Tom Chi  

Don Norman in Jan. Scientific American

December 15th, 2003 by Tom Chi :: see related comic

Hey! I just sat down to read through my January 2004 Scientific American, and there was Don Norman with an article relevant to this week’s discussion. Luckily it also available on-line here. Side note: notice how much Norman’s pose looks like Darwin. Coincidence?Anyway… in the article he states that decision-making in humans is dependent on both cognition and affective (emotional) reasoning. Thus machines should have emotions (i.e. be programmed with affective models) to help them learn and act. From the article:

“I’m not saying that we should try to copy human emotions,” Norman elaborates. “But machines should have emotions for the same reason that people do: to keep them safe, make them curious and help them to learn.” Autonomous robots, from vacuum cleaners to Mars explorers, need to deal with unexpected problems that cannot be solved by hard-coded algorithms, he argues.

To Norman’s credit, researchers have been trying for a long time to make systems deal with unexpected situations. Any approach which offers a new take is probably worth investigating. The approach also falls in line with our tendency to create technology in our own image: the more that technology is patterned after the ways we think, the easier it will be for us to relate to and extend upon it.

On the other side of the argument, perhaps it is simply not useful to model emotions in a machine. Of all the ways that you could teach a vacuum cleaner not to fall down the stairs, would “fear-modeling” really be the best approach?

I’d also like to add one more dimension to this discussion: besides cognition and emotion, I would also add “motor” decision-making to the mix. Motor decisions include: retracting your hand from a hot stove, squinting in bright sunlight, and shivering in a brisk wind. You don’t rationally decide to do these things, nor are they motivated on your emotions. They happen as a near-instant reaction to the environment and they make up a decent percentage of the ‘decisions’ our bodies make each day. For machines to really come into their own in the real world, they will need to a thorough collection of sensors and a hierarchical system of actions ranging from very low-level (”move away from fire” - motor) to high-level (”pick holiday gift” - cognition/emotion).

2 Responses to “Don Norman in Jan. Scientific American”
Bill Seitz wrote:

Related story in current Wired about Rosalind Picard’s work, where he gets a small quote.

“I want computers to have emotions only to help them survive in the world, not as a way of responding to me,” says Don Norman, a professor of computer science at Northwestern and a computer interface expert. “I’d rather have a machine that knows its place. Otherwise, you feel like it’s a used-car salesman.”

Marc's Voice wrote:

Why machines should Fear

Coolio - gotta go read this…

Leave a Reply

OK/Cancel is a comic strip collaboration co-written and co-illustrated by Kevin Cheng and Tom Chi. Our subject matter focuses on interfaces, good and bad and the people behind the industry of building interfaces - usability specialists, interaction designers, human-computer interaction (HCI) experts, industrial designers, etc. (Who Links Here) ?