Computing Machinery and Intelligence

• 1. The 1950 Mind Publication

  ›

  Alan Turing published Computing Machinery and Intelligence in the journal Mind during 1950. This paper introduced his concept of what is now known as the Turing test to the general public for the first time. Researchers in the United Kingdom had been exploring machine intelligence for up to ten years prior to the founding of artificial intelligence research in 1956. The Dartmouth workshop of 1956 is widely considered the birth of AI. Turing himself had been running the notion of machine intelligence since at least 1941. One of the earliest-known mentions of computer intelligence was made by him in 1947. His work Intelligent Machinery from 1948 did not see publication until 1968.

  13

• 2. Replacing Think With Game

  ›

  Turing proposed replacing the question Can machines think with a behavioral test involving an interrogator, human, and machine. He identified a simple and unambiguous concept to substitute for the term think. He delineated the specific machines under consideration before posing a new question he believed he could answer affirmatively. The original Imitation game involved three players: Player A was a man, player B was a woman, and player C played the role of the interrogator. Player C could be of either sex and communicated only through written notes or any other form that did not give away details about gender. Player A tried to trick the interrogator into making the wrong decision while player B attempted to assist the interrogator in making the right one. Turing modified this game to involve a computer being tested alongside a human against a human judge who typed into a terminal. If the judge could not consistently tell which was which, then the computer won the game.

  25

• 3. Digital Machines As Universal Tools

  ›

  Turing focused on digital machinery which manipulates binary digits of 1 and 0 using simple rules. He gave two reasons for this focus. First there was no reason to speculate whether they could exist because they already existed in 1950. Second digital machinery is universal. His research proved that a digital computer can simulate the behavior of any other digital machine given enough memory and time. This insight forms the Church-Turing thesis and the concept of the universal Turing machine. All digital computers are in a sense equivalent according to Turing. He restated the question as asking if a particular digital computer C could play satisfactorily the part of A in the imitation game when provided with adequate storage and speed. The focus was not on whether all digital computers would do well but whether there were imaginable computers which would do well.

  1415

• 4. Nine Objections To Machine Thought

  ›

  Turing addressed nine common objections including religious arguments about immortal souls and mathematical limits from Gödel's incompleteness theorem. Professor Geoffrey Jefferson stated in his 1949 Lister Oration that machines must write sonnets or compose concertos from thoughts and emotions before we agree they equal brains. Turing replied that we have no way of knowing that any individual other than ourselves experiences emotions. Lady Lovelace argued that the Analytical Engine has no pretensions whatever to originate anything. She claimed it can follow analysis but has no power of anticipating analytical relations or truths. Scientific Memoirs edited by Richard Taylor published her notes in 1843. Turing countered that computers could surprise humans where consequences of different facts are not immediately recognizable. He also dismissed arguments from continuity in the nervous system noting analog systems can be simulated to reasonable accuracy given enough computing power.

  161819

• 5. Child Minds And Learning Machines

  ›

  Turing postulated that programming a child mind instead of an adult mind would be more appropriate for creating learning machines. He likened the child to a newly bought notebook due to its simplicity making it easier to program. The process involved breaking down into two parts: programming a child mind and its education process. A learning method involving reward and punishment would select desirable patterns in the mind. This whole process was similar to evolution by natural selection where structure equals hereditary material and changes equal mutations. Natural selection became judgment of the experimenter. Turing noted the ignorance of the teacher regarding the machine internal state during the learning process. Inculcating random behavior would be valuable where multiple correct answers exist or systematic approaches investigate unsatisfactory solutions before finding optimal ones.

  720

• 6. Legacy Through Chess And Simulation

  ›

  An examination of artificial intelligence development reveals that the learning machine took the abstract path suggested by Turing. Deep Blue was a chess playing computer developed by IBM which defeated world champion Garry Kasparov though this remains controversial. Numerous computer chess games can outplay most amateurs today. Turing also suggested providing machines with sense organs money could buy then teaching them to understand and speak English. Such attempts at finding underlying algorithms children use to learn features of the world are only beginning to be made. Some authors have likened this second suggestion to a call for finding simulacra of human cognitive development. The paper has been both highly influential and widely criticized since its introduction in 1950. Philosophers like John Searle later raised controversial arguments such as the Chinese room against Turing's conclusions.

  1119