The Turing test in diagrammatic form.
A Turing test is a proposed way of deciding if a machine can 'think' or, more precisely,
if it can demonstrate intelligence (see nature
of intelligence) at approximately human-level. It is named after Alan Turing who first described in 1950.
In a Turing test, a human examiner would converse in natural language separately with another human being and with a machine (an advanced computer), without knowing which was which. If the examiner couldn't be sure whether the entity to which he or she was communicating was human or artificial, then the machine would be ajudged to have passed the test. It's assumed that both the humans and the machine try to appear human.
The origin of the test is a party game in which guests try to guess the gender of a person in another room by writing a series of questions on notes and reading the answers sent back. In Turing's original proposal, the human participants had to pretend to be the other gender, and the test was limited to a five-minute conversation. These features are nowadays not considered to be essential and are generally not included in the specification of the Turing test. Turing proposed the test in order to replace the emotionally-charged and, for him, meaningless question "Can machines think?" with a more well-defined one. Turing predicted that machines would eventually be able to pass the test. In fact, he estimated that by the year 2000, machines with 109 bits (about 119MB) of memory would be able to fool 30% of human judges during a five-minute test. He also predicted that people would then no longer consider the phrase "thinking machine" contradictory.
It has been argued that the Turing test can't serve as a valid definition of artificial intelligence for at least two reasons: (1) A machine passing the Turing test might be able to simulate human conversational behavior, but this could be much weaker than true intelligence. The machine might just follow some cleverly devised rules. (2) A machine might well be intelligent without being able to chat like a human. Simple conversational programs, such as ELIZA, have fooled people into believing they are talking to another human being; however, such limited successes don't amount to passing the Turing test. Most obviously, the human party in the conversation has no reason to suspect they are talking to anything other than a human, whereas in a real Turing test the questioner is actively trying to determine the nature of the entity they are chatting with.
The Loebner Prize is an annual competition to determine the best Turing Test competitors.