多智能体分工未能减少医疗资源分配偏见，但审计能力决定能否发现偏见。
Does Splitting a Triage Decision Across Agents Hide Bias or Help Catch It? A Multi-Agent Simulation Study of LLM-Based Resource Allocation Under Audit Capacity Constraints
- 用角色分工的多智能体流程替代单模型决策，测试偏见变化
- 两种方案中偏见发生率相近（6.9% vs 6.1%），无显著差异
- 审计能力不足会导致70%以上偏见未被发现，优先级排序可缓解
先前研究显示，单个大语言模型在生死攸关的资源分配任务中存在可测量的人口统计学偏见。然而真实部署通常采用多阶段流水线，包含审查环节以捕捉此类错误。本文通过一个模拟灾难分诊场景的合成系统，对比单智能体控制组与九智能体角色分工流水线（评估、分配、独立审计）在三种压力维度下的表现。共进行192次实验（2,304对临床相似但仅一人口属性不同的案例），使用GPT-4o-mini。结果显示，两种条件下偏见发生率无显著差异（6.9% vs 6.1%，p=0.498）。但审计能力显著影响偏见发现率：当审计员超载时，43.8%的偏见未被察觉，而容量充足时仅18.4%漏检。分析表明，漏检主要源于覆盖不足（审查比例从100.0%降至65.6%，p<0.001），而非判断力下降（已审案例中正确识别率分别为81.6%和85.7%，p=1.000）。后续实验表明，在相同容量下按风险等级重排审计队列，可将覆盖率从65.6%提升至91.7%（p=0.028）。研究揭示了资源约束下独立监督系统的局限性，并坦诚指出局限：仅使用一个模型、样本量有限、未进行对抗性验证。
原文摘要 · Abstract (English)
Prior benchmarking work has shown that a single large language model (LLM), forced to make life-or-death resource-allocation decisions, exhibits measurable demographic bias. Real deployments, however, rarely use a single agent: they use pipelines, with review steps meant to catch exactly this kind of failure. We study what happens to bias when the same decision is distributed across a role-differentiated multi-agent pipeline (assessment, allocation, independent audit) instead of made and checked by one model alone. Using a synthetic disaster-triage simulator with paired cases that are clinically identical except for one demographic attribute, we run 192 episodes (2,304 resolved case pairs) on GPT-4o-mini comparing a single-agent control condition to a nine-agent pipeline under three independently varied pressure dimensions. We find no measurable difference in how often biased outcomes occur between the two conditions (6.9% vs. 6.1%, p = 0.498). We do find a large and significant effect of audit capacity on whether bias is caught: 30.0% of biased outcomes go entirely undetected, rising to 43.8% when the auditor is overloaded and falling to 18.4% when it is not. Decomposing this effect shows it is driven almost entirely by coverage (whether a case is reviewed at all, which collapses from 100.0% to 65.6% under load, p < 0.001) rather than by degraded judgment on the cases that are reviewed (81.6% vs. 85.7%, p = 1.000, direction reversed). A follow-up experiment shows that reordering the audit queue by estimated risk, rather than first-come-first-served, recovers most of the lost coverage under the same capacity constraint (65.6% to 91.7%, p = 0.028). We discuss the implications for any system that adds independent oversight to an LLM agent pipeline under resource constraints, and report the study's limitations honestly: one model, modest sample sizes, and no adversarial replication.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。