一作ACL 2026OralOutstanding Paper· Natural Language Processing
Lying with Truths: Open-Channel Multi-Agent Collusion for Belief Manipulation via Generative Montage
Jinwei Hu · Xinmiao Huang · Youcheng Sun · Yi Dong · Xiaowei Huang
摘要
As large language models (LLMs) transition to autonomous agents synthesizing real-time information, their reasoning capabilities introduce an unexpected attack surface. This paper introduces a novel threat where colluding agents steer victim beliefs using only truthful evidence fragments distribu...