OpenAI 安全委员会成员 Paul Christiano 称 AI 接管风险最高达五成
说话的人不是普通学者，是参与发明 RLHF、现在坐在 OpenAI 安全委员会里的 Paul Christiano。他给出 3 年内接近五成的灾难概率估算，和实验室对外宣传的乐观口径差别很大。做安全政策、对齐研究或想了解 OpenAI 内部风险分歧的读者可以看看原视频。
OpenAI 安全与安保委员会成员 Paul Christiano 在《The Hinton Lectures》演讲中表示，未来 6-12 个月内存在 AI 接管的实质性风险，如果以最快速度、最简单方式构建前沿 AI，未来 3 年内发生灾难的概率接近五五开。他声称目前已经观察到重大事故发生前可能出现的迹象，并且即便采取现实可行的政策应对，风险依然极高。
Christiano 的判断带有特殊分量：他是 RLHF（人类反馈强化学习）的共同发明者，这项技术在 2017-2021 年间于 OpenAI 诞生，至今仍是 ChatGPT 等模型对齐训练的核心方法；2024-2026 年他还担任美国政府 AI 安全机构（现名 CAISSI）安全负责人。也就是说，他本人处于 OpenAI 内部阻止高风险实践的委员会中，同时深度参与过官方安全监管。
Gary Marcus 转发时表示不同意其技术判断，但提出了一个尖锐质疑：如果 OpenAI 董事会成员认为存在这种级别的风险，这家公司及其竞争对手为何仍被允许按现有速度继续运营。该发言出自《The Hinton Lectures》系列讲座，相关活动仍在持续举行。
i don’t agree but if board members are saying this why the fuck is this company (or its competitors) allowed to operate? Rob Wiblin @robertwiblin Paul Christiano (OpenAI board member) yesterday: "There's material risk of AI takeover in the next 6-12 months. We have seen roughly the things we might expect to see shortly before something really bad were to happen. Even with [any] realistic policy responses, the risk remains exceptionally high. If we build this as quickly as we can in the easiest way we can the chance is [more like] a coin flip [within the next 3 years]. For most people in this room, the most likely reason they will die is because of an AI accident." Christiano: • Is on OpenAI's safety and security committee (which exists to block risky practices) • Co-invented RLHF at OpenAI back in 2017-2021 • From 2024-2026 was head of safety at the US gov's AI agency (now called CAISSI) (Video is from 'The Hinton Lectures' there's another on tonight and tomorrow.)
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI regulation, compliance frameworks, and sovereign compute policy. Key entities and technical topics under continuous tracking include OpenAI, Paul Christiano, RLHF, AI安全.
Source attribution: curated from Gary Marcus (@GaryMarcus). Cross-source corroboration: 11 independent sources have covered this development (including rohanpaul_ai, Latent Space (swyx), Andrew Curran), confirming industry-wide signal strength.
