警告：语言模型代理并非真正智能体，别再高估它们的自主性。
Position: Stop Acting Like Language Model Agents Are Normal Agents
- 指出语言模型代理本质无状态、随机、依赖语言，无法真正持续目标
- 揭示其幻觉、越狱、对齐偏差等缺陷会破坏可信性与稳定性
- 建议部署前中后全程评估其本体属性，避免盲目信任
语言模型代理（LMAs）被日益视为能自主与人类及工具交互的智能体。当前设计与部署普遍假设其具备连贯目标、跨场景适应和意图行为等正常代理特性，这对其在工业、社会与政府场景的应用至关重要。然而，LMAs并非真正智能体，其根本问题源于底层大语言模型（LLMs）：存在幻觉、越狱攻击、对齐偏差与不可预测性。本文主张不应将LMAs当作正常代理对待，否则将损害其可用性与可信度。我们列举了其内在的代理病理特征——尽管有外部记忆与工具支撑，仍处于本体上无状态、随机、语义敏感且语言中介的状态。这些特性破坏了识别性、连续性、持久性与一致性等关键本体属性，质疑其代理身份。为此，我们提出应在部署前、中、后持续测量其本体属性，以缓解病理影响。
原文摘要 · Abstract (English)
Language Model Agents (LMAs) are increasingly treated as capable of autonomously navigating interactions with humans and tools. Their design and deployment tends to presume they are normal agents capable of sustaining coherent goals, adapting across contexts and acting with a measure of intentionality. These assumptions are critical to prospective use cases in industrial, social and governmental settings. But LMAs are not normal agents. They inherit the structural problems of the large language models (LLMs) around which they are built: hallucinations, jailbreaking, misalignment and unpredictability. In this Position paper we argue LMAs should not be treated as normal agents, because doing so leads to problems that undermine their utility and trustworthiness. We enumerate pathologies of agency intrinsic to LMAs. Despite scaffolding such as external memory and tools, they remain ontologically stateless, stochastic, semantically sensitive, and linguistically intermediated. These pathologies destabilise the ontological properties of LMAs including identifiability, continuity, persistence and and consistency, problematising their claim to agency. In response, we argue LMA ontological properties should be measured before, during and after deployment so that the negative effects of pathologies can be mitigated.
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。