Overview
- The pilot extends YouTube’s AI likeness detection to selected government officials, political candidates, and journalists, surfacing detected deepfakes for their review.
- Removal is not automatic, as requests are assessed case by case with longstanding protections for parody, satire, and political critique.
- Participants must verify identity with a selfie and government ID; YouTube says this data is used only for verification, not to train generative models, and opt‑outs trigger deletion of setup data.
- The tool first reached roughly 4 million YouTube Partner Program creators last year, with broader access planned along with potential pre‑upload blocking, voice detection, and monetization controls.
- YouTube reports a very small volume of removals to date from creators using the system and is advocating for the NO FAKES Act to codify rights against unauthorized AI recreations.