# Exploring Health Misinformation Detection with Multi-Agent Debate

arXiv: https://arxiv.org/abs/2512.09935 · HTML: https://mlsanity.com/paper/2512.09935/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 14 references that are in the map.
- First posted on arXiv 29 Nov 2025.

## Builds on (most cited first, 9 of 9)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate](https://mlsanity.com/paper/2305.19118/) (2023)
- [Fact or Fiction: Verifying Scientific Claims](https://mlsanity.com/paper/2004.14974/) (2020)
- [HealthFC: Verifying Health Claims with Evidence-Based Medical Fact-Checking](https://mlsanity.com/paper/2309.08503/) (2023)
- [The Truth Becomes Clearer Through Debate! Multi-Agent Systems with Large Language Models Unmask Fake News](https://mlsanity.com/paper/2505.08532/) (2025)
- [Removal of Hallucination on Hallucination: Debate-Augmented RAG](https://mlsanity.com/paper/2505.18581/) (2025)
- [Step-by-Step Fact Verification System for Medical Claims with Explainable Reasoning](https://mlsanity.com/paper/2502.14765/) (2025)
- [Web Retrieval Agents for Evidence-Based Misinformation Detection](https://mlsanity.com/paper/2409.00009/) (2024)
- [EMULATE: A Multi-Agent Framework for Determining the Veracity of Atomic Claims by Emulating Human Actions](https://mlsanity.com/paper/2505.16576/) (2025)

## Abstract (arXiv)

Fact-checking health-related claims has become increasingly critical as misinformation proliferates online. Effective verification requires both the retrieval of high-quality evidence and rigorous reasoning processes. In this paper, we propose a two-stage framework for health misinformation detection: Agreement Score Prediction followed by Multi-Agent Debate. In the first stage, we employ large language models (LLMs) to independently evaluate retrieved articles and compute an aggregated agreement score that reflects the overall evidence stance. When this score indicates insufficient consensus-falling below a predefined threshold-the system proceeds to a second stage. Multiple agents engage in structured debate to synthesize conflicting evidence and generate well-reasoned verdicts with explicit justifications. Experimental results demonstrate that our two-stage approach achieves superior performance compared to baseline methods, highlighting the value of combining automated scoring with collaborative reasoning for complex verification tasks.