Artificial intelligence gets a step closer

Elbot managed the most successful score yet in the annual Turing Test

Artificial intelligence came another step closer to reality this weekend after a computer came within five per cent of passing the Turing Test which evaluates a system's ability to demonstrate intelligence.

The Turing Test is named after mathematician Alan Turing whose 1950 paper Computing Machinery and Intelligence stated that, if enough people cannot reliably differentiate between a human and a machine during a natural language conversation, the machine can be considered intelligent.

No machine has yet managed to deceive the 30 per cent of interrogators required to pass the Turing Test.

However, at this weekend's annual Loebner Prize competition at the University of Reading, one system, dubbed Elbot, managed the most successful score yet, fooling 25 per cent of the judges.

In this year's test, five computer systems were pitted against five judges who were each given five minutes of unrestricted conversation through a terminal to decide which of the entities they were talking to was a human and which was a machine.

The Loebner Prize was created by American businessman Hugh Loebner in 1990 together with the Cambridge Centre for Behavioural Studies, and is an annual competition offering a grand prize of $100,000 (£58,000) and a solid gold medal to the first machine to crack the Turing Test.

Although no machine has yet won the grand prize, each year $2,000 (£1,150) and a bronze medal is awarded to the best entrant.

"Although the machines aren't yet good enough to fool all of the people all of the time, they are certainly at the stage of fooling some of the people some of the time," said Professor Kevin Warwick, of the School of Systems Engineering at the University of Reading, and organiser of this year's test.

"Today's results actually show a more complex story than a straight pass or fail by one machine. Where the machines were identified correctly by the human interrogators as machines, the conversational abilities of each machine was scored at 80 and 90 per cent."

Warwick believes this is a clear indication that computers are getting increasingly good at communicating with humans in a natural and comfortable way, slowly narrowing the divide between man and machine.