Rep. Ro Khanna on AI: What the Legislation Would Ban Is Recursive Self-Improving AI That Can Train Itself

RUSH EXCERPT:
KHANNA: "Not all of them. Obviously, some of the executives are opposed. Now who supports me is a lot of the engineers and researchers. Many of them actually who work at openai and anthropic. Later today I am meeting with a group of 10 to 15 of them at these companies. And what the legislation would ban is recursive self-improving AI. That is basically AI that isn’t being trained anymore by human beings. It’s AI that can train itself. And that is a real risk of losing human control. What the Senator Sanders and Greg kiz ar and Jeffrey hinton, the godfather of AI are saying is until we have guardrails, we can’t let AI train itself and lose control. It’s something we don’t have the capability to do right now but shouldn’t get to the point where we actually are doing it."