video1.0<iframe src="https://www.loom.com/embed/588afd60a82b417d987ad507795031da" frameborder="0" width="1728" height="1296" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>12961728Loomhttps://www.loom.com12961728https://cdn.loom.com/sessions/thumbnails/588afd60a82b417d987ad507795031da-1fb74f6ef1f07c1c.gif89.776This Loom presents EchoProof, a system for detecting AI-generated audio while keeping audio local on the user’s computer. The speaker explains that the model now distinguishes real versus deepfake audio more reliably after training on over 12,000 samples and testing it against examples such as a “genuine microphone” recording and an 11 Labs clip. During development using GCV 4.1 with three agents, validation improved from 95.11% to 95.68% after live runs involving nine cloud calls and eight experiments. The work concludes by expressing hope EchoProof can help stop “serials.”