Scientists have explained why while it is impossible to create artificial intelligence.
Test of cognitive abilities of artificial intelligence was proposed in 1950 by the British mathematician and cryptographer Alan Turing.
Experts in artificial intelligence Kevin Warwick (Kevin Warwick) and Huma Shah (Huma Shah) discovered a serious problem in the standard Turing test. The test determines whether artificial intelligence can think like a human.
Cybernetics found that if the machine will resort to a kind of analogue to the Fifth amendment (US Constitution), that is, to remain silent during the entire test, it will pass the test. In this case the artificial intelligence that simply refuse to answer the judge’s questions, will be recognized as a thinking being. However, to remain silent, every primitive machine which has no intelligence, so the Turing test under these conditions stops working.
Test of cognitive abilities of artificial intelligence was proposed in 1950 by the British mathematician and cryptographer Alan Turing. According to the scientist, it is the only way to find out is whether the machine on the human mind. Also known as the “imitation Game” test offers person-the judge blindly ask written questions to two subjects: to the other person and artificial intelligence. For the answers you need to determine who is who. If this fails, the machine is recognized thinking.
Warwick and Shah examined records of these tests of the loebner competition, in which the artificial intelligence refused to “talk” with the judge. In each case, the judge could not determine with whom he is dealing — with the machine or with the person. The subject could be a person who decided not to answer the questions, artificial intelligence, or machine, is simply incapable of reply. In the tests took place is a third option.
Scientists have named this issue the “Fifth amendment” by analogy with the amendment to the U.S. Constitution, which States that a person shall not be compelled to testify against himself.
The loebner competition is the main platform for testing Turing. It involves the developers of the chat bots and virtual interlocutors. Critics of the competition have indicated that programs which prove to be winners, was designed to specifically to introduce the judges misled, purposely making spelling mistakes and resorting to other tricks.
