返回全部动态
OpenAI 报告 AI 研究智能体进展，首席科学家警告控制风险
原标题：OpenAI reports AI "research interns" and warns about its own pace at the same time
AI 摘要
OpenAI 发布内部数据，显示 AI 智能体已深度参与其研究，并宣称达到“自动化研究实习生”里程碑，同时首席科学家 Jakub Pachocki 发文警告，当前对 AI 系统的控制能力不足，链式思维监控可靠性下降，呼吁加强安全标准。
以上摘要由 AI 生成，可能存在误差。事实请以原文为准。
正文节选
OpenAI reports AI "research interns" and warns about its own pace at the same time OpenAI is releasing internal data showing how much AI agents now drive its own model development, and it says it has hit a self-set milestone on the path toward self-improving AI. Chief scientist Jakub Pachocki pairs that with an unusually blunt warning: no lab has solved control of these systems well enough. OpenAI published two matched texts. One is a blog post with internal metrics on increasingly automated AI
发布时间：2026-09-07 21:05
抓取时间：2026-09-07 21:33
来源机构：THE DECODER