The short video platform TikTok is launching a pilot test of the Likeness Detection feature, designed to protect content creators from unauthorized use of their likeness in deepfakes. In the initial phase, the tool is available to a limited group of American creators, allowing the company to evaluate the effectiveness of the algorithms in real-world conditions.
How does protection against digital clones work?
The Likeness Detection mechanism automatically scans the platform for AI-generated content that contains a specific creator's face. When suspicious material is detected, the author receives a notification and the ability to file a complaint against the deepfake or fake account. Activating the tool requires identity verification through the Jumio service, including a real-time selfie and scanning of an identification document.
An important aspect is the privacy policy: TikTok states that it does not store copies of users' documents, and biometric data is used solely for matching against content on the platform. This reduces the risks of sensitive information leakage, although it leaves questions about the long-term security of such data.
Market context and implications
The launch of Likeness Detection comes amid a rapid increase in the number of deepfakes on social networks, especially using face generation technologies. According to analysts, the number of fake videos featuring well-known bloggers has increased by 230% over the past six months. TikTok, as a platform with an audience of over 1 billion users, is becoming a key target for malicious actors.
My professional opinion: this step is not just about copyright protection, but a strategic response to regulatory pressure from the EU and the US, where laws on labeling AI-generated content are tightening. However, the tool's effectiveness will depend on the speed of deepfake detection and the accuracy of the algorithms — otherwise, we risk false positives that could harm legitimate creators. The market awaits the first test results to assess whether Likeness Detection will become an industry standard or remain a niche experiment.