Research

Paper

AI LLM February 25, 2026

Breaking Semantic-Aware Watermarks via LLM-Guided Coherence-Preserving Semantic Injection

Authors

Zheng Gao, Xiaoyu Li, Zhicheng Bao, Xiaoyan Feng, Jiaojiao Jiang

Abstract

Generative images have proliferated on Web platforms in social media and online copyright distribution scenarios, and semantic watermarking has increasingly been integrated into diffusion models to support reliable provenance tracking and forgery prevention for web content. Traditional noise-layer-based watermarking, however, remains vulnerable to inversion attacks that can recover embedded signals. To mitigate this, recent content-aware semantic watermarking schemes bind watermark signals to high-level image semantics, constraining local edits that would otherwise disrupt global coherence. Yet, large language models (LLMs) possess structured reasoning capabilities that enable targeted exploration of semantic spaces, allowing locally fine-grained but globally coherent semantic alterations that invalidate such bindings. To expose this overlooked vulnerability, we introduce a Coherence-Preserving Semantic Injection (CSI) attack that leverages LLM-guided semantic manipulation under embedding-space similarity constraints. This alignment enforces visual-semantic consistency while selectively perturbing watermark-relevant semantics, ultimately inducing detector misclassification. Extensive empirical results show that CSI consistently outperforms prevailing attack baselines against content-aware semantic watermarking, revealing a fundamental security weakness of current semantic watermark designs when confronted with LLM-driven semantic perturbations.

Metadata

arXiv ID: 2602.21593
Provider: ARXIV
Primary Category: cs.LG
Published: 2026-02-25
Fetched: 2026-02-26 05:00

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.21593v1</id>\n    <title>Breaking Semantic-Aware Watermarks via LLM-Guided Coherence-Preserving Semantic Injection</title>\n    <updated>2026-02-25T05:38:08Z</updated>\n    <link href='https://arxiv.org/abs/2602.21593v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.21593v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Generative images have proliferated on Web platforms in social media and online copyright distribution scenarios, and semantic watermarking has increasingly been integrated into diffusion models to support reliable provenance tracking and forgery prevention for web content. Traditional noise-layer-based watermarking, however, remains vulnerable to inversion attacks that can recover embedded signals. To mitigate this, recent content-aware semantic watermarking schemes bind watermark signals to high-level image semantics, constraining local edits that would otherwise disrupt global coherence. Yet, large language models (LLMs) possess structured reasoning capabilities that enable targeted exploration of semantic spaces, allowing locally fine-grained but globally coherent semantic alterations that invalidate such bindings. To expose this overlooked vulnerability, we introduce a Coherence-Preserving Semantic Injection (CSI) attack that leverages LLM-guided semantic manipulation under embedding-space similarity constraints. This alignment enforces visual-semantic consistency while selectively perturbing watermark-relevant semantics, ultimately inducing detector misclassification. Extensive empirical results show that CSI consistently outperforms prevailing attack baselines against content-aware semantic watermarking, revealing a fundamental security weakness of current semantic watermark designs when confronted with LLM-driven semantic perturbations.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CR'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <published>2026-02-25T05:38:08Z</published>\n    <arxiv:comment>Accepted by The Web Conference 2026 (Short Paper Track)</arxiv:comment>\n    <arxiv:primary_category term='cs.LG'/>\n    <author>\n      <name>Zheng Gao</name>\n    </author>\n    <author>\n      <name>Xiaoyu Li</name>\n    </author>\n    <author>\n      <name>Zhicheng Bao</name>\n    </author>\n    <author>\n      <name>Xiaoyan Feng</name>\n    </author>\n    <author>\n      <name>Jiaojiao Jiang</name>\n    </author>\n  </entry>"
}