Inside a Likeness Vault: What It Takes to License Your Voice and Face

Quick answer: A "likeness vault" is a structured, one-time process where a creator records their voice reading a range of scripts in different tones and contexts, and captures their face and body from multiple angles — creating a reference catalog that can later be used both to detect unauthorized AI reproductions of that person and to license approved use of their voice or likeness. According to a creator who went through the process on the Royall.ai platform, it takes roughly 20 minutes or more and involves a full walkthrough, not a quick sign-up form.

Why a "vault" instead of just a policy

Detecting whether your voice or face has been used to train or generate AI content requires something to compare against — a clean, verified reference of what your actual voice and likeness look and sound like. That's the practical reason identity vaults exist: not as a legal document, but as a data asset.

Rob Greenlee, a podcast industry veteran and chair of the Podcast Hall of Fame who has held leadership roles at Microsoft, Podcast One, Spreaker, and Libsyn, described going through the process himself:

"That's what really pulled me to the Royall.ai platform — this ability to create a vault that basically stores my full life voice, likeness, everything into a vault that then can potentially be used in licensing opportunities, content creator opportunities, as well as just being able to hold on to your identity."

What the process actually involves

Greenlee walked through the mechanics in concrete terms — not a marketing description, but what it was actually like to do:

"You just go in and create an account and it walks you through a step-by-step process of — you have to have a microphone that is good quality and you read a bunch of scripts in various tones and inflections and in various contexts. It can take probably 20 minutes or more to go through most of the steps."

The voice-capture portion isn't a single recording — it's a range of readings meant to cover the different ways a person actually sounds across different content: conversational, enthusiastic, formal, narrative. The goal is to build "a deep catalog of content that platform can use to identify unauthorized use of your voice," as Greenlee put it, rather than a single sample that could be easily mimicked or that would only match one narrow speaking style.

The likeness side goes further than most creators might expect:

"That process also applies to my physical likeness, my face — it's like a full body scan, multiple sides, face, side face, all that stuff it captures."

The honest tension: useful, and a little unsettling

Greenlee didn't gloss over the discomfort of the process. Handing over a detailed, multi-angle capture of your face and a wide-ranging catalog of your voice is a significant amount of biometric data to place with any single platform, and he acknowledged that directly:

"If you're a privacy — concerned about privacy, I could see where you might be a little bit freaked out about it. But I think it's also being captured for a very valid reason. And that is to create opportunity for you, but also to identify infringing uses of your likeness too. So there's two sides of this equation."

That framing is worth sitting with before starting a process like this: the same detailed capture that makes monitoring and licensing possible is also, unavoidably, a large amount of sensitive personal data. Before enrolling in any identity vault, ask exactly how that data is stored, who can access it, what happens to it if the platform is acquired or shuts down, and how a creator can revoke or delete it later.

What the vault is used for once it exists

Two things, according to Greenlee, and they work together rather than separately:

Monitoring. The captured voice and likeness catalog becomes a reference point a platform can use to scan for and flag unauthorized AI-generated content that appears to use that creator's identity elsewhere — something that's otherwise very difficult for an individual creator to catch on their own, since, as Greenlee noted elsewhere in the discussion, "one of the risks is that the content creator doesn't know that it's happening."

Licensing. The same catalog becomes the deliverable in an approved deal — the actual asset a brand, platform, or AI company would license to generate new content featuring that creator's voice or likeness, with the creator's consent and, ideally, compensation built into the terms.

Is it worth doing now, or waiting?

Greenlee's own conclusion, having been through it, was to treat it as a proactive step rather than something to defer:

"I think it being proactive about creating your own likeness vault of sorts and start setting yourself up for being a part of this movement... I would advise is just get in there and create an account and get your likeness into the platform as soon as possible."

The underlying logic: the earlier a creator establishes a verified reference of their own voice and likeness, the earlier they can both detect misuse and participate in the licensing opportunities that are already starting to appear.

Key takeaways

An identity vault captures a creator's voice (via varied script readings) and likeness (via multi-angle face and body scans) to create a verified reference catalog.
The captured data serves two purposes: detecting unauthorized AI use of a creator's identity, and enabling approved licensing deals.
The process takes roughly 20+ minutes and is more involved than a typical account sign-up.
It's reasonable to feel uneasy about handing over this much biometric data — ask specific questions about storage, access, and deletion rights before enrolling.

FAQ

What is an AI identity vault for creators?

It's a structured capture of a creator's voice (multiple tones and scripts) and physical likeness (multi-angle face and body scans) used to create a verified reference that can detect unauthorized AI reproductions and support licensed use of that creator's identity.

How long does the identity vault process take?

Based on one creator's firsthand account, roughly 20 minutes or more to complete the core voice and likeness capture steps.

Is it safe to submit this much biometric data to a platform?

That depends entirely on the platform's data handling practices. Before enrolling, ask how the data is stored and secured, who can access it, what happens to it if the company is acquired or shuts down, and whether you can revoke consent or delete your data later.

Source: Adapted from commentary by Rob Greenlee, podcast industry veteran and Podcast Hall of Fame chair, in the Royall.ai "Who Owns You" webinar series, Episode 3: "AI Identity and the Future of Creator Endorsements."