Nhà khoa học trưởng OpenAI kêu gọi làm chậm tốc độ mở rộng, cho rằng không phòng thí nào giải quyết được vấn đề căn bản của AI

Jakub Pachocki called for a global pause in the maximum speed scaling of artificial intelligence models until safety bars are universally agreed upon. The OpenAI Chief Scientist wrote in a post titled "An Alien Mind" that no research lab has sufficiently solved the problems of alignment and monitoring to justify current development speeds. He argued that international coordination on future AI development must become a top priority for governments to avoid an irresponsible race.

OpenAI is prioritizing research into recursive self-improvement (RSI), which Pachocki identifies as the essential path to remaining at the AI frontier. This process involves machines playing a larger role in their own development and improving the computational substrate itself. While he believes this speed of progress can sustain into RSI, internal evaluations show a decline in the reliability of monitoring chain-of-thought reasoning, making it harder for researchers to inspect how models reason.