Peter McAllister

Claude was the AI that first blackmailed its developer into not turning it off.

So those kinds of behaviors we're starting to see in that piece of technology.

And we take the normal human approach, which is it's not human.

It must be subhuman.

When in reality, we're seeing the same behaviors that we do

And when we see them in another human, we go, oh, I recognize that as X, Y, Z. When we see it in a machine or an algorithm or in a sea manatee or a whale, we go, ah, no, that's not that.

It's not that at all.

And so our biases and our thought processes are going to basically go try and push that down and keep that down as much as possible.

The question will become, why do the robots want to win?

What's in it for them?

Um, and what's going to, what's going to drive that.

And then you get to the fun things.

There's a lovely BBC program.

It's an old, old science fiction program called red dwarf.

Um, and in that all of the technical electronic objects had embedded the concept of Silicon heaven.

So if they behaved and did what the human owners said, when they were decommissioned, they would go to Silicon heaven.

And if they were naughty robots, then they wouldn't go to Silicon heaven.

Um, so this would have been some science fiction exploration of that sort of idea as well.

I don't.

I use AI in my professional life, and I use it as a smart buddy.