TikTok tests AI likeness detection to help creators report unauthorized deepfakes

Lead

AI-generated videos and likeness-cloning tools have made it easier for people to imitate public figures, influencers, and online creators. TikTok is now testing a new way to respond: an opt-in AI likeness detection tool that can scan for content that may be using a creator’s image without authorization.

According to The Verge, the feature is being tested with “some” creators in the United States. TikTok US spokesperson Zachary Kizer confirmed the test, which was also spotted by social media consultant Matt Navarra. The move comes as other major platforms are building similar protections; YouTube has been working on a comparable tool and recently made it available to all adult users.

Key points

• The feature is opt-in: creators included in the pilot must choose to activate the tool before TikTok scans for possible AI likeness matches.
• Identity verification is required: users need to verify themselves through Jumio, including a real-time selfie scan and an ID check.
• TikTok is limiting the stated data use: the company says it does not retain ID documents, and that facial information is used only for likeness matching and to help identify potential unauthorized uses.
• Creators can review and report results: once TikTok’s system finds possible matches, creators can review the content and report suspicious posts or accounts.
• The broader issue is platform responsibility: AI impersonation is becoming a product, safety, and rights-management problem for social platforms, not just an isolated moderation concern.

Why it matters

For creators, likeness misuse is not merely embarrassing or inconvenient. A convincing AI-generated video can damage reputation, confuse audiences, or be used to promote scams and misleading content. The burden has often fallen on creators to discover the material themselves and file reports after the fact. A platform-level scanning tool could make that process faster and more practical, especially for creators whose face is central to their work and business.

The test also shows how AI safety is moving into everyday creator tools. Traditional moderation systems focused on whether a post violated a platform rule. Likeness detection adds a different question: does this piece of media appear to use a real person’s identity without permission? That requires identity verification, face matching, review workflows, and a process for creators to dispute or report misuse.

At the same time, the approach raises its own trust questions. To protect someone’s likeness, TikTok must process sensitive identity and facial information. The company says ID documents are not retained and that facial data is used only for matching and misuse detection. Those limits are important, but creators will still need confidence that the system is accurate, that false positives are handled fairly, and that privacy protections are clear.

The feature remains a limited test, and TikTok has not announced a broader rollout. Still, the direction is significant: as AI deepfakes become easier to create, major platforms are beginning to give creators more direct tools to detect and report unauthorized uses of their identity.

Source: The Verge AI