The paper introduces CertDW, a certified dataset watermark and ownership verification method that remains reliable even under malicious perturbations. By leveraging conformal prediction, it defines two statistical measures—principal probability (PP) and watermark robustness (WR)—to evaluate model stability on benign versus watermarked samples. The authors derive certification conditions linking WR to a PP-based threshold and provide a high‑probability bound on false positives, enabling robust ownership verification when a suspicious model’s WR exceeds the PP values of benign models.
By Ting Qiao, Yiming Li, Jianbin Li, Yingjia Wang, Leyi Qi, Junfeng Guo, Ruili Feng, Dacheng Tao
arXiv:2606. 23335v2 Announce Type: replace-cross 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.
By Nicolas M. M\"uller, Pascal Debus
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.
The paper introduces TP-CRIV, a framework for verifying the identity of AI models through third‑party challenge‑response interactions without requiring white‑box or API access. TP-CRIV operates in a black‑box setting, using fresh, undisclosed challenges and network isolation to ensure that verification relies solely on the claimant’s local model. The authors demonstrate the approach on ten ImageNet‑pretrained CNNs, achieving clear separation between same and cross‑model responses with statistically calibrated thresholds.
By Teruki Sano, Minoru Kuribayashi, Masao Sakai, Shuji Isobe, Eisuke Koizumi, Zhang Zhang, Satoru Matsumoto
The paper introduces "watermarkability," a metric that quantifies how compatible a language model is with a logit‑based watermark that boosts the probability of a specified token set during decoding. By measuring the expected minimum KL divergence needed to achieve a target probability mass, the authors can assess and compare models and watermark rules. They propose a two‑stage approach—first training a low‑rank adapter conditioned on the target set, then applying a minimal uniform logit bias—to reduce this cost, achieving significant perplexity reductions across multiple watermark schedules on OPT‑125M and Qwen3‑8B.
By Zewei Deng, Mohamed Seif, Liyan Xie
The paper introduces TP-CRIV, a framework for verifying the identity of AI models through third‑party challenge‑response interactions. It operates without white‑box or API access, relying only on black‑box inference and fresh, undisclosed challenges to gather empirical evidence of model possession. The authors demonstrate the method on ten ImageNet‑pretrained CNNs, achieving clear separation between matching and non‑matching models with statistically calibrated thresholds.
MarkTune is a model‑embedded watermarking framework that improves the trade‑off between generation quality and detectability. By regularizing toward a high‑quality reference distribution, it decouples quality preservation from proximity to the base model and uses the GaussMark detection statistic as a reward during on‑policy fine‑tuning. Empirical results show that MarkTune enhances the quality‑detectability frontier, maintains downstream task performance, and remains robust against text‑editing and open‑weight attacks across diverse datasets.
By Yizhou Zhao, Zhiwei Steven Wu, Adam Block
arXiv:2501. 15509v5 Announce Type: replace-cross Abstract: Model fingerprinting has emerged as a crucial mechanism for safeguarding the intellectual property of open-source models, offering a non-intrusive approach that requires no modifications to the protected model.
By Shuo Shao, Haozhe Zhu, Yiming Li, Hongwei Yao, Tianwei Zhang, Zhan Qin
arXiv:2606. 18430v1 Announce Type: new Abstract: Statistical watermarks help organizations attribute large language model (LLM) outputs, yet existing detectors often struggle when watermark signals are weak, texts are repetitive, or watermarks are edited.
By Chih-Duo Hong, Yen-Pang Chen, Fang Yu
arXiv:2603. 23171v3 Announce Type: replace-cross Abstract: Providers monitor deployed large language models (LLMs) to detect misuse that they cannot prevent.
By Toluwani Aremu, Daniil Ognev, Samuele Poppi, Nils Lukas
arXiv:2606. 11698v1 Announce Type: cross Abstract: Model watermarking safeguards AI model intellectual property by embedding distinctive knowledge that induces unique behavioral signatures.
By Jian-Ping Mei, Weibin Zhang, Ao Yao, Tiantian Zhu, Jie Xiao
arXiv:2607. 00325v1 Announce Type: new Abstract: A growing body of literature suggests that training data membership inference problems are fundamentally hard tasks in modern language modeling settings.
By John Kirchenbauer, Brian R. Bartoldson, Bhavya Kailkhura, Tom Goldstein