arXiv 2606.058902026-06-04cs.CLcs.AI

与不确定性共处：LLM对LLM模拟对话中人工道德顾问的不确定性支撑策略

Staying with the Uncertainty: Uncertainty-Scaffolding Strategies for Artificial Moral Advisors in LLM-to-LLM Simulated Conversations

Centre for Data Futures, The Dickson Poon School of Law, King’s College London（数据未来中心、迪克森·普恩法学院、伦敦国王学院）；Department of Informatics, King’s College London（信息学院、伦敦国王学院）；LangAI, Center for Language AI Research, Tohoku University（LangAI、语言人工智能研究中心、东北大学）；Neukom Institute for Computational Science, Dartmouth College（计算科学尼科姆研究所、达特茅斯学院）

Salvatore Greco, Hainiu Xu, Jacopo Domenicucci, Yulan He, Sylvie Delacroix

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

研究LLM作为人工道德顾问时，通过三种不确定性策略（视角倍增、张力保持、过程反思）与三种控制条件对比，在模拟对话中探讨如何帮助对话者“与不确定性共处”，发现不同策略在立场改变量上无差异但影响参与质量。

摘要：

LLM越来越多地被部署为各种背景下的人工道德顾问（AMA）：它们应该展现什么样的对话模式？在本文中，我们研究AMA如何帮助其对话者“与不确定性共处”。我们提出了三种不确定性模式（视角倍增、张力保持、过程反思），并将它们与三种控制条件（基线、说服、谄媚）进行比较。用户代理LLM与遵循特定不确定性策略的AMA就伦理困境进行对话，并完成对话前和对话后的问卷调查。我们进一步考察了两种角色提示格式（陈述式和叙述式）的效果。我们发现：（1）没有一个单一模型作为模拟用户代理占主导地位，开放模型通过角色间分歧与人类模糊性对齐，而封闭模型通过角色内对冲对齐；（2）陈述式角色更好地捕捉初始立场多样性，而叙述式角色显示出更现实的信念修正；（3）所有六种AMA策略产生可区分的对话模式；（4）不确定性策略的不同不在于它们产生多少立场改变，而在于它们维持的参与质量。

英文摘要：

LLMs are increasingly deployed as Artificial Moral Advisors (AMA) in a variety of contexts: what kind of conversational patterns should they display? In this paper, we study how AMA can help their interlocutors "stay with the uncertainty". We propose three modes of uncertainty (Perspective-Multiplying, Tension-Preserving, Process-Reflecting) and compare them against three control conditions (Baseline, Persuasive, Sycophantic). A user-agent LLM engages in a dialogue on an ethical dilemma with an AMA following a specific uncertainty strategy, and completes pre- and post-conversation questionnaires. We further examine the effect of two persona prompt formats (Declarative and Narrative). We found that (1) no single model dominates as a simulated user agent, with open models aligning with human ambiguity through between-persona divergence and closed models through within-persona hedging; (2) declarative personas better capture initial stance diversity while narrative personas show more realistic belief revision; (3) all six AMA strategies produce distinguishable conversational patterns; and (4) uncertainty strategies differ not in how much stance revision they produce, but in the quality of engagement they sustain.

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。