Enrolling in Likeness Detection, Step by Step

Most creators who'd qualify for YouTube's likeness detection have never opened the menu it lives in. It's a genuinely free, opt-in protection against unauthorized AI likenesses of your face showing up in other people's uploads — and it asks for something real in exchange: a biometric scan of that same face, held by the platform, to make the matching possible in the first place. Both halves of that trade are worth stating plainly before walking through the setup.

Who's actually eligible

Two gates, both hard requirements rather than guidelines. You must be over 18 years old to set up likeness detection, and you must be a Channel Owner or Manager on the channel you're enrolling — not an Editor, and not a role with narrower permissions than that. If either condition isn't met, the enrollment flow isn't available to you at all, regardless of how much AI-generated content featuring your face already exists elsewhere on the platform.

What enrollment actually asks for

This is the part worth being upfront about rather than glossing past on the way to the setup steps. Enrolling means agreeing to a biometric facial scan, which requires submitting a government-issued ID and recording a brief selfie video of your face — and the verification process may take up to 5 days to complete once both are submitted. That's a real privacy trade: you're handing over an ID and a biometric video specifically so the system has something to match unauthorized likenesses against. The protection is free in the sense that no money changes hands, but it isn't free in the sense of costing nothing — it costs a biometric enrollment, which is worth deciding deliberately rather than clicking through without registering what was just agreed to.

The enrollment path

From YouTube Studio, the feature sits under Content detection, then Likeness, with a Start now button that begins the ID-and-video submission flow. Once submitted, the up-to-5-day verification window runs before the system is actively matching against your enrolled likeness. There's a second entry point in the same section — For review — which is where matches surface once enrollment is active, rather than part of the setup itself.

What happens when there's a match

Once enrolled, a detected match gives you three distinct actions, not one blanket takedown button:

• Submit a likeness removal request — the path for a privacy-violation claim, when your likeness appears somewhere you didn't authorize.
• Submit a copyright removal request — the path when the match also involves unauthorized use of copyrighted material, a related but legally distinct claim from the likeness issue itself.
• Move to archive — removes the match from your active review list without filing either request, for matches you've reviewed and decided not to act on, while the video in question stays live.

Three separate actions matched to three separate outcomes is worth knowing going in, because the right response to a match isn't always a takedown — sometimes it's confirming there's nothing to act on and clearing it from the queue.

Who else can act on your behalf

Enrollment itself is locked to Channel Owners and Managers, but reviewing matches afterward isn't restricted quite that tightly — Managers, Editors, and limited Editors can all review matches and report privacy violations on behalf of any Owner or Manager on the channel. That's a meaningful operational detail for a channel with a production team: the person who enrolled doesn't have to personally triage every match that comes in afterward, even though only they (or another Owner/Manager) could have started the enrollment in the first place.

Why the awareness gap exists

This isn't a hidden feature in the sense of being deliberately buried — it's simply new enough, and specific enough in its eligibility gate, that most creators who'd qualify haven't gone looking for it. A protection this specific to a problem (unauthorized AI likenesses, not general copyright infringement) tends to get discovered reactively, after a creator finds a synthetic version of themselves circulating, rather than proactively as routine channel setup. Given the biometric enrollment is the same regardless of whether you sign up before or after a problem shows up, there's no real advantage to waiting for a reason.

A Versely walkthrough: auditing your own footprint first

Before or alongside enrolling, it's worth knowing how many of your own past generations actually used your face as a reference — a useful check for anyone who's built face-locked content across the kind of creator and business workflows that lean on a consistent avatar, a UGC-style presenter, or a cloned-voice character, and wants to confirm every use was one they authorized and disclosed:

"Search my past generations for anything that used my face as a reference image, across images and videos."

That's search_my_media with a query describing your own likeness and type scoped to images and videos — a meaning-based search over your generation history rather than a manual scroll through months of output. It's a different job from browse_user_library, which covers uploaded reference files and slideshows rather than the generations built from them. Pairing the two — an audit of your own authorized uses, plus enrollment in a platform-side check for unauthorized ones elsewhere — covers both directions of the same underlying concern: your face, used with your consent, and your face, used without it.

Where this connects to a brand deal

If your face is the asset in a paid partnership — a licensed avatar read, a UGC-style ad built from your likeness — the usage rights attached to that specific use matter as much as the platform-side protection does. Likeness detection catches unauthorized copies; a properly scoped usage-rights clause is what keeps an authorized use from quietly outliving the deal it was made for. They're solving adjacent problems, not the same one, and a creator monetizing their own likeness benefits from treating both as routine rather than reactive. The same logic extends to any AI-generated content featuring a real person — disclosure obligations exist independently of whether the use was authorized.

FAQ

Can I enroll if I'm under 18 or an Editor rather than an Owner/Manager? No — both are hard gates on the enrollment flow itself. Under-18 creators and channel roles below Manager can't start the process, regardless of how much unauthorized content might exist.

How long does verification actually take? Up to 5 days from submitting the government ID and selfie video, per YouTube's own stated processing window. The system isn't actively matching against your likeness until that verification completes.

What's the difference between the three actions on a match? A likeness removal request addresses an unauthorized use of your face specifically. A copyright removal request addresses unauthorized use of copyrighted material within the same match. Archiving clears the match from your review queue without filing either request, while leaving the video live.

Does enrolling protect against every kind of unauthorized AI content featuring me? It's specifically a likeness-matching system — built to detect your face appearing in content you didn't authorize. It's not a general-purpose content-monitoring tool for every mention or reference to you that isn't visual.

The setup itself takes minutes once you've decided to do it. The actual decision is upfront: whether trading a government ID and a biometric face scan for a real, ongoing protection is worth it for your specific channel — and for most creators with a public face and enough of a following to be worth impersonating, it is.