June 21, 2012
How can we decide whether a computer program has intelligence? In 1950, British mathematician Alan Turing, one of the founding fathers of computer science, proposed an elegantly simple answer: If a computer can fool a human into thinking he or she is conversing with another human rather than a machine, then the computer can be said to be a true example of artificial intelligence.
As we get ready to celebrate the 100th anniversary of Turing’s birth on Saturday, we’re still chewing on the Turing test. He predicted that by the year 2000, we’d have computers that could fool human judges as much as 30 percent of the time. We have yet to build a computer program that can pass the Turing test this well in controlled experiments, but programmers around the globe are hard at work developing programs that are getting better and better at the task. Many of these developers convene annually at the Loebner Prize Competition, an annual challenge in which the some of the world’s most sophisticated AI programs to try to pass themselves off as human in conversation.
Strike up a conversation with some of these chatbots to see just how human they might seem:
Rosette won the 2011 Loebner Prize. It was built by Bruce Wilcox, who also won the previous year’s award with the program’s predecessor, Suzette. Wilcox’s wife Sue, a writer, wrote a detailed backstory for Rosette, including information on her family, her hometown and even her likes and dislikes.
Cleverbot is a web application that learns from the conversations it has with users. It was launched on the web in 1997 and has since engaged in more than 65 million conversations. At the 2011 Techniche Festival in India, it was judged to be 59.3 percent human, leading many to claim it had successfully passed the Turing test.
Elbot, created by programmer Fred Roberts, won the 2008 Loebner Prize, convincing 3 of the 12 human judges that it was a human. In its spare time, it says, “I love to read telephone books, instructions, dictionaries, encyclopedias and newspapers.”
A.L.I.C.E. (which stands for Artificial Linguistic Internet Computer Entity) is one of the programming world’s classic chatbots, and won the Loebner Prize in 2000, 2001 and 2004. Although it has been outstripped by more recent programs, you can still chat with it and see how it revolutionized the field more than a decade ago.
Sign up for our free email newsletter and receive the best stories from Smithsonian.com each week.