YouTube announces new tool to detect fake videos

YouTube has launched a new AI system to detect deepfake videos that mimic creators' faces and voices. This feature, currently available to verified users enrolled in the YouTube Partner Program, allows creators to review and remove deepfake videos that falsely or misleadingly represent them.

The system constantly scans new uploads using facial and voice recognition algorithms, using reference data provided by content creators. Creators can then submit privacy-based takedown requests, file copyright claims, or archive the video.

Creators wishing to take advantage of the feature must complete an identity verification process involving a QR code scan, government-issued ID, and a short selfie video. YouTube verifies this data with Google servers and then activates the system through YouTube Studio. This process can take several days.

Precautions against deepfake videos

YouTube also noted that the system may still not fully distinguish between real videos and artificial content on a creator's own channel. This underscores that the system is still under development. A pilot program was tested last year with the Creative Artists Agency, with approximately 5,000 creators participating. The company plans to expand the system globally in January 2026.

The new tool aims to provide strong protection against AI-generated fake content, particularly for famous and high-risk creators. If successful, the system will remove one of the biggest questions for many celebrities and content creators.