Ask any stranger, 'Do you have the time?' and they look at their watch. Not many of us would be fazed by that request, or the reply 'Time for what?'

Our brains cope with understanding and responding to such ambiguities of communication, while computers, so far, do not. But one day they will, thanks to artificial intelligence pioneer Doug Lenat.

At the Austin, Texas offices of Cycorp, Lenat and his team have been working on machines that are smart, in the way that humans using common sense are smart.

Machines may one day overtake human intelligence levels because, with the ability to accelerate their own education, there is no limit to how smart they could become.

A youthful looking 52 year-old, Lenat has been described as a Tolkeinesque figure, a Bilbo Baggins-like character whose chubby body doesn't seem to be in proportion.

He's an amiable chap, although not the sort of hi-tech chief executive that has bean bags and table football in all of his offices for the staff.

When I remark on how quiet the Cycorp offices are, despite housing 60 "ontological engineers and blue-collar philosophers", Lenat replies: "Of course. People are at work."

He's also a serious scientist, which makes you wonder how he coped with the US education system, where anyone intelligent or geeky is bullied.

Born in Philadelphia, his family moved to Delaware for some years before returning to his birthplace, where the school district put him into what he calls the "thug or hood classes".

Not because he wasn't bright, but because the powers that be segregated students by proven ability and, since the Lenats had moved so often, he was an unknown quantity.

"I worked much harder in high school than in college because in non-advanced classes the teachers have preconceived notions about how few students were expected to get A grades," he recalls.

Lenat gained a PhD in computer science in 1976. He thought about becoming a mathematician or a physicist, but neither career attracted him.

He then pondered microbiology or artificial intelligence but, in the latter, found that "it was clear that researchers in the field didn't know what they were doing".

So artificial intelligence provided the perfect platform for a man who once said: "How many people have in their lives a two to 10 per cent chance of dramatically affecting the way the world works? When one of those chances comes along, you should take it."

Lenat's contribution to the world is a program called Cyc (as in 'en-cyc-lopaedia'), said to be the world's largest extra-sentient body of common sense and perhaps, one day, this planet's first digital consciousness.

How does he define 'Cyc'?
"We are capturing human consensus reality, the millions of pieces of knowledge that humans assume someone they've just met knows about the world that allows us to be so terse and ambiguous when we communicate with each other. It's the absence of that knowledge which makes computer programs so brittle," he explained.

"Cyc contains knowledge about concepts, rather than just words. So, if you're interested in Turkey it asks if you're interested in the country or the bird. And if it's the latter, whether you wanted recipes for cooking it, or ways to raise live ones.

"Currently, all web searching does is find relevant documents from which you might be able to extract your answer, but sometimes it doesn't do that because it doesn't understand what you're asking. We're talking about simple reasoning with a bit of understanding behind it."

Lenat's team has had to map out a world for Cyc that we take for granted. For example, when you open a bottle it's good to have the neck pointing upwards, or if you want to talk to people it's not a good idea to phone them at night when they're likely to be asleep.

"I would be surprised if you could name something that Cyc doesn't already know, and it knows millions of things. I'm not talking about proper nouns, but about generic concepts," he said.

This is why his team pores over endless pieces of text, be they newspaper stories, encyclopaedias, novels or advertisements, asking: "What did the writer assume that the reader knows about the world which, had they explicitly stated it, would have insulted or confused the reader?"

These 'missing links' represent the formalised common sense which Cyc needs to draw the conclusions that you and I would from the same body of information.

In the time it has taken Cyc to reach completion of its knowledge, it has made some amusing assumptions. It once concluded that everyone born before 1900 was famous, because all the people it knew about who lived then were famous.

That kind of 'cuteness' occurs infrequently in fictional references to super computers. Isn't Lenat concerned that he'll create a computer such as HAL, which killed the crew in Stanley Kubrick's 1968 film, 2001: A Space Odyssey?

Apparently not. It's that kind of brittleness which Lenat claims that he and his team are trying to prevent, by introducing "formalised common sense" to artificial intelligence.

"HAL killed the crew because it was given two contradictory commands. One was never to lie to the crew, the other was to lie to them about their mission," he explained.

"HAL found a solution by killing the crew, because then it wouldn't have to lie to them. Unfortunately it didn't know that it's better to lie to someone than to murder them."

It's this lack of common sense that worries Lenat, with respect to firms designing robots that may be able to "mow your lawn and watch your baby".

"Think about what kind of person you'd allow to be a nanny to your child. You wouldn't want an idiot savant who only knows a small number of narrow areas and might mow the baby, while watching the lawn," he said.

"It's all well and good researching robot bodies and hand-eye co-ordination, but people shouldn't trust robots unless they have common sense. We're investing more in automated systems with power over our lives, and that's an accident waiting to happen if such systems aren't intelligent."

Speaking of babies, does Lenat think of himself as the father of Cyc, given his references to 'he' rather than 'it'?

"I don't feel emotionally attached to the system," shrugs Lenat. "It's not so much that I want my 'baby' to be successful, but to live to see the world change because of it. I want to live in a fairyland where every object you meet can hold an intelligent conversation with you."

Why would that be a good thing? "So that people in the future will benefit from more effective communication and access to information and become more powerful," he suggested.

"By using this sort of mental amplification for advice services, direct marketing, database cleaning, corporate knowledge assets management and a host of other applications, people will be more powerful and potent in the world than we are today."

FURTHER READING

A downloadable, open source, Beta version of Cyc technology, called OpenCyc, is available here. The company website can be found here.

ALTERNATIVE CV: Doug Lenat
Which historical character do you most admire and why?
Isaac Newton. He wanted to change the world and wasn't afraid of long detours to develop the tools he needed - such as calculus - to proceed toward his goal.

How did you manage to get Bill Gates listed as a personal reference?
Bill first contacted me in the mid-1980s. He was one of the first strong supporters of Cyc, and I was part of the team that helped him set up and staff Microsoft Research.

What's the most satisfying thing about what you do?
In the short term, when the system surprises me by correctly deducing something I hadn't thought of. In the long term, the feeling that we are working together on a project that can change the way the world works.

What's the least enjoyable facet of running a business?
Having to tack against the funding winds, balancing the long-term needs of Cyc, to move forward in the right direction, against the short-term needs of paying customers. I dream of the day when our work has reaped enough profits to provide a permanent endowment for development.

What would an enemy say about you?
That I don't embrace their rigid ideology. I'd like to believe that I have a sort of kitchen sink ideology, willing to include practically anything if it works. This offends anyone who happens to worship that particular item. This includes technologies such as chaos theory, neural nets, statistics and so on, and institutions such as academia.

When do you plan to retire, and where to?
Never and nowhere.