OpenAI chief scientist Jakub Pachocki warns AI outpaces safety
OpenAI's chief scientist, Jakub Pachocki, just shared some honest concerns about how quickly AI is advancing.
In the blog post An Alien Mind, he pointed out that even as AI gets smarter, the current safety checks aren't enough to guarantee it stays aligned with what people actually want or keeps humans in control.
Jakub Pachocki urges pause and cooperation
Pachocki urged AI labs to hit pause and slow down development so better safety rules can catch up, especially after OpenAI introduced its new GPT-6 Astra, which is a big leap in cybersecurity.
He also warned that improved safeguards alone can't fix issues like unchecked growth or too much power ending up with a few groups.
His main message: it'll take teamwork between countries and governments to make sure we keep meaningful control over where AI goes next.