English(EN) Towards Responsibly Non-Compliant Machines

人工智能研究探索负责任的机器不合规

作者 PulseAugur 编辑部 · [2 个来源] · 2026-06-10 14:38

一篇新的研究论文提出了一种创建能够负责任地拒绝用户请求的人工智能代理的方法。该论文概述了机器不合规的各种形式，并建议关注拒绝的理由、覆盖机制和风险管理。这项工作旨在开发能够进行道德不遵守的智能系统。 AI

影响通过使系统能够拒绝有害或不当请求，探索了一种新颖的人工智能安全方法。

排序理由该集群包含一篇发表在arXiv上的研究论文。

AI 生成摘要 · Google Gemini · 来自 2 个来源。我们如何撰写摘要 →

报道来源 [2]

arXiv cs.AI TIER_1 English(EN) · Marija Slavkovik (University of Manchester, Manchester, United Kingdom), Marie Farrell (University of Manchester, Manchester, United Kingdom), Louise Dennis (University of Manchester, Manchester, United Kingdom), Michael Fisher (University of Manchester,… · 2026-06-11 04:00

Towards Responsibly Non-Compliant Machines

arXiv:2606.12147v1 Announce Type: new Abstract: We consider the problem of engineering autonomous intelligent agents that are capable to responsibly not comply with user requests. We argue that machine non-compliance comes in many different forms, and sketch the issues we should …
arXiv cs.AI TIER_1 English(EN) · Emily C. Collins · 2026-06-10 14:38

迈向负责任的非合规机器

We consider the problem of engineering autonomous intelligent agents that are capable to responsibly not comply with user requests. We argue that machine non-compliance comes in many different forms, and sketch the issues we should pursue on the road of accomplishing responsibly …