MOC: What do you like or dislike about HAL.
Breazeal: HAL is very different from the kind of intelligent robots that I'm trying to build. I try to build robots that are much more inspired by Biology, and are much more creature like. HAL is more of a classic AI. There are some inconsistencies that I see in the way HAL is portrayed. One very striking thing for me is that assuming HAL has these computational models of emotion, its expression of them is pretty impoverished. So really, all you can tell about what HAL is experiencing at an emotional level is done purely through language. There is no change in vocal tone to indicate an emotional state. There's no expressive feedback through nonverbal channels such as facial expression. HAL is a very eerie kind of thing. Given that HAL has these computational models of emotions, they don't seem to be very socially situated. It doesn't seem to be experiencing empathy for the other people on board. Otherwise it wouldn't be going off murdering them. It doesn't seem to be experiencing compassion.

Emotions don't play a part in shading HAL's information processing, which is important in decision making. You're left thinking Ôwhoever designed HAL's emotional system really blew it." [laughs]

MOC: What is your ideal intelligent identity. If you could solve all of these technological problems, what would you imagine?
Breazeal: Through the process of trying to build intelligent systems I'm really in awe of what we are, as animals. It's completely humbling and awe inspiring. That's my bias. When I try to build intelligent systems, I'm trying to build systems that capture some essence of life. When I think about intelligent behavior, I think about life- like behavior.

MOC: What role do emotions play in intelligence? Is it something that really is desirable?
Breazeal: Yes, if you look at living systems, we've evolved to have emotions obviously because they promoted our survival. Humans are the most emotional, the most expressive creatures on the planet. We also consider ourselves to be the most intelligent. It's all wrapped up in there. The study of emotions was out of favor for quite some time but now its back in favor. We're starting to find out how emotions play such a critical role in how we make decisions, how we form memories, how it biases our perceptions. Emotions play an incredibly important role in our intelligent behavior.

One thing that really struck me about HAL was that it only valued humans as far as their utility. Its own sense of self value was based on how useful it was. I don't think that's how people value them selves. I think people value ourselves in terms of what we give to other people as well. Our sense of self worth is how we can improve other people's lives.

MOC: Do you find that being a woman you have different concerns than your male counterparts?
Breazeal: All I can say is that I don't necessarily think so. Given the success of Kismet, many people are interested in exploring similar kinds of research agenda, men as well as women. There's certainly people who trivialize Kismet because of the infant caregiver metaphor, because it looks cute. They try to dismiss it because of surface appearances. When really the questions the research poses are deeply important.

If research with robots like Kismet opens more women's eyes to thinking "Oh this is cool, this is something I'd be interested in" then I'm all for it. I would love to see more women coming into the field. Maybe one of the reasons there aren't more women in the field is because they don't realize that they can make it what they want it to be. They can take it in a different direction. They just need to see the possibilities.