From jboss@nunic.nu.edu Mon Oct 4 21:07:30 EDT 1993 Article: 19015 of comp.ai Xref: honeydew.srv.cs.cmu.edu comp.ai:19015 Path: honeydew.srv.cs.cmu.edu!bb3.andrew.cmu.edu!news.sei.cmu.edu!cis.ohio-state.edu!magnus.acs.ohio-state.edu!math.ohio-state.edu!howland.reston.ans.net!spool.mu.edu!olivea!pagesat!news.cerf.net!nunic.nu.edu!nunic.nu.edu!not-for-mail From: jboss@nunic.nu.edu (Jim Boss) Newsgroups: comp.ai Subject: Loebner Prize Competition on Artificial Intellegence Date: 29 Sep 1993 09:42:34 -0700 Organization: National University Lines: 54 Distribution: world Message-ID: <28cdtq$9ko@nunic.nu.edu> NNTP-Posting-Host: nunic.nu.edu PLEASE POST SEPTEMBER & OCTOBER, 1993 PLEASE POST CALL FOR SUBMISSIONS: TURING TEST, YEAR III $100,000 LOEBNER PRIZE IN ARTIFICIAL INTELLIGENCE NEW YORK COMPUTER BUFF WINS 1992 CONTEST Humans were pitted against computers once again on December 15th, 1992 in the second annual search for the world's first "thinking" computer. In the contest, judges converse at computer terminals to try to determine which terminals are controlled by people and which by computers. The 1991 winning program, Joseph Weintraub's "whimsical conversation," fooled 5 of 10 judges into thinking it was a person. His 1992 entry, "Men vs. Women," competing against entries from around the country, won the December event by a narrow margin. The contest, named after benefactor Dr. Hugh G. Loebner of New York City, is administered by the Cambridge Center for Behavioral Studies and will be hosted in 1993 by National University in San Diego. Dr. Robert Epstein, a professor at both Boston University and National University, is in charge of the contest. Planning is handled by a committee of scientists and scholars from MIT, Harvard, and other universities, chaired by distinguished philosopher Daniel C. Dennett of Tufts University. Dr. Daniel Bobrow of Xerox serves as liaison to AAAI. The contest was inspired by a paper published in 1950 by the brilliant English mathematician, Alan Turing, one of the fathers of the modern computer. The tests so far are restricted, requiring computers to be conversant on only one topic. The designer of the 1992 winning entry received a bronze medal and a check for $2,000 from the Cambridge Center for Behavioral Studies. $100,000 will be awarded when the winning program can pass an unrestricted test. When that happens, will we have the right to turn off the winning entry? Who should get the prize, the designer or the computer? Could the winning program be considered "intelligent" or "self-aware" or "conscious"? At the 1992 event, Weintraub's program fooled two of eight judges into thinking it was a person. But the human "confederates" -- hidden humans controlling some of the computer terminals -- scored much higher on the average than all of the computers. Commented Epstein, "In natural conversation, computers are still vastly inferior to people. The grand prize is well out of reach at the moment." For further information: Complete transcripts and IBM-compatible diskettes that play the 1991 and 1992 conversations in real-time may be ordered from Rachael Thompson at the Cambridge Center for Behavioral Studies at 617-491- 9020 (fax: 1072). Deadline for the receipt of 1993 submissions is November 1st. If you have already submitted an entry this year, please resubmit at this time. Because of a fire that razed the Cambridge Center for Behavioral Studies in July, most of the contest records were lost. If you want to participate this year, you must submit an updated application form. The 1993 contest will be held in San Diego on Wednesday, December 8th, 1993 (participation is by modem). Application materials may be obtained from the Cambridge Center (617-491-9020, fax: 1072), or National University (619-563- 7144). Refer technical questions to Dr. Robert Epstein, (619-436-4400 [CA],