Researchers Bypass Deepfake Detection With Replay Attacks www.darkreading.com/cybersecu…

As synthetic audio continues to improve, it’s also getting harder for anti-spoofing models to accurately detect.

That’s according to recent research published on June 1 by a team of researchers at German, Polish, and Romanian universities as well as Resemble AI, a vendor that provides AI voice generation tools as well as deepfake detectors. The research team presented how “replay attacks” are able to bypass audio deepfake detections. “By playing and re-recording deepfake audio through various speakers and microphones, we make spoofed samples appear authentic to the detection model,” they wrote.