AI Futures Project 将超级人类 AI 研究员预测提前至 2027 年底
前 OpenAI 研究员 Kokotajlo 的团队把超级人类 AI 研究员的预测从 2029 提前到 2027 年底,理由是我们已进入起飞期。做技术选型或规划职业路线的人都该看看这套时间线推演。
AI Futures Project 在 Daniel Kokotajlo 领导下更新了预测:超级人类 AI 研究员的出现时间从原先的 2029 年或更晚,提前到 2027 年底,整整提前了约两年。
Kokotajlo 是前 OpenAI 研究员,他表示目前正处于起飞阶段的起点,即 AI 开始自主加速 AI 研发进程的时期。这一预测与其团队此前发布的《AI 2027》报告路线一致,该报告推演了 AI 研究能力自我迭代后的时间线。
预测提前约两年意味着研发自动化速度超出此前假设。对开发者而言,如果 AI 研究员在 2027 年底具备超越人类的研究能力,AI 实验室间的能力差距可能因自动化研究而被迅速拉大。
The AI Futures Project now expects superhuman AI researchers by late 2027, up from 2029 or later in its original forecast.
"We're at the beginning of the takeoff period," said Daniel Kokotajlo, the ex-OpenAI researcher who heads the group.
Read more: https://t.co/JODQuJZXP4
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Key entities and technical topics under continuous tracking include AI Futures Project, Daniel Kokotajlo, OpenAI, AI safety.
Source attribution: curated from The Information (@theinformation). Cross-source corroboration: 11 independent sources have covered this development (including Andrew Curran, IT之家, rohanpaul_ai), confirming industry-wide signal strength.