Artificial stupidity, Part 2

Hugh Loebner likes to compare himself to King Lear, "more sinned against than sinning." All he wanted to do was to give away a fortune in order to hasten the day when human toil would be abolished and people could devote their lives to pleasure, and look where it's gotten him. It's taken years off his life, made him the object of ridicule, and cost him a fortune. And now, on top of everything else, he's probably going to have to take the Cambridge Center to court. "Litigation is likely," he told me.

But why sue, I asked him, when the Cambridge Center clearly wants nothing more than to give you back your money and your prize and get out of its hair?

"Generally, the Center has been making noises about being willing to return my money," Loebner replied. "I have not yet responded to the last letter from Farrow (I am composing a reply which will include much of what follows), but I will tell them that I will not simply accept my $125k back. If they want to return the money, I want $375K -- $125K for the prize money, $250K to cover the 13+ years of expenses, time and effort, lost interest, as well as the costs of establishing an alternative foundation to oversee the prize."

To me this seems quixotic. As the Cambridge Center's attorney Farrow put it, "the legal context of the Loebner Prize is one of a completed gift, not an ongoing contract." I cannot imagine Loebner prevailing in court. Certainly if I were on the jury I would say, "Give him his money back. The end."

Indeed it is hard for me to think of Hugh Loebner as King Lear, but it is easy to think of him as Don Quixote. The real Don Quixote, of course, the man in the novel, not in the sentimentalized TV renditions, was not exactly harmless. Don Quixote's disputatiousness was not quaint; he sometimes beat people -- usually friends or innocent bystanders -- to within an inch of their lives. Likewise, Loebner's most abused victims are usually the ones most sympathetic to his contest.

Here's a typical story, this one from Robby Garner, a two-time winner of the bronze Loebner Prize and a member of the 2003 competition committee.

"The company which hosted the 2002 contest is called the Institute of Mimetic Sciences. We initially wanted to carry out the competition live on the Internet, and CCBS was in agreement with us that it would make the most sense since the overwhelming majority of chatterbots now are Web-based software. I was given the task of asking Hugh about it. Hugh went ballistic when I did, and kept repeating his objections to me on the phone for about 15 awkward minutes. After that call, IMS and CCBS came to a compromise that we would still admit Web-based software, but they would have to run 'on site.'"

Don Quixote had a hard time dealing with "consensus reality," and so does Hugh Loebner. Certainly his advocacy of the rights of sex workers and their clients has made him unacceptable to a certain class of sponsors. But I'm not talking about that. That particular political stance, it seems to me, is not standing between Loebner and the kind of success that he wishes for his prize.

Rather it's his willful ignorance of the very technology he's trying to promote -- and the way he insists on micromanaging a contest that clearly would be better off if he were nowhere near it -- that threaten the continuation of the very thing to which he's devoted so much of his life's work.

I myself, in a prior life, managed a usability test when Sun Microsystems switched its underlying Unix technology. Organizing that test was comparable in scope to what should have happened, and didn't happen, in Atlanta last year. It took me four months working full time -- and I had the resources of a world-leading technology company upon which to rely. I know a thing or two about the logistics of these things. When Loebner told me that he could manage the competition himself, from his apartment, part time, I knew that I was talking to a man who was "going through life both very smart and very ignorant."

Don Quixote, of course, is beloved because his adventures are funny, and there is the funny side of Loebner's bluster, too: his way of provoking pie fights wherever he goes.

"I was caught in a bitch fight between Loebner and Minsky," recalled Neil Bishop. "We wanted to recognize Minsky for his work in the field on decision sciences. We know of the past baggage between the two, so I contacted Minsky to request permission to do so. I think he was flattered in some weird way by this request and ultimately gave us permission but not before blasting me for working with Loebner and wanting me to pass on to Loebner that Minsky would be contacting his lawyer to begin a libel and defamation action if his name was not removed from Loebner.net immediately."

I think that's too sweet for words. Not only that, but Bishop himself proposes to join the pie fight:

"Anyway, here is a tidbit for you. We are presently working to put together our own Turing event that will embrace the integrity of the first event that Professor Minsky competed in during the '70s. The prize? The Minsky Award for Intelligent Systems. The winner will receive a grant to support their research. Many details have be already accomplished to bring this about. If you want to stir things up, you can put this in your article."

When you consider that that Bishop organized the most recent Loebner Prize, you can get some kind of idea just how badly the 2002 contest went.

Richard Wallace seemed almost glad that he didn't win last year's contest. He was among several people who told me about Loebner's forceful statement -- it went on for a while, evidently -- that Ariel Sharon, and not Osama bin Laden, was behind the attacks of Sept. 11. In the memories of many of last year's Loebner contest participants, the bin Laden incident stands out more than anything having to do with Alan Turing.

When you add that to all the other silliness, what you come up with is an event far from dignified. I wasn't there, but I get the impression of a swirling chaos with enough vanity and stateless conversations to make artificial chatterbots totally redundant. The more you look at the actual event, the more apt seems Minsky's phrase "obnoxious and unproductive annual publicity campaign."

"I'm not used to being perceived as the most sober participant," Wallace told me, sounding apologetic.

And yet for all this meshugas, I find a nobility in Hugh Loebner. I respect his standing up for prostitutes, surely a devalued class of persons, and I applaud his making explicit the parallels between their persecution and Turing's. I admire the way he has welcomed all comers regardless of pedigree, the way he has stuck to the common sense of the Turing test in repudiation of those who would make it an exercise in "school figures"; I salute the meritocracy he has championed of hackers, free thinkers, eccentrics, and cheezo hobbyists. I think he was right to stick to his guns and insist that the competition be held, and a prize awarded, every year.

And I like his joie de vivre and his ability to laugh at himself (a little of which trait goes a long way, I might cautiously mention to Drs. Minsky and Dennett). For all his talent to drive one to distraction, Hugh Loebner, self-aggrandizing fool though he may be, set out on this enterprise, as did Don Quixote, to help us find something better within us.

In the conclusion to his response to Shieber, Loebner wrote,

"There is a nobility in this endeavor. If we humans can succeed in developing an artificial intellect it will be a measure of the scope of our intellect ... I suggest Loebner's Corollary to Asimov's Laws of Robotics: 'Humans are gods.'"

But I think the last word must go to Dr. Wallace, who improbably enough is perhaps the closest thing in this tale to a Sancho Panza.

"And remember," Wallace wrote in his Slashdot interview, "no one has proved that our intelligence is a successful adaptation, over the long term. It remains to be seen if the human brain is powerful enough to solve the problems it has created."

Story finder (3 ways to search Salon)

Search terms:

Salon the Web

Salon Directory (browse by topic)