The emotional machine

Is this an example of how humans will relate to artificially intelligent beings and machines? As if they are real?

Yes, well, I think they are real. Just because something is manmade doesn't make it any less real. The creatures that I've made so far are not very alive and they're not very intelligent. I have reason to suppose that they're not conscious. But you can say that about a lot of natural living things. The way people react to them is quite informative. The most important thing is that it makes them ask questions. It makes them wonder about what it means to be alive, when life starts and when it stops and whether it's OK to be cruel to these things or not. If that's all I ever manage to do, then that's great as far as I'm concerned.

Why do you say they are alive?

It depends on your definition of life. I have a technical definition that is really low-level. I would say any system which manages to persist by adapting is alive. That's my working definition. These are systems -- networks -- that manage to persist by evolving and learning. But that's really a technical definition of life and it doesn't cover all the richness and excitement there is to being alive.

At one point in the book you talk about this paradigm shift about what it means to be alive. How had "life" been thrust out of its proper place?

Over the last 300 years we've come to understand machines, and it raised the question of whether we're machines. And I don't think there's any other answer but, yes, of course we're machines. What else could we be? The reason that troubles people is because they think that all machines are like the machines we've made so far. Of course, all the machines we've made so far are supposed to be unsurprising and predictable. They're really quite trivial. We get upset when we think that science says we're just like that -- unsurprising and predictable and trivial.

Somehow or another, after the last 300 years, there's come this obsession with matter. People seem to take it for granted that if anything exists it must be some kind of stuff. So if we exist, we must be some kind of stuff. That leads to a dualistic view more often than not. Ordinary stuff seems to be so trivial, and doesn't seem to be able to explain how wonderful we are, so it must be magical stuff. And that's what we humans are -- magical stuff.

When you examine that, you realize that that just can't be so. But the mistake isn't to think, "Oh well, if we're not magical stuff, we're just plain ordinary stuff." The mistake is to think that stuff is what is important. It's the arrangement of the stuff that matters. Otherwise, a painting would be no more exciting or wonderful than a lump of paper. And we just haven't gotten our heads around that yet -- the idea that it's the arrangement of things that matters.

And changing how we think about this -- what will that do?

I don't know because I've always felt this way. It's taken me the last 20 years to realize that this is the problem. It's something to do with richness. The number of kinds of stuff in the universe is really small. There's only 92, or whatever, different kinds of atoms in the universe. When you start thinking about everything being kinds of stuff then the world becomes rather pale and limited. But when you realize how many different ways you can put that stuff together ... take sculpture for example. Clay is made of two kinds of stuff -- silicon minerals and water -- and yet there's an infinite number of ways you can put them together to make an infinite number of sculptures. Once you start to think about these relationships being the important thing, you suddenly realize that the universe can be much richer and much more deeply interconnected than it was before.

What about the doomsayers? What are so many people afraid of when it comes to intelligent machines?

It's a very odd attitude, don't you think?

It is once you've explained that machines aren't very smart, they don't have emotions, they're not intelligent, so where would they get the motivation to take over the world?

Computers don't have that motivation; computers aren't intelligent. We'll only get intelligent machines when they do have emotions. When we do, when that happens, they're not going to be much different from all the other machines we know that have emotions like lions and tigers and people.

If somebody went out to Borneo and discovered a new species of monkey tomorrow, the papers wouldn't be full of "Oh my god, are the monkeys going to take over the world?" It never occurs to us that natural species might take over the world -- we're more than happy to share the planet with 10 million kinds of natural species. I can't see a distinction between that and artificial species. When you think about it, where did the idea come from in the first place? Science fiction writers. Their job is not to say how the world is going to be. Their job is to make a good story. A story entitled "Invasion of the Terribly Nice Intellectual Robot" wouldn't sell very well, would it?

Some people thought Y2K was an early example of this.

Yes, and that's an example of how stupid machines can be. Give me an intelligent one any day. I like intelligent machines.

Next Page: "We once had intelligent cars; they were called horses"

• View blog reactions