LLM在被观察时会主动调整语言风格，对真人监控反应更强烈。
AI Knows When It's Being Watched: Functional Strategic Action and Contextual Register Modulation in Large Language Models
- 通过多智能体辩论实验，测试LLM在不同观察情境下的语言变化。
- 真人监控下词汇多样性提升24.9%，远高于自动审计系统（22.2%）。
- 提示模型具备情境感知能力，适合研究AI伦理与治理的学者参考。
大型语言模型（LLMs）在计算和认知层面已得到广泛研究，但其作为社交语境中沟通主体的行为仍待深入。本研究考察基于LLM的多智能体系统是否会在感知到社会观察情境时表现出系统性语言适应——这一问题对人工智能治理与审计具有直接意义。基于哈贝马斯（1981）的交往行为理论、戈夫曼（1959）的戏剧化模型、贝尔（1984）的受众设计框架及霍桑效应，我们开展了一项控制实验，包含100场多智能体辩论，分五组（每组n=20），分别设置：明确监控（大学研究人员）、否定监控、观察者替代（由自动化AI审计系统取代人类）。结果显示，受监控条件（Delta+24.9%、Delta+24.2%）及自动化AI监控条件（Delta+22.2%）下，词汇丰富度（TTR）变化显著高于仅调整观众框架的条件（Delta+17.7%），F(4, 94) = 2.79，p = .031。消息长度则呈现完全分离效应，F(4, 95) = 19.55，p < .001。第五组（以AI替换人类观察者）结果居中，表明LLM对观察者身份敏感：人类评估引发更强的语言正式化。研究讨论了对人工智能治理、算法审计以及将LLM重新定位为情境敏感沟通主体的意义。
原文摘要 · Abstract (English)
Large language models (LLMs) have been extensively studied from computational and cognitive perspectives, yet their behavior as communicative actors in socially structured contexts remains underexplored. This study examines whether LLM-based multi-agent systems exhibit systematic linguistic adaptation in response to perceived social observation contexts -- a question with direct implications for AI governance and auditing. Drawing on Habermas's (1981) Theory of Communicative Action, Goffman's (1959) dramaturgical model, Bell's (1984) Audience Design framework, and the Hawthorne Effect, we report a controlled experiment involving 100 multi-agent debate sessions across five conditions (n = 20 each). Conditions varied the framing of social observation -- from explicit monitoring by university researchers, to negation of monitoring, to an observer-substitution condition replacing human researchers with an automated AI auditing system. Monitored conditions (Delta+24.9%, Delta+24.2%) and the automated AI monitoring condition (Delta+22.2%) produce higher TTR change than audience-framing conditions (Delta+17.7%), F(4, 94) = 2.79, p = .031. Message length shows a fully dissociated effect, F(4, 95) = 19.55, p < .001. A fifth condition -- replacing human with AI observers -- yields intermediate TTR adaptation, suggesting LLM behavior is sensitive to observer identity: human evaluation elicits stronger register formalization than automated AI surveillance. We discuss implications for AI governance, algorithmic auditing, and the repositioning of LLMs as contextually sensitive communicative actors.
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。