arXiv cs.CY· Hanhui Xu, Jiacheng Ji, Yuhan Wang, Haoan Jin, Jaiming Luo, Mengyue Wu·· 13 小时前AI 评分29
拟人化 AI 交互的伦理治理：以“脆弱性情境”替代“脆弱群体”
Situating Vulnerability: An Ethical Governance Framework for Anthropomorphic AI Interaction
AI 导读
针对拟人化 AI 交互，论文提出伦理治理单元应是“脆弱性情境”而非“脆弱群体”，并划分身体、情感、认知、自主性和关系五类情境、给出七项原则。作者认为统一限制把伤害错置于回复措辞而非关系轨迹，群体定向保护则依赖永远无法完整的名单、把流动属性当作人的静态标签。该框架将 AI 安全从输出控制转向关系性安全。
正文 · 原文
Abstract:Large language models increasingly address users in the register of care: they remember personal details, adopt a first-person emotional voice, and sustain relationships across many sessions. This anthropomorphic mode of interaction generates risks that output-level metrics capture poorly, because the same empathic utterance may console one user, entrap a second, and consolidate a delusion in a third. Governance has responded in two ways. Uniform restriction refuses sensitive requests or suppresses empathic phrasing across all outputs; group-targeted protection profiles vulnerable populations and applies stricter or prohibitive rules to them. This article argues that both fail. Uniform restriction mislocates harm in the lexical form of responses rather than in the trajectory of a relationship; group-based protection depends on a list that can never be complete, treats a fluid property as a static attribute of persons, and converts protective attention into stigmatizing labels. Drawing on philosophical work that treats vulnerability as universal, relational, and differentiated, we propose that the unit of ethical governance should be the vulnerability situation rather than the vulnerable group, develop a typology of five situations - physical, emotional, cognitive, autonomy, and relational - and specify seven principles into requirements that guide response calibration without diagnosing or persistently classifying users. The framework relocates AI safety from output control to relational safety.
| Subjects: | Computers and Society (cs.CY) |
| Cite as: | arXiv:2610.29304 [cs.CY] |
| (or arXiv:2610.29304v1 [cs.CY] for this version) | |
| https://doi.org/10.48550/arXiv.2610.29304 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jiacheng Ji [view email]
[v1]
Wed, 21 Oct 2026 07:00:26 UTC (613 KB)
来源：arXiv cs.CY · arxiv.org