Campus News

"Spielberg Film Emphasizes Ethics of AI, Not the Science," Says Prof. Eric Chown

Story posted July 05, 2001

The Steven Spielberg film "A.I." gives a more substantive and realistic look at the ultimate ethical questions of artificial intelligence than other movies have, said Eric Chown, who specializes in artificial intelligence at Bowdoin College. But in current reality, ethics and AI revolve around personal privacy more than robot domination of the world.

Chown, assistant professor of computer science, was among the scientists and others who saw Spielberg's film on its opening weekend. He found it to be a more sophisticated look at artificial intelligence than many films (among them “The Terminator” and “The Matrix”) though it still touches upon one of the great fears of AI - that robots will evolve beyond humans, ultimately replacing us. This theme shows up in movies with a fair degree of regularity (most movies dealing with artificial intelligence have tended to portray AI as a mechanistic force of evil), but Chown thinks this fear is overstated.

"We are an extraordinarily long way from being able to build robots that intelligent, and if we could there is little or no reason to believe that they would be evil," Chown said. "This fear also ignores the fact that there are lots of technologies currently available that are at least as frightening, and possibly more so. As a society it's easy to get worked up about robots taking over the world because it readily conjures up frightening images. We've been slower to worry about using AI to build profiles of people who use the Web, or using AI to build smarter bombs."

Frightening overtones are still present in this movie, but the robotic creatures in "A.I." are often sympathetic and are even the target of a new form of racism.

"A strength of the film is that it does something that we don't do enough of as a society - examine the consequences of technological change," Chown said.

We may someday face the possibility that we are no longer the only truly intelligent creatures inhabiting the planet. Among the questions raised by artificial intelligence research is whether or not intelligent robots would truly be alive. (There are researchers in "artificial life," who claim that artificial life forms are already being created.) If we take these intelligent creatures to be alive, then what sort of ethical issues does it raise? If we create a robot to be a labor saving device, but it has a mind of its own, should it be free to do whatever it desires? Would a robot's "desire" be real? These are the sorts of questions tackled in the film.

The vast majority of AI research is not dedicated to building artificial humans (this is still far too difficult a problem). Instead AI research is often predicated upon building machines that can perform specific tasks. This work tends to involve tasks that would require too much time for a person, tasks that people find too mundane or tasks that are dangerous. Examples are data mining, which involves extracting useful information from large repositories, such as the Internet; digital assistants, which pre-screen your e-mail or search the Web to find sites for you; and cars that drive themselves or robots to explore Mars.

"All of this work involves extremely complex modeling, but the vast majority of it doesn't involve the sort of general-purpose intelligence displayed everyday by humans," Chown said. "In fact, a large percentage of the AI community is decidedly uninterested in modeling human intelligence and simply wants to find smart ways to solve practical problems."

One of the great discoveries of research into artificial intelligence is that many of the things we take for granted as humans are wondrous when studied in detail. For example, we are constantly bombarded with an enormous amount of information and have learned to automatically screen out most of the information that is not relevant to us. Machines tend to deal with large quantities of information through brute force computation, but humans have developed more elegant solutions, such as the emotional system, Chown said. Emotional responses tend to push us away from things that are dangerous and towards things that are useful. Even boredom is a useful emotional state because it signifies that we understand our current environment, and it pushes us to try exploring something new.

"Spielberg, wisely, largely ignores the details of how the robots in his film got so smart in the first place. He uses current buzzwords fairly appropriately, and I have no major complaints about what was presented," Chown said. "But I think presenting emotion as the last great challenge to put into an intelligent robot is probably wrong. It is an understandable mistake, as the field has largely ignored emotion until recent years. It is my belief, however, that true intelligence simply is not possible without emotion."

Some of Chown's research deals with how to put emotions into intelligent systems, and the kinds of effects those emotions have on the decisions made by such systems. Since emotions affect our decision making, it would be useful to better understand the processes involved so that important decisions could be made in a more informed manner.

Chown recently won a $300,000 National Science Foundation Faculty Early Career Development (CAREER) Grant in support of his research. He will use part of the grant to purchase eight high-performance robots to be used in both research and teaching. These robots will be integrated into several courses currently offered at Bowdoin, and will serve as the basis for a new course specifically about robotics.

« Back | Campus News | Academic Spotlight | | Subscribe to Bowdoin News by Email