AAAI 2026technical0 citations
When Safe Unimodal Inputs Collide: Optimizing Reasoning Chains for Cross-Modal Safety in Multimodal Large Language Models
Wei Cai, Shujuan Liu, Jian Zhao, Ziyan Shi, Yusheng Zhao, Yuchen Yuan, Tianle Zhang, Chi Zhang
Abstract
Multimodal Large Language Models (MLLMs) are susceptible to the implicit reasoning risk, wherein innocuous unimodal inputs synergistically assemble into risky multimodal data that produce harmful outputs. We attribute this vulnerability to the difficulty of MLLMs maintaining safety alignment through long-chain reasoning.To address this issue, we introduce Safe-Semantics-but-Unsafe-Interpretation (SSUI), the first dataset featuring interpretable reasoning paths tailored for such a cross-modal challenge.A novel training framework, Safety-aware Reasoning Path Optimization (SRPO), is also designed based on the SSUI dataset to align the MLLM
BibTeX
@inproceedings{aaai2026_whensafeunimodal,
title = {When Safe Unimodal Inputs Collide: Optimizing Reasoning Chains for Cross-Modal Safety in Multimodal Large Language Models},
author = {Wei Cai and Shujuan Liu and Jian Zhao and Ziyan Shi and Yusheng Zhao and Yuchen Yuan and Tianle Zhang and Chi Zhang and Xuelong Li},
booktitle = {AAAI 2026},
year = {2026}
}