YouTube declares war on deepfakes with new tool that lets creators flag AI-generated video clones (www.techradar.com)

YouTube has rolled out a new deepfake detection tool that lets verified creators opt in to scan the platform for videos that use their face or voice without permission. Modeled on Content ID, the system requires creators to submit a government ID and a short video sample so YouTube can build a biometric baseline; enrolled creators receive alerts in a new Content Detection tab and can review matches, report them, request takedowns under privacy rules, or file copyright claims. The feature is initially limited to YouTube Partner Program members but is likely to expand to more monetized creators. Technically, the tool uses pattern-matching on facial and vocal features rather than full forensic reconstruction, so it can catch many direct impersonations quickly but may miss heavily manipulated, stylized, or low-resolution fakes. The move signals a strategic shift toward treating likenesses like digital assets to be protected, giving creators a proactive remediation route against impersonation and misinformation. It also raises trade-offs: creators must trust YouTube with biometric data and rely on the platform’s enforcement speed, while detection limits mean this is a deterrent, not a complete fix. Compared with Meta and TikTok’s labeling or tagging approaches, YouTube’s system is a more direct, takedown-oriented response to malicious synthetic media.