The Watermark Shortcut: How Provenance Marking Sabotages Audio Deepfake Detection

• URL: http://arxiv.org/abs/2606.23335v1
• Date: Mon, 22 Jun 2026 13:42:31 GMT
• Title: The Watermark Shortcut: How Provenance Marking Sabotages Audio Deepfake Detection
• Abstract summary: Provenance watermarking is increasingly treated as a safeguard for synthetic speech.<n>We identify a previously uncharacterized liability: when synthetic speech is watermarked and human speech is not, detectors trained alongside latch onto the watermark as a spurious "mark fake" shortcut.<n>We show that retraining with the watermark on both classes decorrelates it and restores clean behavior.
• Score: 5.288260916272508
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Provenance watermarking is increasingly treated as a safeguard for synthetic speech, whether built directly into speech-generation models such as Chatterbox, provided through dedicated techniques such as AudioSeal, or deployed by commercial platforms such as ElevenLabs. We identify a previously uncharacterized liability: when synthetic speech is watermarked and human speech is not, detectors trained alongside latch onto the watermark as a spurious "watermark => fake" shortcut. This single feature yields three coupled failures: generalization degradation (model performance deteriorates on unseen data), strip-to-evade (a watermarked fake escapes once unwatermarked), and mark-to-frame (watermarking a real voice flags it as fake). In a controlled white-box experiment, a watermark-trained detector shows all three (for example, mark-to-frame lifts Equal Error Rate from 16% to 75%). In a black-box test of a commercial API, we show that adding a watermark to real speech disguises it as fake. However, this shortcut is fixable: retraining with the watermark on both classes decorrelates it and restores clean behavior. We release experiment data as a paired clean-versus-watermarked corpus (WASP).

Related papers

• Audio Pirates: Black-box Audio Watermark Removal via Diffusion Priors [44.562700127905295]
  We propose DiffErase, a black-box watermark removal attack that assumes no knowledge of the target watermarking scheme.<n>We show that DiffErase consistently removes watermarks while preserving perceptual quality across multiple audio domains.<n>These findings highlight the need for future audio watermarking designs to consider diffusion-based threats.
  arXiv  Detail & Related papers  (2026-05-28T22:07:32Z)
• Hidden in Plain Tokens: Simply Robust, Gradient-Free Watermark for Synthetic Audio [58.612001688217056]
  Inference-time watermarks for autoregressive models are unfit for continuous modalities due to discretization inconsistencies.<n>We propose an elegant solution for powerful and robust watermarking of synthetic audio.
  arXiv  Detail & Related papers  (2026-05-25T15:43:20Z)
• SSCL-BW: Sample-Specific Clean-Label Backdoor Watermarking for Dataset Ownership Verification [8.045712223215542]
  This paper proposes a sample-specific clean-label backdoor watermarking (i.e., SSCL-BW)<n>By training a U-Net-based watermarked sample generator, this method generates unique watermarks for each sample.<n>Experiments on benchmark datasets demonstrate the effectiveness of the proposed method and its robustness against potential watermark removal attacks.
  arXiv  Detail & Related papers  (2025-10-30T12:13:53Z)
• Robustness of AI-Image Detectors: Fundamental Limits and Practical Attacks [47.04650443491879]
  We analyze the robustness of various AI-image detectors including watermarking and deepfake detectors. We show that watermarking methods are vulnerable to spoofing attacks where the attacker aims to have real images identified as watermarked ones.
  arXiv  Detail & Related papers  (2023-09-29T18:30:29Z)
• On the Reliability of Watermarks for Large Language Models [95.87476978352659]
  We study the robustness of watermarked text after it is re-written by humans, paraphrased by a non-watermarked LLM, or mixed into a longer hand-written document. We find that watermarks remain detectable even after human and machine paraphrasing. We also consider a range of new detection schemes that are sensitive to short spans of watermarked text embedded inside a large document.
  arXiv  Detail & Related papers  (2023-06-07T17:58:48Z)
• Invisible Image Watermarks Are Provably Removable Using Generative AI [47.25747266531665]
  Invisible watermarks safeguard images' copyrights by embedding hidden messages only detectable by owners. We propose a family of regeneration attacks to remove these invisible watermarks. The proposed attack method first adds random noise to an image to destroy the watermark and then reconstructs the image.
  arXiv  Detail & Related papers  (2023-06-02T23:29:28Z)
• Tree-Ring Watermarks: Fingerprints for Diffusion Images that are Invisible and Robust [55.91987293510401]
  Watermarking the outputs of generative models is a crucial technique for tracing copyright and preventing potential harm from AI-generated content. We introduce a novel technique called Tree-Ring Watermarking that robustly fingerprints diffusion model outputs. Our watermark is semantically hidden in the image space and is far more robust than watermarking alternatives that are currently deployed.
  arXiv  Detail & Related papers  (2023-05-31T17:00:31Z)
• SepMark: Deep Separable Watermarking for Unified Source Tracing and Deepfake Detection [15.54035395750232]
  Malicious Deepfakes have led to a sharp conflict over distinguishing between genuine and forged faces. We propose SepMark, which provides a unified framework for source tracing and Deepfake detection.
  arXiv  Detail & Related papers  (2023-05-10T17:15:09Z)
• Did You Train on My Dataset? Towards Public Dataset Protection with Clean-Label Backdoor Watermarking [54.40184736491652]
  We propose a backdoor-based watermarking approach that serves as a general framework for safeguarding public-available data. By inserting a small number of watermarking samples into the dataset, our approach enables the learning model to implicitly learn a secret function set by defenders. This hidden function can then be used as a watermark to track down third-party models that use the dataset illegally.
  arXiv  Detail & Related papers  (2023-03-20T21:54:30Z)
• Certified Neural Network Watermarks with Randomized Smoothing [64.86178395240469]
  We propose a certifiable watermarking method for deep learning models. We show that our watermark is guaranteed to be unremovable unless the model parameters are changed by more than a certain l2 threshold. Our watermark is also empirically more robust compared to previous watermarking methods.
  arXiv  Detail & Related papers  (2022-07-16T16:06:59Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.