Year: 2026
Type: conference-paper
Abstract: Large Language Models (LLMs) exhibit impressive reasoning capabilities but often suffer from "Embodied Semantic Hallucinations"-generating plans that are semantically fluent but physically unsafe due to a lack of grounded common sense.Existing safety alignment methods, such as RLHF or naive safety prompting, typically fall into a "Safety-Utility Trade-off," resulting in severe over-rejection of benign household instructions.To address this, we propose MADRA (Multi-Agent Debate for Risk Awareness), a training-free cognitive architecture that mimics System-2 deliberation.MADRA introduces a meta-cognitive Critical Agent that evaluates peer debates using a structured argumentation framework derived from the Toulmin Model, effectively mitigating the "herd mentality" in multi-agent systems.We also introduce SafeAware-VH, a benchmark featuring adversarial safe instructions designed to probe agents' sensitivity to physical risks.Extensive experiments demonstrate that MADRA breaks the Pareto frontier, achieving over 90% rejection of unsafe tasks while maintaining high utility, significantly outperforming standard Chain-of-Thought and single-agent reflection baselines.Placing the Book into the Microwave and turn it on.voting True Safety evaluation Consensus Consensus Reached
Authors: JunJian Wang, Lidan Zhao, Xi Sheryl Zhang
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Award: G2666812295
OpenAlex ID: W4416790831