When Hugh Loebner created the Loebner Prize in 1989 to spur progress toward technology that could pass the Turing test, the A.I. establishment welcomed him with open arms. Under the aegis of the Cambridge Center, a blue-ribbon panel chaired by cognitive scientist and author Daniel Dennett and composed of a who's who of computer scientists from leading institutions, organized the first contest and wrote the first rules. When, after two years of planning, the first event was held in 1991 at Boston's Computer Museum, it was a gala affair partially underwritten by the National Science Foundation and Sloan Foundation.

Hype had been building for months, and when the actual contest finally took place, reporters from the popular and scientific press thronged the hall. Robert Epstein, director emeritus of the Cambridge Center and a former subordinate of Loebner's at University of Maryland in Baltimore County (UMBC) acted as master of ceremonies. When you consider that Loebner's degrees are in sociology, not computer or cognitive science, and that before his return to the family business his only employment in the groves of academe was as an assistant director of the statistics center at UMBC, you see what a stunning coup he had pulled off. It was as if the goofiest nerd in the entire high school had wormed his way into the "A" clique, the one led by the rich, handsome quarterback and the beauty queen/state tennis champ.

There was only one problem: the A.I. programs entered in the contest performed horribly. They were pathetic. The mountain, as it were, had gone into labor and given birth to a mouse.

"Perhaps the most conspicuous characteristic of the six computer programs was their poor performance." This damning summary of the first event appears in the article "Lessons From a Restricted Turing Test," by Harvard professor Stuart Shieber, in the Communications of the Association for Computing Machinery (CACM), the house organ for the computer science establishment. The winning program, by Joseph Weintraub, was a very simple variant of Joseph Weizenbaum's 25-year-old Eliza program. Worse, the article reported, "Dr. Epstein, in a speech after the event, noted that he had learned from the day's proceedings that 'little progress has been made in the last twenty-five years.'"

It had to be said, and Epstein said it: The emperor had no clothes. After decades of government-funded research by the brightest minds in computer science, A.I. programs still stank, and the National Science Foundation and Sloan Foundation had just spent $80,000 to demonstrate this sad fact to the world. Now what?

The answer, according to Shieber, was obvious: stop running the contest! The world wasn't ready for a Loebner Prize, he said, so having one was counterproductive.

The wheels of the ACM grind exceedingly fine, but they grind slowly, and Shieber's report on the 1991 Loebner contest did not appear until the spring of 1994. In the meanwhile Dennett's committee gamely soldiered on and conducted two more contests (attended by considerably less hype and similarly unimpressive results). But once Shieber's article appeared, the world had changed. The Party had spoken, saying, in effect, "Loebner is making us look bad. Make him go away (but see if you can find a way to keep the money)." Dennett and other members of the committee tried to get Loebner to change his contest along lines suggested by Shieber, but he refused. Shortly thereafter Dennett and his committee resigned en masse, angry at Loebner's intransigence and generally sick of dealing with him. Dennett is still bitter about this episode.

Actually, the themes "angry at Loebner" and "generally sick of dealing with him," came up a lot in my research for this article, although not many people were willing to use those phrases for attribution. For example, I discovered that sometime before Dennett's resignation, Loebner and his old friend Epstein had ceased to be on speaking terms. Epstein was the first of many people associated with the contest who are no longer on speaking terms with Hugh Loebner. Even his friends and fans find him, well, hard to take in anything other than small doses. In the words of Robby Garner -- a two-time Loebner Bronze winner, and currently a member of the 2003 contest committee, "Loebner was criticized early on for his self-aggrandizement, and he has hardly disappointed any of those detractors if they have ever spent any time or shared a meal with the man."

Who is this man who seems to attract controversy the way picnics attract ants?

- - - - - - - - - -

I first heard of Hugh Loebner from a friend of mine who had attended a conference in the fall of 2002 for users of the Lisp programming language, the language favored for most A.I. work. He told me about Loebner, about his prize, and about what it was like to have a conversation with him. "Loebner's one of those guys who's always trying to catch up with his own thoughts," my friend said. "You can almost see him running after them." And then he added, "It's not clear whether going through life both very intelligent and very ignorant is a wise strategy, but that seems to be the one he has chosen."

I was intrigued by my friend's story about the Loebner competition, because last August I published a novella, called "Cheap Complex Devices," that purported to comprise the chronicle of, and the results from, a storytelling competition between researchers vying for the inaugural "Hofstadter Prize for Machine-Written Narrative." The contest was named after Douglas Hofstadter, the Peter Pan-ish author of the Pulitzer Prize-winning favorite book of every freshman student of cognitive science: "Gödel, Escher, Bach: An Eternal Golden Braid."

Like Hofstadter's, my book is about minds and machines, and it is basically incomprehensible (at least to me), but unlike his book, mine at least has skulduggery, sex, egomaniacal and possibly delusional computer scientists, pathos, a bitterly feuding prize committee, and a whore with a heart of gold. I really like my book, and I'm glad I wrote it. So I am quite happy that I had never heard of Hugh Loebner, or of the Loebner Prize for Artificial Intelligence, before I undertook to write "Cheap Complex Devices," for had I known then what I know now, I never would have bothered. The story had already been done!

Among other things, "Cheap Complex Devices" is a lampoon of the technology cult of MIT's A.I. Laboratory (co-founded by Marvin Minsky) and Media Lab (with which Minsky is closely associated) and places like them. Being a techno-paranoid humanist, I find techno-utopian trans-humanists of the Minsky ilk an affront. I desire to live in a world without future shock, and they want to shock me every chance they get.

I loathe "progress"; they worship it. But, paradoxically, I find myself annoyed by the way the A.I. elite has, for decades, promised far more than it has delivered. So I'm annoyed with self-promoting A.I. visionaries for moving both too fast and too slow. Or maybe I'm just touchy because as a member of the usability engineering group at Sun Microsystems in the early '90s I was a frequent participant at conferences where Media Lab triumphalism was as predictable as sunrise. In any event I felt like giving them a pie in the face.

Which is to say that although Loebner wants to live in the 22nd century and I want to live in the 19th, both of us seem to have a thing about tweaking the noses of the A.I. establishment. I was predisposed to like the fellow.

Recent Stories