TikTok tests a tool to catch AI clones of its creators

The News

TikTok has begun testing an opt-in tool that scans the platform for AI-made likenesses of its creators and lets them flag those clips to the company for review. The feature was first spotted by social media consultant Matt Navarra and confirmed to The Verge by TikTok US spokesperson Zachary Kizer, who said it is being trialled with "some" creators in the United States.

The mechanics are deliberately gated. Creators who join the test and want to use the tool must first verify their identity with the company, giving TikTok a reference point against which suspected synthetic copies can be matched. The move mirrors YouTube, which has built a comparable likeness-detection system and recently opened it to all adult users on its platform.

It is a narrow launch, but a pointed one. For the first time, one of the world's largest short-video platforms is offering creators a formal channel to police AI impersonations of their own faces and voices.

Why It Matters

The test signals that likeness protection is shifting from a legal afterthought to a core platform feature. For most of the generative-AI boom, the burden of policing deepfakes fell on victims, who had to discover fakes themselves and then chase takedowns through opaque reporting queues. Building detection into the product inverts that arrangement: the platform does the scanning and the creator does the confirming.

The timing is not accidental. When realistic face-swap and voice-cloning tools went mainstream roughly two years ago, platforms responded mostly with policy language rather than working tooling. YouTube's decision to widen access to its own likeness system, and now TikTok's trial, suggest the industry has concluded that detection infrastructure, not terms-of-service clauses, is what actually reassures creators and advertisers.

There is commercial logic beneath the safety framing. Creator trust is the raw material of these platforms' advertising businesses, and if audiences can no longer tell a genuine endorsement from a synthetic one, the value of every sponsored post erodes. Likeness detection is, in that sense, brand-safety plumbing dressed as creator protection.

Indian Angle

The irony for India is sharp. TikTok has been banned in the country since June 2020, so none of this reaches Indian creators directly. Yet the problem the tool addresses is arguably more acute in India than in most markets where TikTok still operates. The 2023 deepfake clip of actor Rashmika Mandanna triggered a national outcry and pushed MeitY to issue advisories to platforms on synthetic media, and Indian courts have since become unusually assertive. The Delhi High Court has granted personality-rights protection to figures such as Anil Kapoor and Amitabh Bachchan, effectively recognising a face and voice as protectable assets.

For Indian creators, who number in the millions across Instagram, YouTube and homegrown apps such as Moj and Josh, the practical question is whether those platforms import features like this one. YouTube's likeness tool already applies to Indian users, and pressure from MeitY, which has moved towards mandatory labelling of AI-made content, makes similar tooling on Meta and domestic platforms increasingly likely.

There is also a regulatory read-across. India's DPDP framework and its evolving IT Rules treat unauthorised synthetic likenesses as both a privacy and an intermediary-liability question. A platform that can technically detect a deepfake has a harder time arguing it could not have known, a shift that could reshape how Indian courts assign responsibility.

FAQ

Who can use TikTok's new tool?

Only "some" US creators enrolled in the test, according to TikTok. Participants must verify their identity with the company before the likeness-scanning feature is switched on, so it is not yet a general rollout to the wider user base.

How is this different from YouTube's system?

Both scan for AI-made likenesses, but YouTube has already opened its version to all adult users on its platform. TikTok's tool remains an opt-in trial with a limited group of US creators at this early stage.

Does this affect creators in India?

Not directly, because TikTok has been banned in India since 2020. But the underlying deepfake problem is very much live in India, and comparable tools on YouTube, and potentially Meta, already touch Indian creators.

Where can I read the original report?

The Verge first reported the test, including comment from TikTok. The link to the full coverage appears in the attribution paragraph below.

This story was reported by The Verge. Read the full original coverage at The Verge.