Skip Navigation LinksCan-Machines-Think



Clic en los enlaces de abajo para leer otros escritos




•Mente y entendimiento

•Naturaleza humana

•Espiritualidad agnóstica





​​​​​Alan M. Turing (1912-1952)

Can machines think?

Alan Turing –mathematician, cryptographer, philosopher, master of logic– is considered the father of artificial intelligence (the simulation of human intelligence in machines) and the theory of computer science (the theoretical foundations of computer science). The previous titles are common knowledge about this man. For the ideas the mathematician conceived for the construction of machines capable of learning, his name could also be added to the list of the pioneers of evolutionary neuroscience (the study of the evolution of nervous systems).

Turing wrote: "We normally associate punishments and rewards with the teaching process. Some simple child machines can be constructed or programmed on this sort of principle. The machine has to be so constructed that events which shortly preceded the occurrence of a punishment signal are unlikely to be repeated, whereas a reward signal increased the probability of repetition of the events which led up to it." According to neuroscientist Dwayne Godwin, these ideas "foreshadowed our modern understanding of how neurons learn and adapt through synaptic plasiticity".

"Some simple child machines can be constructed or programmed on this sort of principle," said Turing in "Computing Machinery and Intelligence", an important and short paper written in 1950, which has been instrumental in the development of artificial intelligence. This 'simplicity' should be 'easy' only for visionary geniuses like him.

In his renowned paper the author discusses the question "Can machines think?" that, over time, would become a permanent topic of discussion among scientists and philosophers. As a master of logic, Turing considered imprecise the way to raise this issue because the verb 'think' and the noun 'machine' are words with multiple and broad meanings.

As an alternative to the abstract question, Turing developed "The Imitation Game" (also the title of a successful movie). “The Imitation Game" became over time in the famous 'Turing test'. In this test, if an interrogator, after making numerous questions to a computer A and a person B, without seeing or knowing which is which, fails to identify who is A and who is B, A must be a machine that thinks. When this happens, the computer wins; otherwise, the inquirer prevails.

Alan Turing predicted that "in about fifty years’ time (that is year 2000) there will be computers playing the game so well that that seven of every ten interrogators will fail to make the right identification of players after five minutes of questioning ". Will the time come when machines consistently defeat ninety–nine percent of participants?

Not yet. In 2011 Watson, an artificial intelligence system developed by IBM with capacity to answer questionnaires posed in natural language, defeated in ' Jeopardy!', a famous contest of questions and answers on American television, the star contestants of that moment. However, there was at least one question that Watson answered so wrongly that any smart interrogator would have detected that the entity responding was a 'uncoordinated' device, not a 'reasonable' person.

The transition from thought to consciousness makes the issue of 'whether machines think' even more complicated. " Not until a machine can write a sonnet, because of thoughts and emotions felt, could we agree that machine equals brain", said in 1949 Geoffrey Jefferson​, a British neurologist, representative of a philosophical current known with the rare designation of solipsism, which rejects any possibility of thinking machines.

Without ignoring the mystery of consciousness, Alan Turing argued that it is not necessary to resolve it to evaluate the capacity of reasoning in computers. “According to the most extreme followers of solipsism, the only way by which one could be sure that a machine thinks is to be that machine,” noted the philosopher.

Perhaps hardware and software designers will never build robots that write sonnets and are aware that they have done so. There is no doubt, however, that at some point, sooner rather than later, there will be machines that will consistently pass, almost one hundred percent of the times, the famous test that bears the name of the hero of this note. Will such equipment have consciousness? That is a good question to pose to intelligent computers.

Atlanta, January 30, 2015