Loebner Prize facts for kids
The Loebner Prize was a special competition in artificial intelligence (AI). It gave awards to computer programs that seemed the most human-like. The contest used a format called the Turing test. In this test, a human judge would chat by typing with both a computer program and a real person at the same time. The judge's job was to figure out which one was the computer and which was the human.
The competition started in 1990. Hugh Loebner created it with the Cambridge Center for Behavioral Studies in Massachusetts, United States. From 2014, the AISB organized it at Bletchley Park in the UK.
The Loebner Prize was also connected with other places like Flinders University, Dartmouth College, the Science Museum in London, University of Reading, and Ulster University. In 2004 and 2005, it was even held in Hugh Loebner's apartment in New York City. Some experts in AI, like Marvin Minsky, thought the prize was just for publicity and didn't really help the field of AI.
For the last competition in 2019, the rules changed. There were no special judges. Instead, the public judged the chatbots, and there were no human competitors. The prize stopped in 2020.
Contents
Prizes for Smart Programs
At first, the prize for the most human-like program was $2,000. It changed over the years, for example, it was $3,000 in 2005 and $2,250 in 2006. In 2008, it was $3,000 again.
There were also two big, special prizes that were never won.
- A $25,000 prize was offered for the first program that judges could not tell apart from a real human. This program also had to convince judges that the *human* was the computer program.
- A $100,000 prize was for the first program that judges could not tell from a real human in a Turing test that included understanding text, pictures, and sounds. The competition was planned to end once this big prize was won.
How the Contest Worked
The rules of the competition changed over the years. Early contests had limited conversations, but after 1995, people could talk about anything.
In 2007, some simple "screening questions" were used. These included questions about the time, general knowledge like "What is a hammer for?", and comparisons like "Which is faster, a train or a plane?". The programs also had to remember parts of the conversation. All words used were from a dictionary for children or teenagers under 12.
In 2008, web-based entries were allowed for the first time. Judges had limited time to talk to the programs. In 2003, they had 5 minutes per program. From 2008 to 2009, they had 5 minutes to talk to both a human and a program at the same time. This time increased to 25 minutes from 2010 onwards.
Why Some Didn't Like It
Many experts in AI didn't like the Loebner Prize for several reasons.
Some thought it was just a way to get attention. Marvin Minsky even jokingly offered a prize to anyone who could stop the competition. Hugh Loebner then joked that Minsky was helping to sponsor the event!
Critics also said the rules encouraged judges to make quick decisions. The conversations were often very short, sometimes only a few minutes. Also, early on, the conversations were limited to one topic chosen by the computer, which made it easier for programs to use simple tricks.
Programs in the competition often didn't try to truly understand things. Instead, they used basic tricks, like those used by early chatbot programs such as ELIZA. Being good at tricking people was often rewarded.
Past Contests
2006 Contest
In 2006, the contest was held in London. The winner was a program named 'Joan', which was part of Jabberwacky. Both were created by Rollo Carpenter.
2007 Contest
The 2007 competition took place in New York City. No computer program passed the Turing test that year. The winner was Robert Medeksza, who created Ultra Hal. He received $2,250.
2008 Contest
The 2008 competition was held in the UK. Five programs competed in the finals.
- Elbot, created by Fred Roberts, won the prize. It tricked three out of twelve judges into thinking it was human. This was very close to the 30% needed to say a program passed the Turing test.
- Eugene Goostman and Ultra Hal each tricked one judge.
A journalist from The Times, Will Pavia, was a judge and was fooled by Elbot and Eugene.
2009 Contest
The 2009 competition was held in Brighton, UK. The prize was $3,000. David Levy won that year.
2010 Contest
The 2010 competition was the 20th time the contest was held. It took place in Los Angeles, USA. The winner was Bruce Wilcox with his program Suzette.
2011 Contest
The 2011 competition was held in the United Kingdom. The prize was $4,000. Bruce Wilcox won again with his program Rosette. That year, a group of junior judges (kids!) also participated. Their results were different from the main judges.
2012 Contest
The 2012 competition was held at Bletchley Park in England. This was to celebrate 100 years since Alan Turing was born. The prize was $5,000. Mohan Embar won with his program Chip Vivant. For the first time, the conversations were shown live online.
2013 Contest
The 2013 competition was held in Northern Ireland, UK. The winner was Steve Worswick with his program Mitsuku. Mitsuku also tied for first place in the Junior Loebner Prize Competition that year.
2014 Contest
The 2014 competition was held at Bletchley Park. It was filmed live by Sky News. Bruce Wilcox won again with his program 'Rose'. He received $4,000 and a bronze medal.
2015 Contest
Bruce Wilcox won the 2015 Loebner Prize again with his program 'Rose'.
2016 Contest
The 2016 Loebner Prize was held at Bletchley Park. Mitsuku won first place.
2017 Contest
The 2017 Loebner Prize was also held at Bletchley Park. This was the first time a new way of sending messages was used. Mitsuku won first place again.
2018 Contest
The 2018 Loebner Prize was the last time it was held in its traditional Turing Test style at Bletchley Park. Mitsuku won first place for the third time in a row.
2019 Contest
The 2019 Loebner Prize was held at the University of Swansea. The contest changed a lot. Instead of a few judges, members of the public, including schoolchildren, could talk to the bots. They knew the bots were not human. Seventeen bots took part. Steve Worswick won for a record fifth time with Mitsuku. This earned him a spot in the Guinness Book of Records!
A special jury also voted for their favorite bots.
- Most humanlike chatbot:
- 1st place: Mitsuku
- 2nd place: Uberbot
- Best overall chatbot:
- 1st place: Mitsuku
- 2nd place: Uberbot
Winners of the Loebner Prize
Year | Winner | Program |
---|---|---|
1991 | Joseph Weintraub | "Whimsical Conversation" (PC Therapist) |
1992 | Joseph Weintraub | PC Therapist |
1993 | Joseph Weintraub | PC Therapist |
1994 | Thomas Whalen | TIPS |
1995 | Joseph Weintraub | PC Therapist |
1996 | Jason Hutchens | HeX |
1997 | David Levy | Converse |
1998 | Robby Garner | Albert One |
1999 | Robby Garner | Albert One |
2000 | Richard Wallace | Artificial Linguistic Internet Computer Entity (A.L.I.C.E.) |
2001 | Richard Wallace | Artificial Linguistic Internet Computer Entity (A.L.I.C.E.) |
2002 | Kevin Copple | Ella |
2003 | Juergen Pirner | Jabberwock |
2004 | Richard Wallace | Artificial Linguistic Internet Computer Entity (A.L.I.C.E.) |
2005 | Rollo Carpenter | George (Jabberwacky) |
2006 | Rollo Carpenter | Joan (Jabberwacky) |
2007 | Robert Medeksza | Ultra Hal |
2008 | Fred Roberts | Elbot |
2009 | David Levy | Do-Much-More |
2010 | Bruce Wilcox | Suzette |
2011 | Bruce Wilcox | Rosette |
2012 | Mohan Embar | Chip Vivant |
2013 | Steve Worswick | Mitsuku |
2014 | Bruce Wilcox | Rose |
2015 | Bruce Wilcox | Rose |
2016 | Steve Worswick | Mitsuku |
2017 | Steve Worswick | Mitsuku |
2018 | Steve Worswick | Mitsuku |
2019 | Steve Worswick | Mitsuku |
See also
In Spanish: Premio Loebner para niños
- List of computer science awards
- Artificial intelligence
- Glossary of artificial intelligence
- Robot
- Artificial general intelligence
- Confederate effect
- Computer game bot Turing Test