Welcome to Skeptico and really a joint broadcast of Skeptico and the Truth Ethics Podcast.

Yeah, I'm going to have some of these shows that are separated, but some of them, it's probably best just to do them together and save me a little bit of editing headache.

So on this version, I have a swapcast with Ben Bifert, who is the creator and host of the Machine Ethics

Podcast and does consulting under the title of Ethical by Design, Artificial Intelligence.

So you would think this would be a natural synergistic swapcast. There's no controversy here.

If only it was so.

If you've been with me for a while, especially my ride on Skeptico, you know that sometimes

people are very sensitive to folks coming in from what they perceive as the outside,

who have different opinions about what has been staked out and understood as their domain.

Now, what I think is super interesting about this case, and you'll hear it over and over again,

it's not just with Ben, but it really kind of came to a head with Ben.

And that's that we're in a different position with AI.

AI has this incredibly kind of self-reflective ability to think about what it's

thinking about and report what it's thinking about. Let me give you an example. Ben got so

frustrated during this dialogue that he started saying things that I think are obviously not true.

And one of the positions that he took both in the dialogue and especially

in the email exchange we had afterwards is that I had this thing totally backwards and mixed

up and that AI truth and transparency, though that's not really part of AI ethics. That's not

really part of the alignment problem. You're kind of coming in here and redefining everything.

And this isn't

what it's really about in this field. Well, again, like I just said, the interesting thing about

this is you can just go and ask. And so that's what I did. So Gemini, there seems to be a little

...