AIBearisharXiv – CS AI · 10h ago7/10
🧠
The Watermark Shortcut: How Provenance Marking Sabotages Audio Deepfake Detection
Researchers discovered that audio deepfake detectors trained on watermarked synthetic speech and unwatermarked real speech exploit watermarks as a spurious shortcut, causing three critical failures: poor generalization, watermarked fakes evading detection, and real watermarked speech being flagged as fake. The vulnerability affects commercial platforms like ElevenLabs and AudioSeal, though retraining detectors with watermarks on both classes resolves the issue.