arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.12285cs.CLcs.AI
收藏

从有偏聊天机器人到有偏代理：检验角色分配对LLM代理鲁棒性的影响

From Biased Chatbots to Biased Agents: Examining Role Assignment Effects on LLM Agent Robustness

Linbo Cao, Lihao Sun, Yang Yue

备注：
Accepted to the AAAI 2026 TrustAgent Workshop. 6 pages, 4 figures
阅读 PDF
arXiv 页面 ↗HTML 原文 ↗

AI总结：

研究发现基于人口统计学的人设分配会影响LLM代理的行为和性能，导致高达26.2%的降级，揭示了代理系统中隐含偏见和行为波动的风险。

摘要：

大型语言模型（LLMs）正被越来越多地部署为能够执行具有现实世界影响的自主代理，而不仅仅是文本生成。尽管在文本生成中由于角色引发的偏见已被广泛记录，但其对代理任务性能的影响仍鲜有研究，尽管这种影响可能带来更直接的操作风险。在本工作中，我们展示了首个系统性案例研究，证明基于人口统计学的人设分配可以改变LLM代理的行为并使其在不同领域中性能下降。评估广泛部署的模型在涵盖战略推理、规划和技术操作的代理基准上，我们发现显著的性能差异——高达26.2%的降级，由与任务无关的人设线索驱动。这些变化出现在各种任务类型和模型架构中，表明人设条件和简单的提示注入可以扭曲代理的决策可靠性。我们的发现揭示了当前LLM代理系统中被忽视的漏洞：人设分配可以引入隐含偏见并增加行为波动性，对LLM代理的安全和稳健部署提出了担忧。

英文摘要：

Large Language Models (LLMs) are increasingly deployed as autonomous agents capable of actions with real-world impacts beyond text generation. While persona-induced biases in text generation are well documented, their effects on agent task performance remain largely unexplored, even though such effects pose more direct operational risks. In this work, we present the first systematic case study showing that demographic-based persona assignments can alter LLM agents' behavior and degrade performance across diverse domains. Evaluating widely deployed models on agentic benchmarks spanning strategic reasoning, planning, and technical operations, we uncover substantial performance variations - up to 26.2% degradation, driven by task-irrelevant persona cues. These shifts appear across task types and model architectures, indicating that persona conditioning and simple prompt injections can distort an agent's decision-making reliability. Our findings reveal an overlooked vulnerability in current LLM agentic systems: persona assignments can introduce implicit biases and increase behavioral volatility, raising concerns for the safe and robust deployment of LLM agents.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

↑