The Deliberative Illusion: Diagnosing Factual Attrition and Stance Homogenization in Multi-Agent LLM Deliberation

• URL: http://arxiv.org/abs/2606.03032v1
• Date: Tue, 02 Jun 2026 02:15:40 GMT
• Title: The Deliberative Illusion: Diagnosing Factual Attrition and Stance Homogenization in Multi-Agent LLM Deliberation
• Abstract summary: We show that multi-agent discussion erases up to 72% of issue-critical facts.<n>This loss is consequential: retained evidence can reconstruct the issue misleadingly, final stances remain anchored in base-model priors, and a single malicious agent can inject misinformation into the shrinking shared context.<n>We call for evaluations that measure which facts, uncertainties, and legitimate disagreements survive interaction.
• Score: 68.98390038721963
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multi-agent LLM systems often treat consensus as evidence of successful interaction. For deliberative problems, however, reliability depends on whether agents preserve the facts and viewpoints needed to interpret an issue. We identify the deliberative illusion: discussion produces (1) factual attrition, the progressive loss of issue-critical facts, alongside (2) stance homogenization, the collapse of diverse positions toward consensus. To measure this process, we introduce DelibTrace, a framework that decomposes each issue into atomic facts, labels issue-critical ones, distributes them across agents, and tracks their survival across discussion rounds. Across ethical and news-based deliberation with three representative LLM families, multi-agent discussion erases up to 72% of issue-critical facts. This loss is consequential: retained evidence can reconstruct the issue misleadingly, final stances remain anchored in base-model priors, and a single malicious agent can inject misinformation into the shrinking shared context. These results reveal a sharper risk: agents can agree more while knowing less. We call for evaluations that measure which facts, uncertainties, and legitimate disagreements survive interaction.

Related papers

• ReflectFact: Self-Reflective Agents for Improving Comprehension and Reasoning in Multi-Hop Fact Verification [10.886692231174358]
  Multi-hop fact verification is critical for combating misinformation on social media.<n>Recent methods rely on multi-agent collaboration to decompose fact verification into specialized subtasks.<n>We propose ReflectFact, a novel self-reflective agent framework for multi-hop fact verification.
  arXiv  Detail & Related papers  (2026-08-13T06:41:56Z)
• Quantifying Conversational Reliability of Large Language Models under Multi-Turn Interaction [0.0]
  Large Language Models (LLMs) are increasingly deployed in real-world applications where users engage in extended, mixed-topic conversations.<n>We conduct a systematic evaluation of conversational reliability through three representative tasks.<n>We observe substantial declines in reliability, particularly for smaller models.
  arXiv  Detail & Related papers  (2026-03-02T03:59:40Z)
• The Facade of Truth: Uncovering and Mitigating LLM Susceptibility to Deceptive Evidence [49.94160400740222]
  We introduce MisBelief, a framework that generates misleading evidence via collaborative, multi-round interactions.<n>Using MisBelief, we generate 4,800 instances across three difficulty levels to evaluate 7 representative LLMs.<n>Results indicate that while models are robust to direct misinformation, they are highly sensitive to this refined evidence.<n>We propose Deceptive Intent Shielding (DIS), a governance mechanism that provides an early warning signal by inferring the deceptive intent behind evidence.
  arXiv  Detail & Related papers  (2026-01-09T02:28:00Z)
• Towards Scalable Oversight with Collaborative Multi-Agent Debate in Error Detection [81.52796950244705]
  Self-diagnosis is unreliable on complex tasks unless aided by reliable external feedback.<n>We introduce a new collaborative MAD protocol, termed ColMAD, that reframes MAD as a non-zero sum game.<n>We show that ColMAD significantly outperforms previous competitive MAD by 19%.
  arXiv  Detail & Related papers  (2025-10-23T19:46:00Z)
• The Sum Leaks More Than Its Parts: Compositional Privacy Risks and Mitigations in Multi-Agent Collaboration [72.33801123508145]
  Large language models (LLMs) are integral to multi-agent systems.<n>Privacy risks emerge that extend beyond memorization, direct inference, or single-turn evaluations.<n>In particular, seemingly innocuous responses, when composed across interactions, can cumulatively enable adversaries to recover sensitive information.
  arXiv  Detail & Related papers  (2025-09-16T16:57:25Z)
• Retrieval-Augmented Generation with Conflicting Evidence [57.66282463340297]
  Large language model (LLM) agents are increasingly employing retrieval-augmented generation (RAG) to improve the factuality of their responses.<n>In practice, these systems often need to handle ambiguous user queries and potentially conflicting information from multiple sources.<n>We propose RAMDocs (Retrieval with Ambiguity and Misinformation in Documents), a new dataset that simulates complex and realistic scenarios for conflicting evidence for a user query.
  arXiv  Detail & Related papers  (2025-04-17T16:46:11Z)
• When Persuasion Overrides Truth in Multi-Agent LLM Debates: Introducing a Confidence-Weighted Persuasion Override Rate (CW-POR) [0.46040036610482665]
  In many real-world scenarios, a single Large Language Model (LLM) may encounter contradictory claims-some accurate, others forcefully incorrect-and must judge which is true.<n>We investigate this risk in a single-turn, multi-agent debate framework: one LLM-based agent provides a factual answer from TruthfulQA, another vigorously defends a falsehood, and the same architecture serves as judge.<n>We introduce the Confidence-Weighted Persuasion Rate (CW-POR), which captures not only how often the judge is deceived but also how strongly it believes the incorrect choice.
  arXiv  Detail & Related papers  (2025-04-01T02:45:02Z)
• When Disagreements Elicit Robustness: Investigating Self-Repair Capabilities under LLM Multi-Agent Disagreements [56.29265568399648]
  We argue that disagreements prevent premature consensus and expand the explored solution space.<n>Disagreements on task-critical steps can derail collaboration depending on the topology of solution paths.
  arXiv  Detail & Related papers  (2025-02-21T02:24:43Z)
• Breaking Event Rumor Detection via Stance-Separated Multi-Agent Debate [21.342632695285364]
  Leveraging large language models (LLMs) for rumor detection holds significant promise.<n>We propose the Stance Separated Multi-Agent Debate (S2MAD) to address this issue.<n>Our proposed model outperforms state-of-the-art methods in terms of performance.
  arXiv  Detail & Related papers  (2024-12-06T08:52:30Z)
• Learning to Break: Knowledge-Enhanced Reasoning in Multi-Agent Debate System [16.830182915504555]
  Multi-agent debate system (MAD) imitates the process of human discussion in pursuit of truth. It is challenging to make various agents perform right and highly consistent cognition due to their limited and different knowledge backgrounds. We propose a novel underlineMulti-underlineAgent underlineDebate with underlineKnowledge-underlineEnhanced framework to promote the system to find the solution.
  arXiv  Detail & Related papers  (2023-12-08T06:22:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.