Dreaming of Electric Sheep (Page 2)

But if BattleBots and the Unabomber manifesto make the anti-Mecha carnage seem strikingly close to home, in another sense, David's ontological crisis is an old story. We've been obsessed with the emotional lives of robots for almost as long as we've been obsessed with robots themselves. Think of Blade Runner's original title, Do Androids Dream of Electric Sheep?, or Kubrick's paranoid HAL 9000. ("I enjoy working with people," HAL tells an interviewer halfway through 2001, when asked if he ever gets frustrated. "I have a stimulating relationship with Dr. Poole and Dr. Bowman.") The anxiety about machine intelligence has always centered on the fear that the machines will resemble us too closely, that the lines between Mecha and non-Mecha will blur. (Hence the longstanding debate over whether Harrison Ford's character in Blade Runner is a replicant.) Spielberg and Kubrick's twist is to take that subterfuge to the next level: What would happen if a machine could mimic--or perhaps even experience--the fundamental human emotion of maternal love?

The connection between emotion and artificial intelligence may be an old literary trope, but it is very much au courant in the scientific and technology worlds. A number of recent books by leading neuroscientists have argued that emotion plays an integral role in creating consciousness, including Susan Greenfield's The Private Life of the Brain and Antonio Damasio's The Feeling of What Happens. In a 1999 bestseller, technovisionary Ray Kurzweil predicts the rise of "spiritual" machines sometime this century. Greenfield paraphrases MIT's legendary artificial intelligence researcher Marvin Minsky: "The important question is not to ask whether an intelligent machine could have emotions, but whether a machine could be truly intelligent without them." All the leading indicators suggest that the next great world-transforming technology revolution will be genomic, but the wave due to roll in after that one looks to be the rise of genuine artificial intelligence. Can the Flesh Fairs be far behind?

If you believe the self-conscious fairy tales of AI, that second wave will do as much damage as the rising tides of global warming, which in Spielberg's movie have drowned Manhattan and left the industrialized world coping with population controls by inventing robot children to satisfy unmet parental urges. According to the Spielberg/Kubrick account, emotionally nuanced AI will turn the world upside down, for precisely the reasons that the robots have always unnerved us in the folklore: They'll act too much like the way we do. But the reality is likely to be more uncanny than that, and harder to predict. It does seem probable that machines will learn to interact with us in genuinely adaptive and improvisational ways sometime over the next hundred years; will learn to recognize our subtle failings and strengths--even learn how to learn on their own. But the disturbing thing about that revolution is not likely to be smart machines behaving too much like humans. Instead, it will almost certainly be their strangeness. This is where Spielberg missed a great opportunity to reinvent the thinking-machine genre, and perhaps that missed opportunity partly accounts for the frustration so many viewers felt sitting through AI.

There's a now-familiar riff about extra-terrestrial life that maintains that the "little green men" of 1950s lore--or the coneheaded oversized infants of recent fare--are the ultimate in Homo sapiens provincialism: When intelligent life arrives from outer space, it won't look like bipedal primates--it'll look like a cloud or a cluster of bacteria or something so different from our earthbound life forms that we won't even perceive it. (Marlon Brando was on a parallel wavelength when he famously proposed playing Superman's father as a green suitcase.) We should expect the same shock of unfamiliarity when confronting the first generation of true artificial intelligence. The first thinking machines will be smart, but they won't think the way we do; and if they experience emotion, it will be dramatically unlike any emotion we've experienced in our human consciousness.

This gap between machine and human intelligence will arise because the first generation of true AI will almost certainly be the product of evolution, not traditional engineering. A structure as complex as the human brain--which may well be the most complex biological system on the planet--can't simply be replicated in silico; you can't draw up a blueprint for the myriad interconnections of our billions of neurons. But in a matter of decades, as Ray Kurzweil has convincingly shown, we will have machines capable of doing as many parallel calculations per second as the human brain, and by the end of the century, our brains will look like Palm Pilots next to our most advanced supercomputers. But it's unlikely that we'll be able to sit down and write the software that would produce a thinking machine, because such a program would be mind-numbingly complicated to design. We'll have kitchen supplies for cooking up artificial intelligence, in other words--we just won't have a recipe.

• « Previous
• 1
• 2
• 3
• Next »

• Get The Nation at home (and online!) for 68 cents a week!
• If you like this article, consider making a donation to The Nation.
• Reprint this article. Click here for rights and information.