#65 – Katja Grace on Slowing Down AI and Whether the X-Risk Case Holds Up
Hear This Idea
00:00
The Problems With Goal-Directed AI Systems
I'm worrying since will the goals just be really bad? One reason to think that might be fine is that we have reasons to expect AI systems even now seem reasonably good at understanding, even barely complex goals and values. Maybe there are stories where they still don't end up in some sense sharing those values or just caring about that at all. I feel like the concerns that there have been are maybe somewhat different in a world of large language models. It's like pretty unclear. I mean, it seems like they don't actually have goals probably. And if in the end they're acting agentically because they're sort of role playing an agent but they're kind of doing it aloud.
Transcript
Play full episode
