August 12th, 2006

ozarque figure

Robots again; who's in charge?

Early in August there was a discussion in this journal about robots/robotics [see, for example, http://ozarque.livejournal.com/288891.html ], set off by my report about a panel on the subject at Conestoga. One of the major questions I was interested in was whether robots designed to do caretaking tasks should look and/or sound like human beings. And I posted a review [at http://ozarque.livejournal.com/289184.html ] of a book on research showing that people respond to robots that do sound like human beings by anthropomorphizing. Saying (on page 3) that over the course of our evolution "humans have become voice-activated with brains that are wired to equate voices with people..." and notes that even when we are entirely aware that the voice we're hearing is coming from a machine, we continue to behave as if its souce were a human being.

And indefatigable42 added something critically important -- something linguistic -- to the mix, making it clear in a comment that what we need is a cognitive shift from "The robot is doing things for you" to "You are using the robot to do things."

Since then, Frances Green has sent me some interesting materials on the subject that I'd like to share with you.

1. Research shows that we humans "equate voices with people." But what if the robot not only doesn't have a human-like voice but has a wheelchair for a body?

Frances sent me a copy of "And they call it robot love" -- an interview with Mari Velonaki by Rachel Nowak, on pp. 48-49 of New Scientist for January 14, 2006 -- about an exhibition titled Fish-Bird "starring two love-struck robots that look like decrepit wheelchairs." (They not only look like decrepit wheelchairs, that's what they're made of.) These robots don't talk at all, so it can't be their voices that are provoking reactions; instead, they print out messages and let them drop on the floor.

On page 49, asked how people have been reacting to the exhibit, Velonaki says: "It is bizarre. People will talk to them, pat them and ask them questions. ... Children... kiss them and encourage them to print more messages. At one venue... they crawled about on the floor to be on the same level as the robots. This was incredible. In 10 years of work I have never seen anything like this. It appears that people project human attributes onto the robots and relate quite closely to them even though they look like wheelchairs."

2. Next, there's "Make sure your android went to finishing school," by Paul Marks, on pp. 30-31 of New Scientist for February 18, 2006, which is about the potential dangers of robots. (Accidental dangers, not deliberate.) On page 30 Marks explains that a standard robotics safety measure is to store the robot in a "work cell" equipped with sensors that turn off its power supply if a human being enters that cell. Which would be fine, he says, "if robots were going to stay in their work cells, but they are not: over the next decade humanoid robots will begin appearing in our homes and offices. In the vanguard of this new market are Japanese firms... which plan to launch domestic robots that will assume roles as varied as receptionists, security guards, entertainers, hospital porters, tourist guides and cleaners."

The only thing in the article that moves even slightly in the direction of "You are using the robot to do things" is an observation on page 31 from a researcher in Denmark, who suggests that when we say a robot is safe for everyday use we should mean that "only in the same sense that a car and driver are safe" for everday use.

3. And then, there's "Almost human," a special section on pp. 38-49 of the February 4, 2006 New Scientist on "the most stunning advances in humanoid bots," which it says "will not only be able to clean the house, do the dishes and take out the garbage, but also to play with children, help care for the elderly and even explore the farthest reaches of space and perform repairs or search-and-rescue missions in hazardous sites on Earth." It describes an array of amazing robots, all of whose designers are obviously doing everything they can think of to make them as humanlike as possible.

It seems to me that this focus on making robots that can't be distinguished from human beings is the worst possible strategy for getting from "the robot is doing something for me" to "I am using the robot to do something"; on the other hand, I wonder if the story about the robots made from wheelchairs is an indication that both human appearance and human voice are irrelevant....
ozarque figure

Recommended link; robots....

Recommended -- at http://www.aaai.org/AITopics//newstopics/resources3.html -- a huge page of brief stories on every kind of AI topic, including many robot/robotics stories; the first one on the page, dated December 21, 2004, is about a robot squirrel. Clicking on "Current News" will take you to http://www.aaai.org/AITopics//html/current.html , which starts with a story from August 12, 2006 about robots for "home or the battlefield."