The short video platform TikTok is launching a pilot test of a new tool that fundamentally changes the rules of the game in the fight against deepfakes. This is the Likeness Detection feature — a system capable of automatically identifying content created using artificial intelligence that involves a specific author's face.
Currently, the option is only available to a limited group of creators in the United States. The main goal of the tool is to give bloggers the ability to quickly respond to fraudulent activities: from filing a complaint about a deepfake to blocking a fake account that uses their likeness.
The activation process requires passing a strict identity verification. Authors must provide a real-time selfie and a scan of their ID through the Jumio service. It is important to note that, according to the company, copies of documents are not stored on TikTok's servers, and biometric data is used solely to find matches with content generated by neural networks.
This step is not just a technical innovation, but a response to the growing threat of reputational and financial losses for influencers. Amid the avalanche-like growth of AI fraud, where attackers create realistic videos with celebrities' faces to advertise scam projects, such mechanisms are becoming critically important.
My expert assessment: TikTok's initiative demonstrates market maturity: platforms are beginning to move from passive moderation to proactive protection. However, the key question is how effectively Likeness Detection will work with high-quality deepfakes. Jumio's technology is good for initial verification, but in the long term, a decentralized registry of faces will be needed to avoid a single point of failure in security.