Vis-Poison is a novel attack that poisons multimodal retrieval-augmented generation systems by inserting attacker-controlled images as visual evidence, without altering any textual metadata. The attack uses an automated multi-agent approach to create visually plausible poisoned images and has been tested on two multimodal RAG pipelines, four embedding models, and six generation models. In black-box settings, Vis-Poison achieves an end-to-end success rate between 40.16% and 65.40% against 30,000-entry knowledge bases, and remains effective against various multimodal large language models with an average success rate above 60%.
arXiv:2511.18921v2 Announce Type: replace Abstract: Backdoor attacks undermine the reliability and trustworthiness of machine learning systems by injecting hidden behaviors that can be maliciously ac...
arXiv:2604. 05809v2 Announce Type: replace-cross Abstract: This paper presents Text-Guided Backdoor (TGB), an adjustable backdoor attack against multimodal pretrained models that uses natural-word triggers, namely words that can naturally occur in ordinary textual inputs.
arXiv:2606. 10742v1 Announce Type: cross Abstract: External memory has become a core component of modern web agents, enabling long-horizon reasoning through the retrieval of past experiences.
arXiv:2603. 29418v2 Announce Type: replace-cross Abstract: Although multimodal large language models (MLLMs) are increasingly deployed in real-world applications, their instruction-following behavior leaves them vulnerable to prompt injection attacks.
arXiv:2607. 15657v1 Announce Type: cross Abstract: Multimodal AI agents increasingly rely on persistent long-term memory to ground generation in past visual and textual episodes.
The paper introduces COMIC, a reference‑aware safety gate designed for multimodal large language models (MLLMs). COMIC detects the operation requested by a user, identifies visual targets through OCR and open‑vocabulary proposals, and evaluates safety on explicit operation‑target pairs, using max‑risk aggregation and quality‑aware routing to decide whether to allow or block a request. Experiments on several open‑source MLLMs and jailbreak benchmarks show that COMIC improves robustness while maintaining benign utility and efficiency.
arXiv:2609.29099v1 Announce Type: cross Abstract: Multimodal training relies on image-text corpora collected from external sources, creating opportunities for attackers to poison the data. Stealthy a...
The paper investigates how Vision Language Models (VLMs) can be fooled by tiny, human‑imperceptible changes to images. It introduces a gradient‑based attack that targets only the vision encoder, reducing computational cost while still effectively disrupting both untargeted and targeted multimodal interpretations. Experiments on open‑source VLMs such as Qwen2.5‑VL, Granite‑Vision, FastVLM, and Phi‑3.5‑Vision demonstrate that these small perturbations can dramatically alter the models’ textual outputs.
The paper investigates how Vision Language Models (VLMs) can be fooled by small, human‑imperceptible changes to images. It introduces a gradient‑based attack that targets only the vision encoder, reducing computational cost while still effectively disrupting both untargeted and targeted multimodal alignment. Experiments on open‑source VLMs such as Qwen2.5‑VL, Granite‑Vision, FastVLM, and Phi‑3.5‑Vision demonstrate that these perturbations can significantly alter the models’ textual outputs.
arXiv:2601. 12263v2 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) integrate visual and textual knowledge into unified representations that increasingly underpin modern retrieval and recommendation systems.
The paper introduces MemJack, a memory‑augmented multi‑agent framework that automatically generates jailbreak attacks on Vision‑Language Models (VLMs) using benign natural images as visual anchors. MemJack discovers visual anchors, camouflages them semantically, evaluates responses, repairs via reflection, and replans dynamically, forming a closed‑loop attack pipeline. The authors also create MemJack‑Bench, a dataset of over 113,000 interactive multimodal jailbreak trajectories, and show that MemJack achieves a 71.48% attack success rate against Qwen3‑VL‑Plus, reaching 90% under extended budgets, outperforming other baselines on natural‑image evaluation.
Text-to-image diffusion models have achieved high visual fidelity and broad adoption, but remain vulnerable to safety violations when adversaries exploit them to synthesize illicit content. Existing alignment paradigms, from input sanitization to structural feature pruning, are largely organized around unsafe concepts explicitly exposed during filtering, editing, or localization.