Thoughts on Turing’s Imitation Game
Alan Turing introduced the concept of the imitation game in 1950 in Time magazine, while he was working in the Computer Science department at The University of Manchester (1948 to 1954). In doing so, he raised a seemingly timeless question: are there or will there ever be machines which can successfully imitate humans? And what does this idea of imitation actually mean? Rather than give any definitive answer in his 1950 article, Turing explores his question by responding to a series of “objections” from a range of perspectives such as theological and mathematical. He even gives the example of a “heads in the sand objection”, suggesting that we are all to some extent fearful of the idea of machine thought. The “Argument from Consciousness” quotes Professor Geoffrey Jefferson, who became the UK’s first professor of neurosurgery at The University of Manchester in 1939. Jefferson argues that a machine would have to be able to feel emotions before it could be said to be thinking. Rather than seeking to answer his own imitation game question in a literal way based on the technology available in 1950, Turing’s article is a philosophical and theoretical exercise, and he suggests that we should not simply ask what is possible at any particular point in time, but whether it is possible to imagine computers at any point in the future which might be able to convince in the imitation game:
we are not asking whether all digital computers would do well in the game nor whether the computers at present available would do well, but whether there are imaginable computers which would do well.
Turing describes his theoretical imitation game as a form of Q&A between an interrogator communicating with a human and a digital computer in separate rooms, with a printer available to print out their answers; spoken voices would be a giveaway! The interrogator asks questions which the human and computer must answer, with the computer’s “aim” to convince the interrogator that it is the human being and therefore win the game.
Turing finishes his article by introducing the idea of the “child machine”, comparing how a machine might be able to learn to the way that a child learns through experience, according to certain rules. He introduces the possibility that a machine may, in future, have the ability to learn and therefore change or develop the rules of its behaviour, in the way that human beings do. And of course, we now understand that “machine learning” is possible, whether or not it successfully imitates the learning of living organisms; we use machine learning daily now through web searches, speech recognition and perhaps soon in self-driving cars. (Continue reading this post on The Imitation Game)
Curator Clare Gannaway posts on her blog, The Imitation Game, ahead of the exhibition she is curating due to open at the gallery in Spring 2016.