TikTok’s next AI trust problem is no longer just labeling what was made by machines. It is helping creators find synthetic versions of themselves.
As spotted by Matt Navarra, TikTok is starting to test an opt-in AI likeness detection tool that scans the platform for unauthorized AI likenesses and lets creators report them to the company. The test was first spotted by social media consultant Matt Navarra, and TikTok US spokesperson Zachary Kizer confirmed that it is initially being tested with “some” creators in the US.
That makes this more than another AI safety toggle. It moves part of the deepfake response closer to the people most likely to spot the damage first: the creators whose faces, voices, and identities are being copied.
A tool for the deepfake problem creators can actually see
The basic promise is simple: creators opt in, TikTok scans for AI likenesses, and suspicious uses can be reported. The feature is still a test, and TikTok has not framed it as a broad public rollout. But the shape of the product says a lot about where platform safety is heading.
For years, deepfake enforcement has mostly sounded like a moderation problem. Platforms write policies, remove content after reports, and add labels when AI-generated media is detected or declared. But creator likeness is more personal, more commercial, and often more urgent. A fake video does not need to go viral to cause damage. It only needs to look convincing enough to confuse an audience, mislead a fanbase, or attach a creator’s identity to something they never said.
That is why the opt-in detail matters. TikTok is not just trying to detect “AI content” as a broad category. It is testing a mechanism around identity. The creator becomes part of the detection loop, not just the person filing a complaint after the fact.
The Verge also notes that YouTube has been working on a similar feature, which points to the larger platform race now forming around synthetic identity. As generative AI gets easier to use, creators are becoming both the fuel and the target: their likenesses help train audience expectations, and those same likenesses can be copied, remixed, and pushed back into the feed without consent.
Creator protection is becoming platform infrastructure
This is the part brands should watch too. If creator marketing depends on trust, then unauthorized AI likenesses are not just a creator safety issue; they are a campaign integrity issue. A synthetic clip can blur what is endorsed, what is parody, what is fraud, and what is just another piece of content moving through the system.
TikTok’s test suggests platforms are starting to treat likeness protection as a core trust layer, not a special case for celebrities. That matters because TikTok is built around recognizable faces. Its culture rewards personality, familiarity, and repetition. The more a creator’s face becomes an asset, the more valuable it becomes to copy.
There is still friction here. A detection tool only helps if it finds enough real misuse, avoids drowning creators in false positives, and leads to fast enforcement. But the direction is clear: AI safety is moving from abstract labels into personal rights management inside the feed.
For creators, the strategic shift is control. For platforms, it is responsibility. If TikTok wants creators to keep turning themselves into media businesses, it also has to help protect the identity those businesses are built on.