Machine talkers fail Turing test

Artificial intelligence took a step closer to becoming a reality on Sunday as machines edged closer to passing a key test.

This weekend saw artificial conversations entities (ACEs) compete for the 18th annual Loebner Prize, a US$100,000 jackpot for a machine that can fool judges into thinking it's human - a prize that has eluded entrants in the contest's 18-year history.

To clinch the prize, ACEs have to convince around a third of volunteers, a mixture made up mainly of computer scientists and journalists, that they are having a conversation with a person.

Every one of the five top ACEs tested at the University of Reading last week managed to convince at least one person into believing they were talking to a human rather than a machine.

But while one ACE managed to fool a quarter of the people during text chats, none managed to hit the crucial 30 per cent threshold laid down by British mathematician Alan Turing in 1950.

Turing said that a machine could be "attributed with intelligence" if it managed to have a "text-based conversation indistinguishable from a human".

Test organiser professor Kevin Warwick, from the University of Reading's school of systems engineering, said in a statement: "This has been a very exciting day with two of the machines getting very close to passing the Turing test for the first time.

"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.

"Today's results actually show a more complex story than a straight pass or fail by one machine."

The program Elbot, created by Fred Roberts, was named as the best machine in this year's event and was awarded the US$3,000 Loebner Bronze Award.

Speaking to Builder AU sister site silicon.com recently, prize founder Hugh Loebner told silicon.com he believes the Turing Test will eventually be passed.

"Oh I think so - I don't know about the timescale," he said.