It will always be easier to make organic brains by unskilled labor than to create a machine-based artificial intelligence. That joke about doing things the old-fashioned way, which appears in the book version of 2001: A Space Odyssey, still has an undeniable ring of truth. The science-fiction masterpiece will probably be remembered best for the finely honed portrait of a machine that could not only reason but also experience the epitome of what it means to be human: neurotic anxiety and self-doubt.

The Heuristically programmed ALgorithmic Computer, a.k.a. HAL, may serve as a more fully rounded representation of a true thinking machine than the much vaunted Turing test, in which a machine proves its innate intelligence by fooling a human into thinking that it is speaking to one of its own kind. In this sense, HAL's abilities--from playing chess to formulating natural speech and reading lips--may serve as a better benchmark for measuring machine smarts than a computer that can spout vague, canned maxims that a human may interpret as signs of native intelligence.

Surprisingly, perhaps, computers in some cases have actually surpassed writer Arthur C. Clarke's and film director Stanley Kubrick's vision of computing technology at the turn of the millennium. Today's computers are vastly smaller, more portable and use software interfaces that forgo the type of manual controls found on the spaceship Discovery 1. But by and large, computing technology has come nowhere close to HAL. David G. Stork, who edited Hal's Legacy: 2001's Computer as Dream and Reality, a collection of essays comparing the state of computing with HAL's capabilities, remarks that for some defining characteristics of intelligence--language, speech recognition and understanding, common sense, emotions, planning, strategy, and lip reading--we are incapable of rendering even a rough facsimile of a HAL. "In all of the human-type problems, we've fallen far, far short," Stork says.

Even computer chess, in which seeming progress has been made, deceives. In 1997 IBM's Deep Blue beat then world champion Garry Kasparov. Deep Blue's victory, though, was more a triumph of raw processing power than a feat that heralded the onset of the age of the intelligent machine. Quantity had become quality, Kasparov said in describing Deep Blue's ability to analyze 200 million chess positions a second. In fact, Murray F. Campbell, one of Deep Blue's creators, notes in Hal's Legacy that although Kasparov, in an experiment, sometimes failed to distinguish between a move by Deep Blue and one of a human grandmaster, Deep Blue's overall chess style did not exhibit human qualities and therefore was not "intelligent." HAL, in contrast, played like a real person. The computer with the unblinking red eye seemed to intuit from the outset that its opponent, Discovery crewman Frank Poole, was a patzer, and so it adjusted its strategy accordingly. HAL would counter with a move that was not the best one possible, to draw Poole into a trap, unlike Deep Blue, which assumes that its opponent always makes the strongest move and therefore counters with an optimized parry.

The novel of 2001 explains how the HAL 9000 series developed out of work by Marvin Minsky of the Massachusetts Institute of Technology and another researcher in the 1980s that showed how "neural networks could be generated automatically--self-replicated--in accordance with an arbitrary learning program. Artificial brains could be grown by a process strikingly analogous to the development of the human brain." Ironically, Minsky, one of the pioneers of neural networks who was also an adviser to the filmmakers (and who almost got killed by a falling wrench on the set), says today that this approach should be relegated to a minor role in modeling intelligence, while criticizing the amount of research devoted to it. "There's only been a tiny bit of work on commonsense reasoning, and I could almost characterize the rest as various sorts of get-rich-quick schemes, like genetic algorithms [and neural networks] where you're hoping you won't have to figure anything out," Minsky says.ý

Comments