Research
MM-PoisonRAG: Disrupting Multimodal RAG with Local and Global Knowledge Poisoning Attacks
Publication date not yet evaluated · Added · 3 reports aclanthology.org
Page published
Coverage timeline
Why it matters
MM-PoisonRAG demonstrates that multimodal RAG pipelines can be systematically poisoned, giving defenders concrete evidence of an attack surface in retrieval-augmented MLLM deployments.
MM-PoisonRAG is a research paper presenting local and global knowledge poisoning attacks against multimodal retrieval-augmented generation (RAG) systems used by multimodal large language models (MLLMs). The authors show that MLLMs' reliance on retrieval exposes them to poisoned knowledge that can disrupt factual grounding and manipulate outputs.