YouTube Rolls Out AI Likeness Detection Tool to Prevent Deepfakes

[MacRumors]

YouTube today began rolling out a new AI likeness detection feature, which lets creators detect, manage, and request the removal of unauthorized videos that use AI to generate or alter the creator's facial likeness.

According to YouTube, the feature is meant to safeguard identities and prevent audiences from being misled by deepfakes.

The likeness detection tool is available in YouTube Studio under a content detection tab. After completing an identity verification process that requires a photo ID and a selfie video, creators will be alerted if there are any AI-generated videos that use their likeness. YouTube Studio will show a list of videos with titles, channel, views, and dialogue, along with an option to request a removal.

The tool supports likeness removal requests for AI videos, and copyright removal requests in case someone has used copyright-protected content without permission.

YouTube creators that are members of the YouTube Partner Program will get access to the likeness detection tool over the next few months. In a statement to TheWrap, YouTube said that the first creators selected to use the feature are those that "may have the most immediate use for the tool." All monetized creators will have access by January 2026.

Article Link: YouTube Rolls Out AI Likeness Detection Tool to Prevent Deepfakes

 

[MacUserFella]

But they can’t stop bot comments?

 

[Love-hate 🍏 relationship]

Awesome

 

[sleeptodream]

Very nice. They should have this for everyone that wants it, not just monetized youtubers

 

[Johnny907]

Ok. Can they do something about THIS nonsense next, please.
We tightly control access for our five year old, but she does have a Fire Kids tablet for road trips and games as a reward. Somehow THIS popped up despite being logged into her YT Kids account with restrictions applied, and it took an hour to calm her down and explain how AI isn’t real and those weren’t real puppies, and several discussions since about what was wrong with them being black. Fun times. We’ve been discussing ways to remove her access to YT without it feeling like a punishment because she did something wrong, because neither of us trust anything Google has to say given AI content was supposed to be blocked on Kids accounts.

 

[jz0309]

why stop at deepfakes? how about notification that AI has been used to create/modify a video?

 

[ThisIsMike]

Ahhhh welcome to the Pandora’s box. Enjoy your stay.

Oh, and why isn’t this tool for EVERYONE on YouTube? They’re sure happy enough to send anyone a take down notice for possible copyright violation, they can’t do the same here?

Don’t worry, I’m sure when it’s available to everyone, you’ll have to pay for it. So we’ll be paying big tech for the solution to the problem THEY created. Effing brilliant.

 

[diblasio]

Just share ur biometric data with friggn GOOGLE and you’ll be deep-fake free in no time! They PROMISE it won’t be used to train their own crappy chat bot! 🥴

 

[ProbablyDylan]

Very nice. They should have this for everyone that wants it, not just monetized youtubers

This is more about Google protecting their investments (those big youtubers) than it is about preventing AI impersonation.

 

[SpotOnT]

Anything produced with AI should be legally required to contain a digital watermark.

Then let people have away with whatever deepfakes they want. Everyone will know it is just an animation.

 

[boswald]

Some of the best and brightest engineers in the world and yet they didn’t anticipate something like this. Unbelievable.

 

[MysteriousStain]

What further horrors await us and our descendants?

 

[hagjohn]

I've been giving the thumbs down to the content that I see.

 

[zenmacx]

The naïveté of those here who want this...

Is mind boggling.

 

[loyalraccoon]

Don’t be fooled, this is a ruse to stop any satire or parody of people in power Google is in bed with.

 

[HiVolt]

why stop at deepfakes? how about notification that AI has been used to create/modify a video?

Or being able to simply block those videos? But youtube loves making money off them so that will never happen.

 

[Kabeyun]

Next up: AI figures out how to circumvent YouTube deep fake detection tool.

 

[burgman]

What further horrors await us and our descendants?

Nuclear dirt naps, Carrington events, Kessler Syndrome, Cheetos Lip Balm is rereleased.

 

[burgman]

I've been giving the thumbs down to the content that I see.

That doesn’t do what you think, just like downvotes on comments. They’re engagements. The 30 second rule stopping watching does ding them.

 

[ikramerica]

Ok. Can they do something about THIS nonsense next, please.
We tightly control access for our five year old, but she does have a Fire Kids tablet for road trips and games as a reward. Somehow THIS popped up despite being logged into her YT Kids account with restrictions applied, and it took an hour to calm her down and explain how AI isn’t real and those weren’t real puppies, and several discussions since about what was wrong with them being black. Fun times. We’ve been discussing ways to remove her access to YT without it feeling like a punishment because she did something wrong, because neither of us trust anything Google has to say given AI content was supposed to be blocked on Kids accounts.

Lesson learned. Never trust google. Never leave kid unsupervised with internet.

Remove access to youtube and switch to a kids account on a streamer or downloading videos.

 

[v3rlon]

Can they make it stop all the AI crap videos AND adds?
Anything that is all AI voice over or all AI video gets a flag, and we can select to not have those show up in our feed.
It wouldn't save us from AI script, but at least a human would have to read it.