ChatGPT 4 passes the Turing test

ChatGpt 4 surpassed the famous Turing test. A further confirmation of how, in this historical period, the boundaries between artificial intelligence e human intelligence become less labile every day.

The fourth version of the famous chatbot developed by OpenAI was analyzed with a personality test known by the name of OCEAN Big-5. A test that takes into account cinque typical behavioral traits of the human being.

The experts of Stanford School of Humanities have tested ChatGPT with ranging tools from psychology to behavioral economicse.

And theAI provided answers in line with those of more than 100,000 real people, selected from 52 different countries. With a little one note of demerit.

In fact, ChatGPT obtained mediocre results regarding the trait called “ability to please”. Artificial intelligence has therefore passed the Turing test, but it does not yet seem capable of arouse the sympathy of others.

What is the Turing test and how does it work

Il Turing test it was created by the English mathematician Alan Turing. And it premiered in 1950within an article of Mind magazine dedicated to computers and intelligence.

The purpose of the Turing test is to test whether and how much a machine be able to develop his own intelligence. To conceive the idea of ​​intelligence applied to a machine, Turing started from the so-called imitation game.

The Turing test was developed and rethought During the years. But his original idea starts from a scheme he foresees three elements indicated with le letters A, B and C.

The test predicts that C recognizes who between A and B is a man and who instead it is donna. Taking into account a series of answers provided in typed form. And with a complication related to tasks of A and B. A in fact must try to trick C, while B must help him.

Turing imagined that, during the imitation game, a car replaced A. And he theorized that an intelligent machine would lead to C maintaining the same identification percentage prior to replacement.

Other machines that passed the Turing test

ChatGPT4 it is not the first machine ever to have passed (or cheated) the Turing test. A famous precedent in this sense is represented by the computer Eugene Goostman in 2014.

Eugene Goostman is a software developed for hold conversations and it was the first computer in history to pass the Turing test, convincingly more than a third of his interrogations humans.

In fact, the cleverbot convinced around 33% of its judges that it was and13 year old preadolescent. An experiment that was documented in front of none other than the Royal Society of London.

Otherwise exactly the same Eugene Goostman he had already come very close to passing the test in both 2005 that in 2012. But, on both occasions, he managed to convince just under 30% of interrogations.

Leave a Reply

Your email address will not be published. Required fields are marked *