YouTube tightens grip on deepfakes with celeb-focused AI tools

YouTube is rolling out an update to its AI likeness detection system to include celebrities, giving stars and their representatives a way to flag and remove deepfake videos. The expansion follows rising concerns over manipulated content featuring public figures without consent. The tool, first introduced last year, now covers thousands of verified accounts. It scans uploaded videos for synthetic or altered footage that mimics a person’s face or voice. When detected, the system sends alerts to account holders or designated contacts. Users can then request removal under YouTube’s policy against impersonation and privacy violations. Deepfakes have increasingly targeted high-profile individuals, from politicians to musicians. Last month, a viral clip falsely showed a major singer endorsing a product. The incident led to calls for stronger platform controls. YouTube says the AI model has improved accuracy, reducing false positives by 30% since testing began. The move aligns with broader industry efforts. Meta and TikTok have also rolled out similar tools in recent months. Experts warn that without regulation, deepfakes could erode trust in online media. YouTube plans to expand detection to more languages and regions by year-end. Creators and rights groups have welcomed the update. One talent manager told TechCrunch the system cuts response time from days to hours. Still, critics argue it may not catch all cases, especially with rapidly evolving AI techniques. Source: techcrunch.com