, so it’s your call, not some faceless bureaucracy's.

Listener: It sounds like a lot of walls. Which I get, protection is good, but does this make it impossible for people to use these tools for legitimate stuff? Like, what if I *want* to be in a commercial but I don't want to fly to a studio?

Host: That is actually the door this whole system is trying to open: licensing. The idea is that nobody licenses what they can freely steal. By making your likeness scarce and protected, it suddenly has market value. A brand could pay to use your AI avatar as a spokesperson, or a community could license your character for an event. It turns potential theft into a deal where you set the terms and the revenue flows to you.

Listener: That makes sense. But this only works if every site uses it, right? If I protect my face on one site, what stops someone from using it on another?

Host: That’s the most interesting part of the report. There’s already a live IP Protection API that other platforms can use. Before a developer mints an AI character, they can call this API to check if a face or voice is protected. It tells them right then and there if a license is required. It's meant to work across property lines because impersonation doesn't stop at the edge of one website.

Listener: So we’re moving from 'hide and hope' to a world where our identity is actually machine-readable and defensible.

Host: Exactly. We can't go back to the old days of obscurity. The way forward is making ownership a part of the infrastructure. If you're a creator, you might want to look into the myOrbit creators page to see how those verified claims work—and if you’re a developer building these tools, the IP Protection API is already live and waiting for you to check those keys.