The Turing Test

The Turing Test is a test created by British scientist , Alan Turing, "that tests whether a machine can exhibit intelligent behavior that is indistinguishable from that of a human". According to Dictionary.com, "a proposed test of a computer's ability to think, requiring that the covert substitution of the computer for one of the participants in  a keyboard and screen dialogue should be undetectable by the remaining human participant". In the modern day an A.I. capable of passing a Turing Test, would be capable of doing something like carrying on Facebook conversations with a person without them knowing they were talking to a machine.

Originally the turing test had nothing to do with artificial intelligence. When It was originally proposed in the paper, "The Imitation Game" it was meant to see if a woman could convince a third party judge (in a separate location) that she was in fact a man. The test worked very similar to how it works today (pictured above) but with a woman in place of the computer. The idea of the Turing Test is that if a group of judges is as likely to pick the human participant as they are the computer it proves that the computer is an accurate simulation of human.

Over the course of the 1970s many natural language processing (NLP) programs were created in an attempt to pass the Turing Test. One of the most famous, ELIZA, which we will discuss in more depth later, simulates a Rogerian psychotherapist. Many of these NLP programs simply scanned sentences for keywords and responded with questions accordingly never really creating a unique statement or response. Even in the modern day programs like Siri and Cortana utilize similar scans.

Present day the validity and applicability of the Turing Test has been in question, with many scientists proposing revisions or a completely new version of the test. Nonetheless the Turing Test is still extremely influential in the philosophy of Artificial Intelligence.