Week of May 9, 2001

Here we are in the year 2001, and - despite the movie's predictions-- computers have yet to develop minds of their own. How close are we to developing machines that can simulate human thought? This week on the Infinite Mind, we look at the latest research on Artificial Intelligence. Guests include Brian Aldiss, writer of the short-story, "Super Toys Last All Summer Long," which is the inspiration for Steven Spielberg's upcoming movie, AI: Artificial Intelligence; Dr. Peter Norvig, co-author of the standard textbook on AI, Artificial Intelligence: A Modern Approach and Chief of the Computational Sciences Division at NASA's Ames Research Center in California; Dr. Rosalind Picard, founder and director of the Affective Computing Research Group at the MIT Media Lab; and Dr. Marvin Minsky, a pioneer in the field of AI, who's now a professor in the MIT Media Lab and was co-founder and, for many years, director of the MIT Artificial Intelligence Lab.

Host Dr. Fred Goodwin begins with an essay about the range of opinions researchers hold about what might be possible within AI. Some believe computers will ultimately be able to simulate or surpass human intelligence. Others question whether this would be possible and focus, instead, on making machines that will be better able to help us solve complex problems, without a consciousness of their own. Dr. Goodwin sees a similar split within neuroscience. Some neuroscientists believe we will eventually be able to uncover all the secrets of the brain, and thus will find answers to the age-old questions such as "Where do we come from? Why are we here? Where are we going?" Others think science will never be able to reveal these secrets of humanity. For now, whether or not one believes science can uncover - and recreate -- the mysteries of the brain is, for now at least, a position of faith, not fact.

Next, The Infinite Mind's Marit Haahr speaks with British science-fiction writer Brian Aldiss. His books include Report on Probability A, The Saliva Tree, and Billion Year Spree: The History of Science Fiction. His 1969 short story, "Super Toys Last All Summer Long" is the inspiration for Steven Spielberg's upcoming movie, AI: Artificial Intelligence. Aldiss reads from "Super Toys" - the story tells of a couple in the future who own robots in lieu of children; it touches on the difficulties one might have communicating with a machine that believes it's real. Aldiss comments that he believes science fiction pushes science forward - as soon as some new discovery is made, writers seize upon it, and suddenly the possibilities seem much closer and more likely.

Click here to order Supertoys Last All Summer Long: And Other Stories of Future Time at Barnes and Noble.com.

Our next guest is co-author of THE textbook on AI, Artificial Intelligence: A Modern Approach. Dr. Peter Norvig is Chief of the Computational Sciences Division at NASA's Ames Research Center in California. His team conducts research and development in autonomous spacecraft, human-centered computing, automated software engineering, automated learning, and neural networks. Dr. Norvig discusses the origins of AI, from the early days of computers in the 1950s. Although some of the original predictions about machines with consciousness have yet to come true, he believes the field is doing well, considering it's only 50 years old. He says the original goal of AI was to achieve processing at the level of the human brain, but now many researchers are concentrating, instead, on just achieving better performance, without necessarily trying to mimic the way human beings think. The focus has shifted from trying to replace human beings on the job to creating intelligent agents that can help people perform complex tasks. One example is bots - software that continually searches the Internet looking, for instance, for good shopping deals. In his work at NASA, Norvig is looking to create spacecrafts that have intelligence on board, so that on a long and difficult mission - say, to Mars - the rover or probe would not have to depend on ground control for step-by-step instructions.

To contact Dr. Norvig, write to: Dr. Peter Norvig; Chief, Computational Sciences Division; NASA Ames Research Center; Moffett Field, CA 94035-1000. Or visit www.norvig.com.

Click here to order: Artificial Intelligence: A Modern Approach by Dr. Peter Norvig at Barnes and Noble.com.

Next, Dr. Rosalind Picard joins us to talk about a new branch of AI. Dr. Picard realized that AI researchers had ignored one important aspect of intelligence - emotion. She's founder and director of the Affective Computing Research Group at the MIT Media Lab, where she's now working on developing computers with emotional intelligence. She's found that people interact with computers in much the same way that they interact with other people - for example, we are nicer to a computer face-to-face than we are behind its back. And just as we might become frustrated with a person who is unresponsive to our emotions, we become frustrated with computers, since they do not react when we become angry or annoyed with them. This not only leads to unhelpful actions (like kicking our computers), but also reduces productivity. She's working on ways to help computers register our emotions, so that we become less irritated with them.

To contact Dr. Picard, write to: Dr. Rosalind Picard; Director of Affective Computing Research; Associate Professor of Media Arts and Sciences; M.I.T. Media Laboratory, E15-392; 20 Ames Street; Cambridge, MA 02139. Or visit www.media.mit.edu.

Click here to order: AFFECTIVE COMPUTING by Rosalind Picard at Barnes and Noble.com

.

Next, reporter William Speed Weed looks at the possibility of using computers to conduct therapy. A handful of psychologists have started using computers to treat people with panic attacks, eating disorders, stress, depression, and phobias - and with great success. These tools are not "intelligent" yet, but some scientists are working on developing a "therapist in a box" that could be used by NASA for a manned mission to Mars.

Dr. Goodwin next interviews Dr. Marvin Minsky, who is one of the founders of the field of artificial intelligence, and, since the 1950s, has been at the forefront of AI research. He's currently a professor in the MIT Media Lab and was co-founder and, for many years, director of the MIT Artificial Intelligence Lab. His book, The Society of Mind, lays out his ideas on the structure and functioning of the human mind. Recently, he has been working on giving machines the capacity for commonsense reasoning. He thinks researchers have made the mistake of trying to program computers to come up with the best solution to a problem, rather than to do what the human brain does - coming up with five or ten pretty good solutions. That way, when one solution doesn't work, the computer can try another. He believes that AI researchers should not give up on creating machines that can surpass human intelligence, but, until we create a computer with a child's capacity to learn and adapt, we will not achieve true AI. Furthermore, he thinks that the only way we will evolve beyond our current state is to enhance our brains with the help of computers.

To contact Dr. Minsky, write to: Dr. Marvin Minsky; M.I.T. Media Laboratory; 20 Ames Street; Cambridge, MA 02139. Or visit www.media.mit.edu.

Click here to order THE SOCIETY OF MIND by Marvin Minsky on Barnes and Noble.com.

Finally, commentator John Hockenberry gives his thoughts on AI, evolution, and his "pal" Marvin Minsky.

· Back to the The Infinite Mind Index