Unsafe by Reciprocity: How Generation-Understanding Coupling Undermines Safety in Unified Multimodal Models

• URL: http://arxiv.org/abs/2603.27332v1
• Date: Sat, 28 Mar 2026 16:28:39 GMT
• Title: Unsafe by Reciprocity: How Generation-Understanding Coupling Undermines Safety in Unified Multimodal Models
• Abstract summary: We investigate whether cross-functionality reciprocity itself constitutes a structural source of vulnerability in Unified Multimodal Models (UMMs)<n>We propose RICE: Reciprocal Interaction-based Cross-functionality Exploitation, a novel attack paradigm that explicitly exploits bidirectional interactions between understanding and generation.
• Score: 53.41258113970795
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent advances in Large Language Models (LLMs) and Text-to-Image (T2I) models have led to the emergence of Unified Multimodal Models (UMMs), where multimodal understanding and image generation are tightly integrated within a shared architecture. Prior studies suggest that such reciprocity enhances cross-functionality performance through shared representations and joint optimization. However, the safety implications of this tight coupling remain largely unexplored, as existing safety research predominantly analyzes understanding and generation functionalities in isolation. In this work, we investigate whether cross-functionality reciprocity itself constitutes a structural source of vulnerability in UMMs. We propose RICE: Reciprocal Interaction-based Cross-functionality Exploitation, a novel attack paradigm that explicitly exploits bidirectional interactions between understanding and generation. Using this framework, we systematically evaluate Generation-to-Understanding (G-U) and Understanding-to-Generation (U-G) attack pathways, demonstrating that unsafe intermediate signals can propagate across modalities and amplify safety risks. Extensive experiments show high Attack Success Rates (ASR) in both directions, revealing previously overlooked safety weaknesses inherent to UMMs.

Related papers

• Unleashing Multimodal Large Language Models for Training-free HOI Detection in the Wild [49.77540427384148]
  We present AgentHOI, a training-free, agentic framework that transfers the generalist multimodal reasoning capabilities of foundation models to HOI detection in the wild.<n>To address the challenges of incomplete interaction discovery and ambiguous localization in complex scenes, we introduce two key mechanisms.<n>AgentHOI achieves superior performance over state-of-the-art supervised and weakly supervised methods in real-world settings.
  arXiv  Detail & Related papers  (2026-07-15T14:30:20Z)
• UMSS: Towards Unsupervised Multi-modal Semantic Segmentation [6.785303759588842]
  We make the first attempt to address the problem of Unsupervised Multimodal Semantic (UMSS)<n>We propose UniM2, a novel framework built on DINOv3 that transforms conventional fusion methods into consistent performance gains.<n>Our key idea is to learn a unified latent space driven by Cross Modal Correspondence Synergy (CMCS) to extract intrinsic shared semantic cues.
  arXiv  Detail & Related papers  (2026-07-14T05:44:29Z)
• Beyond Surface Artifacts: Capturing Shared Latent Forgery Knowledge Across Modalities [52.189426539304065]
  This paper introduces a paradigm shift that redefines multimodal forensics from conventional "feature fusion" to "modality generalization"<n>We propose the first modality-agnostic forgery (MAF) detection framework. By explicitly decoupling modality-specific styles, MAF precisely extracts the essential, cross-modal latent forgery knowledge.
  arXiv  Detail & Related papers  (2026-04-09T03:35:21Z)
• When Safe Models Merge into Danger: Exploiting Latent Vulnerabilities in LLM Fusion [15.004295056225002]
  Model merging has emerged as a powerful technique for combining specialized capabilities from multiple fine-tuned LLMs without additional training costs.<n>We present TrojanMerge, a framework that embeds latent malicious components into source models that remain individually benign but produce severely misaligned models when merged.
  arXiv  Detail & Related papers  (2026-04-01T08:32:46Z)
• Diagnosing and Repairing Unsafe Channels in Vision-Language Models via Causal Discovery and Dual-Modal Safety Subspace Projection [21.411964269762432]
  We present a comprehensive framework for diagnosing and repairing unsafe channels within Large Vision-Language Models (LVLMs)<n>We first perform causal mediation analysis to identify neurons and layers that are causally responsible for unsafe behaviors.<n>We introduce a dual-modal safety subspace projection method that learns generalized safety subspaces for both visual and textual modalities.
  arXiv  Detail & Related papers  (2026-03-28T11:31:16Z)
• Quantifying the Gap between Understanding and Generation within Unified Multimodal Models [66.07644743841007]
  GapEval is a benchmark designed to quantify the gap between understanding and generation capabilities.<n>Experiments reveal a persistent gap between the two directions across a wide range of UMMs.<n>Our findings indicate that knowledge within UMMs often remains disjoint.
  arXiv  Detail & Related papers  (2026-02-02T14:19:37Z)
• Beyond Single-Agent Safety: A Taxonomy of Risks in LLM-to-LLM Interactions [0.0]
  This paper examines why safety mechanisms designed for human-model interaction do not scale to environments where large language models interact with each other.<n>We propose a conceptual transition from model-level safety to system-level safety, introducing the framework of the Emergent Systemic Risk Horizon (ESRH)<n>The paper contributes (i) a theoretical account of collective risk in interacting LLMs, (ii) a taxonomy connecting micro, meso, and macro-level failure modes, and (iii) a design proposal for InstitutionalAI, an architecture for embedding adaptive oversight within multi-agent systems.
  arXiv  Detail & Related papers  (2025-12-02T12:06:57Z)
• Uni-MMMU: A Massive Multi-discipline Multimodal Unified Benchmark [69.8473923357969]
  Unified multimodal models aim to jointly enable visual understanding and generation, yet current benchmarks rarely examine their true integration.<n>We present Uni-MMMU, a comprehensive benchmark that unfolds the bidirectional synergy between generation and understanding across eight reasoning-centric domains.
  arXiv  Detail & Related papers  (2025-10-15T17:10:35Z)
• COSMO-RL: Towards Trustworthy LMRMs via Joint Safety and Stability [101.80200069234377]
  We present COSMO-RL, a mixed reinforcement learning framework that trains LMRMs under multimodal, multitask, and multiobjective signals.<n>Our approach aims to let safety and capability grow together in one stable pipeline rather than competing during alignment.
  arXiv  Detail & Related papers  (2025-10-05T13:30:03Z)
• Co-AttenDWG: Co-Attentive Dimension-Wise Gating and Expert Fusion for Multi-Modal Offensive Content Detection [0.0]
  Multi-modal learning has emerged as a crucial research direction.<n>Existing approaches often suffer from insufficient cross-modal interactions and rigid fusion strategies.<n>We propose Co-AttenDWG, co-attention with dimension-wise gating, and expert fusion.<n>We show that Co-AttenDWG achieves state-of-the-art performance and superior cross-modal alignment.
  arXiv  Detail & Related papers  (2025-05-25T07:26:00Z)
• Safety Alignment Can Be Not Superficial With Explicit Safety Signals [8.297367440457508]
  Recent studies on the safety alignment of large language models (LLMs) have revealed that existing approaches often operate superficially.<n>This paper identifies a fundamental cause of this superficiality: existing alignment approaches presume that models can implicitly learn a safety-related reasoning task during the alignment process.<n>By explicitly introducing a safety-related binary classification task and integrating its signals with our attention and decoding strategies, we eliminate this ambiguity.
  arXiv  Detail & Related papers  (2025-05-19T20:40:46Z)
• SoK: The Security-Safety Continuum of Multimodal Foundation Models through Information Flow and Game-Theoretic Defenses [58.93030774141753]
  Multimodal foundation models (MFMs) integrate diverse data modalities to support complex and wide-ranging tasks.<n>In this paper, we unify the concepts of safety and security in the context of MFMs by identifying critical threats that arise from both model behavior and system-level interactions.
  arXiv  Detail & Related papers  (2024-11-17T23:06:20Z)
• Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment [103.05005690990271]
  Mixture of insighTful Experts (MoTE) is a novel framework that combines reasoning chains and expert mixtures to improve self-alignments.<n>MoTE significantly improves model safety, jailbreak resistance, and over-refusal capabilities, achieving performance comparable to OpenAI's state-of-the-art o1 model.
  arXiv  Detail & Related papers  (2024-05-01T15:06:05Z)

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.