His Goal: Smarter Machine

Ellis Hennican, Newsday, 28 March 1999

There is intelligence, and then there is intelligence. Hugh Loebner has a foot in both worlds.

Inventor, art collector, philanthropist, businessman and 58-year-old gadfly-about-New York, Loebner has done more than almost anyone to calculate the differences between the two kinds of brains - artificial and real. But now he's doing something really smart: He is making sure he will be remembered after he's gone.

He will, anyway, if computers don't progress too quickly in the meantime.

Let me explain.

Among his various and eccentric pursuits, Hugh Gene Loebner is the Loebner behind the Loebner Prize, the premier international competition for artificial intelligence.

And right now, Hugh is sitting in the living room of the same sprawling Upper West Side apartment he grew up in. Which, you'll have to admit, is pretty smart right there. Just imagine the kind of rent-control deal he must have. Much of the wall space is covered with art. Sculptures and exquisite micro-mosaics rest here and there. And Hugh is leaning forward in a comfortable-looking chair, a bundle of bouncing energy.

"The idea of the contest is to write a computer program whose responses are indistinguishable from a human being's," he says. "This is known as the Turing Test."

British mathematician Alan Turing was a legendary code-breaker in World War II and a computer-science pioneer. In 1950, he postulated, to considerable public hand-wringing, that computers would one day acquire abilities that rival human intelligence.

But how will we know when they have?

Turing had a suggestion. Let a panel of judges type questions to a human being and a computer, both of them hidden behind an opaque screen. Then see if the judges can tell which is which.

If they can't, an important line has been crossed: Computers can "think" like humans.

It was Loebner's idea, in 1990, to hold such a contest every year - and award a prize to the programer whose computer scored best. He has underwritten most of the financial costs, with some support from the National Science Foundation, the Sloan Foundation, the Grumman Corp. and others.

But no computer has beaten the humans yet.

"Not even close," Hugh Loebner says "Not yet."

But the first one that does, he promises, will get the Grand Prize: $100,000 and an 18-carat gold medal. The year that happens, the Loebner Prize will never be awarded again.

This whole business of giving prizes, however, turns out to be far more complicated than anyone would have guessed, far more demanding than the mere pressure to keep writing checks.

Indeed, even the vaunted Loebner Prize has bounced around a bit in its search for an institutional host. In the early years, the prize was administered by the Cambridge Center of Behavioral Studies in Massachusetts. More recently, it's been hosted by Flinders University of South Australia. And in a one-time stop, Dartmouth College has it for the year 2000.

"But now, the big news," Loebner was saying up in his apartment at week's end, "We have assured the prize a permanent home, more permanent even than I."

The Loebner Prize is going to London. That city's prestigious Science Museum has agreed to host the contest for the next 44 years, unless the Grand Prize is won first, of course.

"And if I'm still here when the 44 years are up," Loebner said, "they'll host it for another 10. That ought to see me to the end."

But first, some down-to-earth business still has to get done.

Which is why, on this nice afternoon in the early spring of 1999, Hugh Loebner was getting up from his chair and walking his guest around the apartment, pointing out various pieces of art.

"Instead of putting my money in stocks and bonds, this is where it went," he said.

To Jeanette Pasin Sloan's huge oil painting of glistening bowls, "Revere Revere."

To Donald Rolle Wilson's precious "Beverly the Gorilla."

To several Thai Varick wire sculptures of biplanes, dragons and bulls.

"Christie's has agreed to take the micro-mosaics," Loebner said. "You know micro-mosaics? They're Victorian decorative art. There's a beautiful snuffbox. A tiny table. A bracelet. They'll auction all of them."

All in support of artificial intelligence.

"To fully fund the prize," its namesake patron said, "we need $125,000. Forty-five thousand is already accounted for. So I have to come up with another $80,000."

What better way to raise it than an art sale.

"I'm thinking of selling everything," he said. "Right down to the bare walls. I said to myself, `Why not start fresh for the new millennium? Why not start over again?' That seemed like an intelligent idea."

Of either type.