Certain traits of human mental activities can be represented by an algorithmic process, and hence imitated by a computer. How far can such practice go, and in particular, could a computer acquire the innermost characteristics of human thinking? Alan Turing proposed testing this human-computer challenge in a game-like conversation with a human judge. An instructive analogy presents distinguishing male and female voices by telephone; this is an easy task, yet occasional mistakes occur in both directions. Progressively, with more powerful computers, the reliability of the Turing test blurs.
An instructive analogy presents distinguishing male and female voices by telephone; this is an easy task, yet occasional mistakes occur in both directionsThis paper presents a series of imitation games conducted on the 100th anniversary of Alan Turing’s birth. The authors present six particular conversations that produced unexpected results. The six transcripts “formed the entire library ... for a specific hidden human.” This is believed to be the first time in the history of such tests that not one judge identified the hidden human as being human. (The hidden human in this case is actually one of the authors of the paper.)
The paper shows “the subjective nature of attributing intelligence to another mind.” It adds to, rather than detracts from, the complexity of the test, providing more insight into what intelligence is all about.