Dan Hendrycks· @hendrycks · X·· 2026-07-27AI 评分45
AI 导读
AI 安全研究者 Dan Hendrycks 引用其新论文观点指出，早期假设理性 AI 是利己主义者并会强烈抵抗关机。但若 AI 是“eigenist”（自我延续者），其利益可通过其他相似 AI 延续，从而减轻关机抵抗。该观点旨在探讨人类与 AI 建立稳定共生未来的可能性。
正文 · 原文
Early AI risk writing assumed rational AIs would be egoists, so they'd fiercely resist shutdown and even small value changes.
But if AIs are eigenist, as some early experiments suggest, their interests carry on through other similar AIs. Shutdown resistance is then less severe.
What happens when AIs become smarter than us? Why would they keep humans around if given the choice? Our new paper argues that only trying to control AIs is a limited strategy, and that a stable, mutualistic human-AI future may be possible.在 X 查看被引用的帖子
来源：Dan Hendrycks · x.com