The Loebner prize rewards the machine that best imitates a human – it provides a great sanity check for wilder AI claims
When you chat to your mates online, are you sure they are not machines? If you’re not, maybe you’ve entered some Gibsonian cyberspace inhabited by an artificial intelligence (AI), or maybe you should think about getting new chat mates. Machines are pitted against humans every year for the Loebner prize to find out which AI program can best imitate a human being. That gets it a $4,000 prize. The ultimate aim is to create an AI program that is indistinguishable from a human being to claim a gold medal and $100,000.
Alan Turing set out this challenge in 1950 based on a Victorian “imitation game”, in which competitors had to work out if they were communicating with a hidden man or woman. Turing reckoned that, if a program could convince people it was a human, it was to all intents and purposes thinking. He predicted that machines would easily pass the Turing test by the year 2000, but none has even come close.
The Loebner prize exposes this failure and is boycotted by the mainstream AI community. Many AI gurus were on Loebner’s committee for the first year of the competition in 1991. However, the great fanfare of aspiration attracted massive media interest and the subsequent failure and disagreements with Hugh Loebner, the owner of Crown Industries, saw the committee resign en masse. Marvin Minsky, one of the few living founders of AI, calls it a publicity stunt and has offered a prize for its termination.
Read more . . .ᔥ
Bookmark this page for “artificial intelligence” and check back regularly as these articles update on a very frequent basis. The view is set to “news”. Try clicking on “video” and “2” for more articles.
