TikTok is testing a instrument that lets creators discover and report AI-generated movies that use their face with out their permission.
The opt-in characteristic, known as Likeness Detection, is being examined with a restricted group of creators within the United States.
It was first noticed by social media guide Matt Navarra, who wrote in a put up on X on Friday (July 17) that TikTok was testing the characteristic with “some US creators.”
“It surfaces AI-generated content material that will use a creator’s face with out consent,” Navarra wrote.
The take a look at is proscribed to a small group of US creators, TikTok US spokesperson Zachary Kizer informed The Verge.
To make use of the instrument, creators first confirm their id by Jumio, a third-party identity-verification service.
The method includes a real-time selfie scan and a authorities ID verify.
As soon as a creator is verified, TikTok‘s techniques scan the platform for AI-generated movies that seem to make use of that individual’s face.
“TikTok doesn’t retain ID paperwork, and facial info is used just for likeness matching and to assist determine potential unauthorized makes use of of a creator’s likeness.”
Zachary Kizer, TikTok (through The Verge)
Creators can then overview any matches and report the posts or accounts they consider are utilizing their likeness with out authorization.
Kizer stated that “TikTok doesn’t retain ID paperwork, and facial info is used just for likeness matching and to assist determine potential unauthorized makes use of of a creator’s likeness.”
The characteristic is TikTok‘s newest transfer in opposition to deepfakes and AI impersonation.
It’s separate from the platform’s current work to label AI-generated content material.
Days earlier, on July 10, TikTok stated it had labeled greater than 3 billion movies as AI-generated content material, utilizing a mixture of Content material Credentials, creator labeling instruments and invisible watermarking.
The identical day, TikTok stated it was becoming a member of the Steering Committee of the Coalition for Content material Provenance and Authenticity (C2PA), the physique behind the Content material Credentials normal, which TikTok says it was the primary video platform to undertake two years in the past.
The usual flags when a bit of content material has been generated or edited with AI.
TikTok added that it had eliminated greater than 86 million pretend accounts within the first quarter of 2026, and that it could start testing improved detection techniques for accounts that put up AI-generated spam.
“At TikTok, we consider AI can rework how individuals share their creativity, uncover new passions, and keep secure on our platform when used transparently and responsibly,” the corporate stated.
On the music aspect, TikTok has already been policing manipulated audio.
Its distribution arm, SoundOn, in April started utilizing content-recognition agency ACRCloud‘s Spinoff Works Detection service to intercept unauthorized uploads earlier than they attain streaming companies.
That system additionally requires uploaders to confirm their id with picture ID, with flagged tracks despatched for human overview.
TikTok is just not the primary platform to supply creators this type of safety.
YouTube started rolling out its personal likeness-detection instrument in October 2025, initially for a restricted set of creators.
In April, YouTube prolonged the instrument to celebrities and expertise businesses, no matter whether or not they have a YouTube channel.
These companions included CAA, UTA, WME and Untitled Administration. In Might 2026, YouTube widened the instrument once more, opening it to all customers aged 18 and over.
YouTube says its system works like Content material ID, its copyright-matching know-how, besides that it scans for a creator’s face fairly than for copyrighted materials.
The YouTube instrument at present detects faces solely, and the corporate says it goals to increase it to voice in 2026.
The instruments are arriving because the music business escalates its personal struggle in opposition to AI deepfakes and impersonation.
In March, Sony Music requested streaming companies to take down greater than 135,000 tracks it stated had been made by fraudsters utilizing generative AI to impersonate artists on its roster.
Dennis Kooker, President, International Digital Enterprise & US Gross sales at Sony Music Leisure, has stated such deepfakes trigger “direct business hurt to official recording artists.”
Additionally in March, Spotify started piloting an opt-in characteristic, Artist Profile Safety, that lets artists approve releases earlier than they seem on their profiles.
Rival service Deezer has stated it receives practically 75,000 totally AI-generated tracks a day, about 44% of its each day uploads.
The music business’s concern over AI voice-cloning dates to the 2023 “pretend Drake” observe, which drew tons of of hundreds of streams earlier than companies pulled it.
For now, TikTok‘s Likeness Detection instrument stays a take a look at, out there to a small group of US creators, and the corporate has not stated when – or whether or not – it’ll roll out extra broadly.Music Enterprise Worldwide
