Elbot Fools Some People, Wins 2008 Loebner Prize
Writing by Evan Ackerman on Monday, 13 of October , 2008 at 1:07 am
Every year, the Loebner Prize is given out to the piece of artificial intelligence software (commonly called a chatbot) that is best able to make real humans think that it is human. The competition was inspired by the Turing test (named after Alan Turing) which basically says (and I’m paraphrasing for brevity) that if you can’t tell the difference between a machine and a human (intelligent human), you’re obligated to call the machine intelligent. This gets into all kinds of murky philosophical issues, but based on this year’s results, it’ll be a little while before we have to deal with all that.
The competition is simple: a panel of judges engages in a text chat on a split screen with both a computer program and a human. After five minutes, the judge has to decide which conversation was which. This year’s winner, Elbot, was able to fool three of twelve human judges into thinking that it was a real person, which if you think about it, is no mean feat, especially since it had to be more human than the real human chatting on the split screen.
You can try talking to Elbot for yourself here, although I think this version is a year or two less advanced than the version which was competing over the weekend. I still find it to be fairly intelligent, though… When I asked Elbot if it was sad that it only fooled three humans, it responded: “I would rather be an unsatisfied robot than a satisfied human.”
Sigh, me too.
[ Loebner Prize ]
Comments (1)
Category: Artificial Intelligence, Competitive
- Add this post to
- Del.icio.us -
- Digg
Comment by amichail
Made Tuesday, 14 of October , 2008 at 11:49 am
Check out this Web 2.0 approach to chatbots: http://chatbotgame.com.
Just as Deep Blue brute-forced it in chess with speed, the idea behind the Chatbot Game is to brute-force it with a huge number of user-submitted Google-like chat rules.