Hi, everyone. So as you know, I've been doing a ton of these shows on AI. And while I think these shows are a perfect fit for Skeptico, they're also kind of not a perfect fit for some of the people I'm trying to pull into this dialogue that are part of the AI community. So what I've done is created a separate podcast called AI Truth Ethics that really, I think, more

succinctly captures what I'm trying to do with this AI thing.

And hopefully it'll lead people back to the larger Skeptico project.

But as some of you are aware, you know, when you introduce somebody to skeptico cold, it can be kind of jarring.

And that's what I'm running into. So anyways, the kickoff of that new podcast has three short

episodes, which I think are really great. I was really happy to do them because I think it does

distill down the essence of the case I've been making

about AI. So I present those to you here. They're just kind of one right after another,

but I definitely think you'll get what's going on. Welcome to AI truth ethics. I'm Alex Sikaris,

and this is the first podcast in this series. I've been doing a number of these shows

on AI ethics and AI dialogues that I've been publishing, but I've been doing them over on my other

channel, skeptico.com. So I want to kind of reboot the dialogue over here and talk to more

of an AI audience. And I'd really like to engage with you

and hear more what you think about the idea I'm putting out around truth, being a core part of

AI ethics and being something that is under explored. So I start with this offering here on kind of

a fundamental issue with AI ethics, and that that is alignment as we're calling it.

Poor alignment of the vertebrae.

Reject antenna alignment.

Engine alignment.

The alignment problem is like we're going to make this incredibly powerful system

and like it'd be really bad if it doesn't do what we want.

So if I find cases where you are not being truthful,

...