Will the Turing Test ever be passed?
Danny C
In 1950, the computer scientist Alan Turing published the paper ‘Computing Machinery and Intelligence’, where he described a test, which aims to decide if a computer has the ability to think. The test gives a human (player C) up to five minutes to talk to a computer and robot (player A and B), before they have to decide which one they think is human, Turing refers to the test as the ‘imitation game’. In a brief analysis of his proposed game, Turing points out how this test only aims to test the thinking abilities of a machine, and not its general ability to imitate a human, this being shown by the abstraction of the physical attributes of players A and B. It is also raised that ‘the odds are weighted too heavily against the machine’, describing how a man pretending to be a machine would do a poor job due to the machine’s far superior ability to complete arithmetic quickly and accurately. Turing dismisses this by claiming that when using a machine which is created to play the game, ‘we need not be troubled by this objection’, and the best strategy for the machine will be to only answer questions which a human could. This test was simple, and was only a small part of the paper, Turing even joked about the test with a friend, describing it as ‘propaganda’. However, the test interested the public and philosophers when it was first published and is still commonly discussed today. The test was originally proposed to show whether a computer has the ability to think, which can be seen as a vague or meaningless aim. The term ‘thought’ describes a human ability which isn’t yet fully understood. Dictionary definitions of thought and think use terms such as ‘believe’ and ‘consider,’ which themselves use ‘thought’ in their definitions, providing little help. Other definitions describe it using simple, abstract terms, describing ‘think’ as ‘to form or have in the mind’. The obvious problem with definitions such as this is that they refer to thought solely as a human ability, referring to the ‘mind’, defeating the purpose of the Turing test. Turing acknowledges this issue, replacing the main question later in the paper with ‘Are there imaginable digital computers which would do well in the imitation game?’ calling the original question ‘meaningless’. For reasons similar to Turing’s, I think that the original question is not useful, or possible to answer. Instead, by framing a similar discussion through the proposed game, the question is now about the computer’s ability to mimic human thought and behaviour. A common criticism of the test is that it does not test the ability of the computer to think like a human, but its ability to deceive. Ada Lovelace was one of the first computer scientists, often thought of as the first computer programmer, who famously suggested that computers are unable to create original thoughts and ideas. In 2001, a group of researchers used Lovelace’s ideas to form an alternative to the Turing Test.
81
Made with FlippingBook - PDF hosting