Date: Feb 06, 2019 3:00pm-4:30pm

Location: Dunton Tower: Room 2203

Speaker: Josh Redstone

Title: Three Perspectives on the Turing Test

Abstract:

Alan Turing proposed the Imitation Game nearly seventy years ago to replace the question, “Can machines think?” (Turing 1950). In this talk, I will show that since Turing first introduced the Imitation Game – now known as the Tring test – it has been viewed in one of two ways: as an experimental test of whether a machine, i.e. a digital computer, can be said to think; and, as a thought-experiment directed at testing one’s intuitions about the conceivability of thinking machines. I illustrate these two perspectives with the aim of introducing a third, novel approach toward the Turing test. This approach applies the Turing test as an experimental apparatus for learning about how people engage cognitively and emotionally with intelligent machines.

My discussion will proceed as follows. I show that the Turing test as an experimental test of machine intelligence serves as a kind of other-directed experimental apparatus. Similarly, the Turing test as thought-experiment serves as a kind of other-directed conceptual apparatus. The reason I argue these are other-directed is that they are concerned with whether a machine can be said to have a mind based upon its verbal behaviour. But I wish to change the focus of Turing’s test toward the people who interact with machines, and not the machines themselves. Here, the test is meant to enable cognitive scientists to better understanding why people perceive or judge that a machine thinks or does not think, not to answer the tricky philosophical question, “Can machines think?” Thus, this application of Turing’s test is “self-directed” or “reflexive.” I call this the Reflexive Experimental Turing Apparatus. I shall devote the remainder of my discussion toward explaining why this application of Turing’s test in fact counts as a variant of the Turing test, and toward highlighting what I believe cognitive scientists can learn by applying it as an experimental apparatus.