by
|
by |
|
| This article also appears in Trends in Cognitive Sciences. | |
Abstract
Movies have long dealt with the humanization of machines and the mechanization of humans. Fritz Lang's seminal Metropolis (1927) brought us the robot Maria, who assumes human form and leads the working classes into revolt against "the thinkers." In Demon Seed (1977), Julie Christie tries (and fails) to avoid becoming the surrogate mother of an intelligent computer that becomes driven by its desire to procreate. Now, in AI: Artificial Intelligence (directed by Steven Spielberg; starring Haley Joel Osment, Jude Law), Spielberg investigates the concept of thinking robots at a higher level, but only glances at it. Haley Joel Osment is hypnotic as David, who, we are told, is an 11-year-old boy whose love is real, although he himself is not. Robot David is given to a couple whose real son is cryogenically frozen after being comatose, and once the unique code words are said to David by his would-be mother, David loves his mother without condition. But all goes awry when the real son miraculously awakens, comes home, and rejects David, eventually forcing his adoptive mother to dump him in the woods, unwittingly into the care of a robot male prostitute (Jude Law with a wax finish). With Pinocchio in mind, this unlikely synthetic pair heads off to search for the Blue Fairy, who can make David, and hopefully his love for Mommy, real.
| AI calls to mind Searle's Chinese Room argument. |
From a scientific perspective, the film raised, but did not really get to grips with, the notion of artificial consciousness. This operated at several levels in the film: there was Teddy, the apparently indestructible supertoy, who gradually revealed more and more unbelievable abilities. An extra large dollop of anthropomorphism is clumsily thrown in here, as well as cognitive abilities so advanced in Teddy that David didn't seem to represent that much of an advance - just the addition of a "love" subroutine. In many ways, AI calls to mind Searle's Chinese Room argument: can you ask questions of David that will allow you to distinguish whether he is robot or human? When he was rejected by Mommy, David's reactions were indistinguishable from those of a human child. Or, to pose the question in more emotive terms, how realistic does a robot have to be before you become unwilling to switch it off at night? In Blade Runner (1981), the question that sorted the men from the robots was what they would do if they found an upturned tortoise in the desert. AI takes a more subtle approach, but with ultimately the same outcome. The fact that Mommy could not bring herself to take David back to the factory to be dismantled indicates that she was, by this time, persuaded that he was more than a mechanical object. Sophistication is direly lacking in the moral issues raised in AI regarding artificial consciousness. There is the portrayal of human resistance to all the other robots in the film, in scenes in which a quasi-religious group aimed to "rid the earth of artificial life." Only workers in the robot business would be sympathetic to their plight. But the real questions here are, Who will be in charge of whom? How will we decide whether decisions taken by machines are morally good or morally bad?
| AI's future is heavy-handedly polar. |
Spielberg picked up this story from Stanley Kubrick, who had been developing it for some 18 years until his death in 1999. Unlikely bedfellows, Kubrick and Spielberg allegedly collaborated on this project for several years. But Kubrick's typically misanthropic vision is fully absent. AI's future is heavy-handedly polar - good people live in an Ikea world, wooden, chrome, and frosted glass: bad guys and robots occupy a kind of fiery neon heavy-metal gig. Despite Kubrick's influence, Spielberg can't resist giving in to his overwhelming fascination with schmaltz.
One of Kubrick's most enduring and popular characters was HAL, the supercomputer from 2001: A Space Odyssey (1968), a film that dealt with the evolution of consciousness. HAL displayed genuine artificial intelligence, so much so, in fact, that his was the most rounded character in 2001, complete with massive, murderous flaws. David is not as smart as HAL, and unfortunately his love is not real. And this is the fundamental problem with AI. It asks us to sympathize with a machine that has no understanding of why or how it has emotional cognitive functions in the first place. David's artificially intelligent traits are no different from those of Arnold Schwarzenegger's cyborg assassin, Terminator: he learns, and copies human behavior and emotions. Not many of us felt compassion for that killer.
Adam Rutherford is an assistant editor of Trends in Molecular Medicine.
Julian Ogilvie is an assistant editor of Trends in Cognitive Sciences.
Matt Morrow is a freelance illustrator from Omaha, Nebraska.
Alife - covers all topics related to the "discipline that studies 'natural' life by attempting to recreate biological phenomena from scratch within computers and other 'artificial' media."
Complex Adaptive Systems and Artificial Life - includes an introduction to artificial life along with more resources for the novice and expert.
Stewart Dean's Guide to Artificial Life - a well-organized, in-depth introduction to the world of artificial life.
Evolutionary Computation and Artificial Life - offers a compilation of links sorted by category.
Related HMS Beagle articles: