On May 15, 2012 the annual Loebner Prize contest was held at the University of Reading, England. It is a contest held each year that awards prizes to the computer program considered by the judges to be the most human-like. During the contest, a human judge is faced with two computer screens. One is controlled by a computer, the other is controlled by a human being. The judge types questions to the two screens and receives answers back. The judge gets five minutes with each screen. Based upon how the conversations go, the judge must decide which screen is being controlled by a person, and which one is being controlled by the computer.
This sort of test is referred to as a “Turing test.” Alan Turing was an English mathematician, logician and cryptographer. Turing is often considered to be the father of modern computer science. He was born a hundred years ago today, on June 23, 1912.
During World War II Turing worked at Bletchley Park, the British code breaking center. He ultimately wound up heading the section responsible for German naval cryptanalysis and devised a number of techniques for breaking German ciphers, most notably for breaking the one used by the Nazi Enigma machine. His ideas and work lead directly to the modern computers that we use every day.
Turing wrote a paper in 1950 called “Computing Machinery and Intelligence.” He began the paper by posing the question, “Can machines think?” Rather than defining the terms “machine” and “think” he modified the question to “Can machines do what we (as thinking entities) can do?” To test the question, he proposed a modification of a then popular party game known as the “imitation game” where a man and woman were put in separate rooms. The guests then tried to tell them apart by writing a series of questions and then reading the typewritten responses. In the game, the man and the woman were trying to convince the guests that they are each other. So Turing proposed recreating the imitation game, but with a machine in one room and a person in the other: “We now ask the question,” he wrote. “‘What will happen when a machine takes … part … in this game?’ Will the interrogator decide wrongly as often when the game is played like this as he does when the game is played between a man and a woman? These questions replace our original, ‘Can machines think?'” Turing suggested that when a machine can fool thirty per cent of the people, then it would be hard to deny that a machine can think like a person.
Hugh Loebner began the Loebner contest in 1990 in conjunction with the Cambridge Center for Behavioral Studies in Massachusetts. Over the years it has been associated with Flinders University, Dartmouth College, and the Science Museum in London. It is now associated with the University of Reading in England.
Within the field of artificial intelligence, the Loebner Prize is somewhat controversial. Some consider it little more than a publicity stunt. Still, over the years, no program has yet to beat the Turing test and this year was no exception–which is precisely what experts in the field of artificial intellegence would expect just now.
Three prizes make up the Loebner award. Each year, there is a 5000 dollar prize awarded for the most human-seeming of all the programs entered for that year. In 2008, for instance, the prize was won by the program Elbot, programmed by Fred Roberts. It managed to fool 25 per cent of the judges (three out of twelve). In his 1950 article, Turing had predicted that by the end of the twentieth century, computers would have a 30 per cent chance of being mistaken for a human being in five minutes of text-based conversation. Turing was obviously overly optimistic, underestimating just how difficult creating true artificial intelligence would be.
The second possible prize, yet to be awarded, is a 25,000 dollar prize for the computer program that judges cannot distinguish from a real human in a text-only Turing test, and that can convince judges that the other (human) entity they are talking to simultaneously is a computer. This prize will only be awarded once.
The third and final possible prize, yet to be awarded, is a 100,000 dollar prize for the computer program that judges cannot distinguish from a real human in a Turing test that includes deciphering and understanding text, visual, and auditory input. This prize will only be awarded once, after which, the Loebner prize will be dissolved.
Most computer scientists believe that the prize will ultimately be won, and that in fact, someday soon computers will become both intelligent and self-aware. As to when this event will happen, predictions have ranged from anywhere between the next six months to twenty years in the future.
Right now, the most powerful computers have nearly the storage capacity and complexity of the human brain. However, they still operate considerably slower. Of course, even when the hardware is up to human capacity, the software will probably still be lacking. It just is not easy to reverse engineer a human brain. As yet, no computers are self-aware.