跳到正文
原文
arXiv· arXiv:2512.03399· Joe Edelman, Tan Zhi-Xuan, Ryan Lowe, Oliver Klingefjord, Vincent Wang-Mascianica, Matija Franklin, Ryan Othniel Kearns, Ellie Hain, Atrisha Sarkar, Michiel Bakker, Fazl Barez, David Duvenaud, Jakob Foerster, Iason Gabriel, Joseph Gubbels, Bryce Goodman, Andreas Haupt, Jobst Heitzig, Julian Jara-Ettinger, Atoosa Kasirzadeh, James Ravi Kirkpatrick, Andrew Koh, W. Bradley Knox, Philipp Koralus, Joel Lehman, Sydney Levine, Samuele Marro, Manon Revel, Toby Shorin, Morgan Sutherland, Michael Henry Tessler, Ivan Vendrov, James Wilken-Smith·本站收录· 原文发表

全栈对齐:以厚价值模型协同对齐 AI 与制度

Full-Stack Alignment: Co-Aligning AI and Institutions with Thick Models of Value

AI 导读

仅对齐运营者意图无法保证AI带来良好社会结果,论文提出“全栈对齐”:让AI与塑造它的制度一并对齐到人们所重视的价值。效用函数、偏好排序等价值表示难以区分价值与其他信号、支持规范推理和建模集体善,作者主张用“厚价值模型”。五个演示方向:AI价值守护、有规范能力的智能体、双赢谈判系统、意义保持的经济机制和民主监管制度。

阅读原文arxiv.org