Saturday, November 03, 2007

HCI and Human Robot Interaction

These last weeks I have met people from the Human Robot Interaction research field (Jodi Forlizzi, CMU and Peter Kahn, UW, and our new colleague Matthias Scheutz, IU). It has been highly interesting and I have learnt and realized things that I had not thought about. It is clear that the two fields are moving closer to each other with a growing overlap. If HRI has primarily been addressing the internal workings of a robot, HCI has been all about interaction. When HRI now actually can build robots with quite interesting and sophisticated qualities, the design challenges between the fields are becoming similar. Interaction design is becoming more aware of the dynamic and interactive (even intelligent) environment that people experience as part of their reality. Maybe the only difference is that robots move and environments do not! A "smart" home is a form of non-movable but highly interactive robot. Interaction with a robot with some kind of human persona is getting more common, as well as interaction with environments that have human characteristics when it comes to interaction (if not in their appearance). Anyhow, all this brings the two fields together in a way that is exciting and promising and, I think, good for both fields. So, let's study human interaction with robots or environments in whatever form and shape they come. We have only seen the beginning....By the way, this is also what Don Norman's new book is about (see my last post!).

5 comments:

Kevin Makice said...

Hey, I like the new e-digs. Nice to see that default theme design retired.

I enjoyed Peter Kahn's presentation, even if I agree with David's concerns about the experiment creating some of the findings. I would have liked to see his results compared with some of the in situ studies he mentioned.

It was impressive, though, the degree of coding and the amount of attachment detected between people and conversational objects. My biggest question is where is the line between "robot" and embodied agent drawn. Autonomy? Physicality? There is some research on how people react to embodied virtual agents. Great talk, with lots of inspiring ideas to ponder.

Erik Stolterman said...

Hi Kevin

Yes, I agree, there are many unanswered and interesting questions in this area, and I truly believe that HCI has to pay more attention to them and see them as at the very core of interaction.

Erik

Webfoot said...

It was very striking to me how similar this experiment was to work in Social Psychology. The room, the hidden conspirators, the set-up.
So I was concerned about seeing social relations in social set-ups too.
But, how else? The validity issue here never existed for social psychology because it was about humans in the first place. I've been thinking about it all weekend....

John Waterworth said...

Hi Erik,

I stumbled across your blog and couldn't resist responding to your last post, where you suggest that perhaps the only difference is that robots move and environments don't. It's an interesting thought, and rooms are certainly getting smarter. But I think we should not forget what a fundamental difference there is between things that move and those that don't!

From a biological perspective, it's the difference between plants and animals. Without movement in the environment, we really don't need brains at all.(this is reminiscent of the joke about the similarity between professors getting tenure and the sea squirt - which hunts around to find a place to fix itself permanently, and then devours its own brain because it doesn't need it any more)

Some think that the roots of our consciousness and intentionality lie in locomotion. And of course we become attached more easily to self moving things and creatures (it's a pity the term automobile is taken) in a way that is less common with plants and inanimate objects.

I think that something fundamental happens when organisms, or machines, start to move around under their own control. Movement brings unpredictability - shifts of context that bring with them changes in the specifics of interaction that are very hard to anticipate and so design for in advance.

- John

Erik Stolterman said...

Hi John

Thanks for your insightful comment. And I agree with you, there is a fundamental difference between moving and non-moving things. At the same time, I did not say it was a "simple" difference :-) I think there might be another aspect too, which is when non-moving things become "smarter" and can act anywhere, so it is not only your house, but actions in your house can lead to actions and interaction by other non-moving things. In a way when interactivity is moving between and across things, it becomes quite unpredictability. But, I think that humans has a completely different way to approach and understand and interact with moving things, which has to do with the fundamental issue of locomotion that you mention.

Thanks
Erik