TikTok is testing a deepfake detector: how the platform will protect bloggers from AI copies
Social platform TikTok is launching testing of a new security tool — Likeness Detection. This optional mechanism is designed to detect AI-generated content that uses a creator's face without their consent. Currently, the feature is available to a limited group of creators in the United States.
The tool works as follows: the system automatically scans the platform for AI-generated videos and images that feature a specific blogger's face. When such content is detected, the creator receives a notification and can file a complaint against the deepfake or fake account.
Verification Procedure
To activate the feature, users must undergo a strict identity verification process through the Jumio service. The process includes two stages: a real-time selfie and scanning of an identity document. This is necessary for the system to accurately match the creator's biometric data with content distributed online.
It is important to note the privacy policy: the company states that it does not store copies of users' documents. Biometric data is used exclusively for finding matches and is not shared with third parties. This reduces the risks of sensitive information leakage, although it raises questions about the long-term storage of biometric templates.
Expert opinion: This step by TikTok is a timely response to the avalanche-like growth of deepfakes on social networks. However, the tool's effectiveness will depend on the speed of complaint processing and the accuracy of recognition algorithms. In an environment where AI generation is becoming increasingly realistic, passive detection may not keep up with malicious actors' activity. The platform will need to implement proactive measures, including labeling all synthetic content at the upload stage.