mlsanityData to 7 Oct 2026

Survey of Adversarial Robustness in Multimodal Large Language Models

arXiv 2503.13962 ↗First posted 18 Mar 2025cs.CVcited by 14

On the map: Multimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness

Abstract (arXiv)

Multimodal Large Language Models (MLLMs) have demonstrated exceptional performance in artificial intelligence by facilitating integrated understanding across diverse modalities, including text, images, video, audio, and speech. However, their deployment in real-world applications raises significant concerns about adversarial vulnerabilities that could compromise their safety and reliability. Unlike unimodal models, MLLMs face unique challenges due to the interdependencies among modalities, making them susceptible to modality-specific threats and cross-modal adversarial manipulations. This paper reviews the adversarial robustness of MLLMs, covering different modalities. We begin with an overview of MLLMs and a taxonomy of adversarial attacks tailored to each modality. Next, we review key datasets and evaluation metrics used to assess the robustness of MLLMs. After that, we provide an in-depth review of attacks targeting MLLMs across different modalities. Our survey also identifies critical challenges and suggests promising future research directions.

Key facts. Cited by 14 papers in the map (data through 7 Oct 2026). Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 52 of its 61 references that are in the map. First posted on arXiv 18 Mar 2025.

Builds on

52 of 61 references.

Closest ancestors · 24

Same direction, newest first

  1. 2024Visual Adversarial Attack on Vision-Language Models for Autonomous Driving27 Nov
  2. 2024A Survey of Attacks on Large Vision-Language Models: Resources, Advances, and Future Trends10 Jul
  3. 2024White-box Multimodal Jailbreaks Against Large Vision-Language Models28 May
  4. 2024SpeechGuard: Exploring the Adversarial Robustness of Multimodal Large Language Models14 May
  5. 2024Efficient Generation of Targeted and Transferable Adversarial Examples for Vision-Language Models Via Diffusion Models16 Apr
  6. 2024Unbridled Icarus: A Survey of the Potential Perils of Image Inputs in Multimodal Large Language Model Security8 Apr
  7. 2024Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models14 Mar
  8. 2024B-AVIBench: Towards Evaluating the Robustness of Large Vision-Language Model on Black-box Adversarial Visual-Instructions14 Mar
  9. 2024ImgTrojan: Jailbreaking Vision-Language Models with ONE Image5 Mar
  10. 2024VL-Trojan: Multimodal Instruction Backdoor Attacks against Autoregressive Visual Language Models21 Feb
  11. 2024Robust CLIP: Unsupervised Adversarial Fine-Tuning of Vision Embeddings for Robust Large Vision-Language Models19 Feb
  12. 2024Test-Time Backdoor Attacks on Multimodal Large Language Models13 Feb
12 more
  1. 2024Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast13 Feb
  2. 2024Shadowcast: Stealthy Data Poisoning Attacks Against Vision-Language Models5 Feb
  3. 2024Safety of Multimodal Large Language Models on Images and Texts1 Feb
  4. 2024Inducing High Energy-Latency of Large Vision-Language Models with Verbose Images20 Jan
  5. 2024MLLM-Protector: Ensuring MLLM's Safety without Hurting Performance5 Jan
  6. 2023On the Robustness of Large Multimodal Models Against Image Adversarial Attacks6 Dec
  7. 2023MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models29 Nov
  8. 2023FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts9 Nov
  9. 2023How Robust is Google's Bard to Adversarial Image Attacks?21 Sep
  10. 2023Abusing Images and Sounds for Indirect Instruction Injection in Multi-Modal LLMs19 Jul
  11. 2023Visual Adversarial Examples Jailbreak Aligned Large Language Models22 Jun
  12. 2023On Evaluating Adversarial Robustness of Large Vision-Language Models26 May

Other references · 28

  1. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  3. 2016Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answeringcited by 2,498
  4. 2022Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answeringcited by 1,881
  5. 2015Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Modelscited by 1,352
  6. 2019OK-VQA: A Visual Question Answering Benchmark Requiring External Knowledgecited by 1,114
  7. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!cited by 1,106
  8. 2023Vision-Language Models for Vision Tasks: A Surveycited by 834
  9. 2023SafetyBench: Evaluating the Safety of Large Language Modelscited by 189
  10. 2023Image Hijacks: Adversarial Images can Control Generative Models at Runtimecited by 165
  11. 2023On the Adversarial Robustness of Multi-Modal Foundation Modelscited by 142
  12. 2024LLaVA-Phi: Efficient Multi-Modal Assistant with Small Language Modelcited by 128
16 more
  1. 2024Jailbreaking Attack against Multimodal Large Language Modelcited by 128
  2. 2024Dissecting Adversarial Robustness of Multimodal LM Agentscited by 118
  3. 2023AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Datasetcited by 79
  4. 2024T2VSafetyBench: Evaluating the Safety of Text-to-Video Generative Modelscited by 66
  5. 2023Jailbreaking GPT-4V via Self-Adversarial Attacks with System Promptscited by 62
  6. 2022Grounding Answers for Visual Questions Asked by Visually Impaired Peoplecited by 47
  7. 2024Breaking Down the Defenses: A Comparative Survey of Attacks on Large Language Modelscited by 42
  8. 2024Voice Jailbreak Attacks Against GPT-4ocited by 39
  9. 2024Physical Backdoor Attack can Jeopardize Driving with Vision-Large-Language Modelscited by 30
  10. 2024VideoQA in the Era of LLMs: An Empirical Studycited by 29
  11. 2024Towards Transferable Attacks Against Vision-LLMs in Autonomous Driving with Typographycited by 17
  12. 2024FMM-Attack: A Flow-based Multi-modal Adversarial Attack on Video-based LLMscited by 14
  13. 2024LlamaPartialSpoof: An LLM-Driven Fake Speech Dataset Simulating Disinformation Generationcited by 14
  14. 2024TrustNavGPT: Modeling Uncertainty to Improve Trustworthiness of Audio-Guided LLM-Based Robot Navigationcited by 11
  15. 2024Beyond Raw Videos: Understanding Edited Videos with Large Multimodal Modelcited by 10
  16. 2024PG-Attack: A Precision-Guided Adversarial Attack Framework Against Vision Foundation Models for Autonomous Drivingcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 10; 2026: 4.

14 papers.

Sort:

By area:

  1. 2025MLLMs are Deeply Affected by Modality Biascited by 38
  2. 2025A Survey on Autonomy-Induced Security Risks in Large Model-Based Agentscited by 28
  3. 2025Flattery in Motion: Benchmarking and Analyzing Sycophancy in Video-LLMscited by 20
  4. 2025MVI-Bench: A Comprehensive Benchmark for Evaluating Robustness to Misleading Visual Inputs in LVLMscited by 14
  5. 2025Robustness in Large Language Models: A Survey of Mitigation Strategies and Evaluation Metricscited by 12
  6. 2025StableToken: A Noise-Robust Semantic Speech Tokenizer for Resilient SpeechLLMscited by 4
  7. 2025Contextual Image Attack: How Visual Context Exposes Multimodal Safety Vulnerabilitiescited by 2
  8. 2026RemoteShield: Enable Robust Multimodal Large Language Models for Earth Observationcited by 2
  9. 2025Robust Diagram Reasoning: A Framework for Enhancing LVLM Performance on Visually Perturbed Scientific Diagramscited by 1
  10. 2025LLaVAShield: Safeguarding Multimodal Multi-Turn Dialogues in Vision-Language Modelscited by 1
  11. 2026Universal Adversarial Attacks against Closed-Source MLLMs via Target-View Routed Meta Optimizationcited by 1
  12. 2025Adversarial Robustness in Zero-Shot Learning:An Empirical Study on Class and Concept-Level Vulnerabilitiescited by 0
2 more
  1. 2026Beyond Visual Safety: Jailbreaking Multimodal Large Language Models for Harmful Image Generation via Semantic-Agnostic Inputscited by 0
  2. 2026POPS: Recovering Unlearned Multi-Modality Knowledge in MLLMs with Prompt-Optimized Parameter Shakingcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Adversarial Attacks on Multimodal Large Language Models: A Comprehensive Survey
  2. 2026Investigating Adversarial Robustness of Multi-modal Large Language Models
  3. 2025Investigating Vulnerabilities and Defenses Against Audio-Visual Attacks: A Comprehensive Survey Emphasizing Multimodal Models
  4. 2025Transferability of Adversarial Attacks in Video-based MLLMs: A Cross-modal Image-to-Video Approach
  5. 2025Pushing the Limits of Safety: A Technical Report on the ATLAS Challenge 2025
  6. 2025Multifaceted Evaluation of Audio-Visual Capability for MLLMs: Effectiveness, Efficiency, Generalizability and Robustness
  7. 2025Towards Robust Multimodal Large Language Models Against Jailbreak Attacks
  8. 2026Forbid Your Attention: Fooling Multimodal Large Language Models by Selectively Removing Intrinsic Focus in Spectral Domain