Alan Turing was the British mathematician, cryptographer and prototypical computer scientist who, some say, did as much as Winston Churchill to save Western civilization from the Nazis. He was also perhaps the most influential thinker of the 20th century -- in the sense that his ideas are the foundation upon which is built computer science and thus our entire digital world. He started the modern discussion about minds, machines and artificial intelligence. Turing was gay and did not hide it; for this he was shunned by his peers and forced by the government to undergo hormone "therapy." He killed himself in 1954, and it is widely believed that he did so to escape persecution over his sexual preference.
The Turing test is the canonical benchmark by which we humans will know that computers have caught up with us in the smarts department. In 1950 -- barely two years after the construction of the first machine that could reasonably be called a computer -- Alan Turing proposed a simple test to determine whether machines could think. If you were conversing with an entity and you could not tell whether that entity was human or merely human-made, then whatever you were conversing with was at least as intelligent as you were. Turing predicted that computers would pass this test by the year 2000.
Long known to historians of the computer, the Turing test emerged from obscurity and became part of popular culture in 1966, when Joseph Weizenbaum's simple 200-line Eliza program, which used a few simple tricks to generate bland responses to human-posed questions, fooled people into thinking they were conversing with an intelligent being.
Although the challenge was at first embraced by the academic A.I. community, passing the Turing test -- which proved to be a rather more difficult nut to crack than some prominent A.I. people had said it would be -- has long since fallen out of fashion as a legitimate goal or benchmark among "real" A.I. researchers. The A.I. establishment has for more than a decade put more energy into explaining why the Turing test is irrelevant than it has into passing it.
The Loebner competition is designed to provide a real-world opportunity for bots to pass the Turing test. And even though the academics are loudly boycotting the Loebner competition today, that doesn't mean it wants for entrants.
There are three categories in the Loebner Prize. The grand prize of $100,000 and a gold medal bearing the likeness of Hugh Loebner will go to the first non-human entity that passes what Loebner calls the "full" Turing test (although the criteria for what constitutes "full" are a little vague.) A silver medallion and $25,000 will go to the winner of a somewhat more rigorously defined "limited" Turing test. And a bronze medallion and $2,000 are awarded annually to whichever program is judged best according to the rules in force at that time.
Contestants have included two-time winner Robby Garner, a self-taught computer programmer and robot builder; Benji Adams, the creator of the Web site Personality Forge, home of hundreds of chatterbot "personalities"; and perhaps the most famous (or infamous) of them all, two-time winner Dr. Richard Wallace, founder of the ALICE Foundation. Dr. Wallace, who, by his own admission suffers from a debilitating mental illness -- which he mitigates with daily doses of medicinal marijuana -- is known not only for his unorthodox theories about consciousness and intelligence but also for the restraining order that keeps him off the University of California campus in Berkeley. Something about the Loebner Prize seems to draw eccentrics out of the woodwork, and chaos itself is the very essence of the annual ritual.
The 2002 contest, hosted in Atlanta by the Institute of Mimetic Science, had the most participants ever. More than 40 bots were entered, from which, by some apparently occult process, a final field of eight was selected.
The event itself, by most accounts, was conducted with all the dignity and pomp of the finale of the Marx Brothers' "Duck Soup," although it perhaps lacked some of Freedonia's essential seriousness. Every person with whom I spoke about it said that last year's contest was an utter fiasco, with unclear rules, inconsistent judging, arbitrary fiats by an opaque prize committee, petulant prima donnas, and last-minute changes of venue that prevented most entrants from even discovering where the contest was taking place until after it had happened.
There are strongly divergent views about the cause of those problems, and in asking some of the principals a few questions about it I soon felt that I had set myself up as Judge Wapner, with each of the aggrieved parties anxious to get their story out to the world. Benji Adams was the most forthcoming; his 26-page memo, in its righteousness and obsessive attention to every slight -- replete with chronologies and an e-mail trail documenting the alleged incompetence or malfeasance of the Loebner competition committee -- reminded me of nothing so much as Kenneth Starr's report to Congress.
By tradition, three things happen at the conclusion of every Loebner contest: The winners take their prizes and run for the nearest exit, Hugh Loebner basks in glory, and the hosting organization takes a solemn oath: "Never again."
Over the 11 years of its existence the Loebner competition has taken place in nine locations, including the Boston Computer Museum, the London Museum of Science, Dartmouth College, Flinders University of South Australia and, during a particularly dark period in the prize's short history, in the billiard room of the private Salmagundi Club in New York City, of which Loebner is a member.
Despite the differing locations, the contest has always been conducted under the auspices of the Cambridge Center for Behavioral Studies, an obscure Massachusetts-based nonprofit that apparently doesn't do much beyond running the Loebner competition and yet, as we'll see, would dearly love to have nothing to do with the contest at all.
There's no waiting list of prestigious institutions clamoring for the honor of physically hosting the contest these days. Rather, the Loebner competition has proven so stressful to its host that it's seldom invited back, and each year the Cambridge Center must scramble to find a new home for it, as a social service agency might scramble to find a foster home for a troublesome adolescent. Three years ago the Cambridge Center seemed to have found a permanent home for its problem child: With great fanfare, the London Museum of Science and the Cambridge Center jointly announced that the museum would be home to the Loebner competition for the next 44 years. After one dose of the Loebner competition, however, the museum had changed its mind and annulled the agreement. Once again the Cambridge Center was shopping for a suitable foster home.
Every year finding another institution to host the Loebner competition becomes a little harder; "selling" the Loebner has become, for the Cambridge Center, a job akin to selling the Brooklyn Bridge. And yet each year, one way or another, Loebner himself has made sure that the show goes on.
Today, however, Hugh Loebner is on the brink of initiating legal action that may finally do what the combined efforts of his detractors have been unable to do for a decade, which is to make him take his money and go home. For years Loebner and the Cambridge Center for Behavioral Studies have been stuck in a dysfunctional marriage of sorts. But now Loebner and the center are flirting with divorce, and it may be a messy one. If they do wind up in court, which Loebner tells me is likely, it's hard to see how the prize will survive.
In fact, although the University of Surrey (U.K.) has announced that it will host the 2003 competition next October, there's always a possibility that it will back out. The last Loebner bronze medallion may already have been won, and the silver and gold medallions may wind up for sale on eBay before anyone has earned them.
It wasn't always like this.
Next page: A blue-ribbon panel, a gala event, and a winning program as dumb as a bag of hammers