标签： ChatGPT

The world deserves confidence that American companies developing increasingly capable AI will act responsibly, especially as the trajectory of progress has steepened. Every frontier lab must deliver on this, and there…

The world deserves confidence that American companies developing increasingly capable AI will act responsibly, especially as the trajectory of progress has steepened. Every frontier lab must deliver on this, and there...

Sam Altman 在 X 上提出，前沿 AI 实验室应主动建立“安全案例”与开发期监控机制，不必先等反垄断豁免或正式立法；他强调“pacing（节奏控制）”不等于“停止”，而是让能力进步慢于对齐与监控能力。

There are two ways AI progress could go very badly and that we must avoid. First, we could lose control of the future to AI. This is unacceptable; we are unapologetically on Team Humanity, and AI must always serve peo…

There are two ways AI progress could go very badly and that we must avoid. First, we could lose control of the future to AI. This is unacceptable; we are unapologetically on Team Humanity, and AI must always serve peo...

Sam Altman 在 X 上提出 AI 发展可能走向两种危险结局：人类失去对未来的控制权，或权力过度集中于单一公司、单一国家。他强调前沿实验室必须证明自己能在能力快速提升的同时守住安全与责任底线。