IBM is to divert a massive chunk of its $5bn (£3.5bn) annual research budget into an artificial intelligence project called "Eliza". The project, described as "multi-billion, multi-year", won the name because its aim is to create a computer with the intelligence equivalent to that of a lizard within a few years. Longer-term targets are to challenge human intelligence levels by making a computer with the processing power of the human nervous system.
Eliza will draw on IBM research, from the world's national laboratories and from the wider IT research community through academic grants. Industry estimates the whole scheme is probably worth more than $2bn.
The project is part of an emerging area of IT development known as "autonomic computing", which attempts to improve the performance of computers and networks by emulating functions that appear in the natural world. IBM intends to send 75,000 copies of its "autonomic manifesto" to the best technical minds in the world in an effort to involve as many researchers as possible.
The fundamental idea is that computers will become vastly more efficient if they are given the "intelligence" to perform a range of tasks without being told to do so.
Paul Horn, senior vice-president of IBM Research, says: "Consider the autonomic nervous system: it tells your heart how many times to beat and checks your blood's sugar and oxygen levels. It monitors your temperature and adjusts your blood flow and skin functions to keep it at 98.6 degrees. But most significantly, it does all this without any conscious recognition on your part."
A spokesperson for IBM says what has held back artificial intelligence projects is that they have focused too tightly on building "thinking machines", akin to the robot in the new Stephen Spielberg film AI, and by emulating things such as touch and sight, have concentrated on the wrong functions of the body.
But the Eliza project has not been set up purely to prove a point. Mr Horn believes autonomic computing is the IT industry's best hope against a major threat it faces: complexity. Over the past two decades, he says, the IT industry has been so obsessed with building smaller, faster and cheaper machines, that it has not stopped to think about making them easier to use. In short, computers are becoming too complicated for everyone's good. By building a computer that regulates itself, repairs itself, anticipates problems and does it all in an intelligent way, users will be able to get far more from their platforms.
Some of IBM's early developments in the autonomic field are encouraging. The company has developed a server that does not bother its operator until all possible solutions to a problem have been tried. If it finds itself stuck, it calls its human master on the phone for help.
Though this might send a chill down the spine of anyone worried about machines taking over the world, Mr Horn's vision is firm. "It's time to design and build computing systems capable of running themselves, adjusting to varying circumstances, and preparing their resources to handle most efficiently the workloads put upon them."Reuse content