A researcher at the University of Geneva has analyzed the famous Turing test. She concludes that it is anti-woman.
In 1950, a famous mathematician by the name of Alan Turing devised a test that continues to be used almost 70 years later. The test asks whether a machine has the ability to impersonate a human. If a human cannot detect that he is talking to a machine, then the test has been passed and the machine has “won”.
This is how Turing originally designed his test.
Imagine three rooms, each connected via a computer screen and keyboard to the others. In one room is a man, in the second a woman and in the third a person called “the judge”. The judge’s job is to decide which of the two people talking to him via computer is the man.
In the first phase of the test, which is not used today, the man tries to help the judge, offering whatever evidence he can to prove his humanity. The woman’s job is to deceive the judge, in the hope that the judge will falsely identify her as the man.
Note from the outset how the woman is portrayed as treacherous. She’s only there to deceive the man.
In the second phase of the test, Alan Turing replaces the man with a computer without the judge’s knowledge. The aim now is to find out who is the computer and who is the woman.
Why did Turing decide to replace the man and not the woman?
Isabelle Collet, researcher and professor at the University of Geneva, a computer scientist, who in 2005 defended a doctorate in educational sciences at the Université Paris-Ouest Nanterre La Défense on La masculinisation des études d’informatique, explains that “for Turing, a woman […] can’t pretend to be a man, it won’t work. The man, on the other hand, can. The computer just has to be on a woman’s level.”
That’s right! Everyone knows that a woman is so brainless that she could never manage to pass herself off as someone else! And especially not a man!
Do you think I’m exaggerating? Alan Turing couldn’t really have thought that?
Here’s what he’s quoted as saying by Isabelle Collet: “When a woman talks, I feel like a frog is gushing out of her mouth.”
That doesn’t sound to me like the words of a man with a lot of respect for the opposite gender.
Why are we talking about a test that’s almost 70 years old? Because it’s still used and valid today.
The Turing test is still an essential step when evaluating artificial intelligence (AI). In a 2017 interview with The Wired, Ray Kurzweil, who runs an automatic writing research group in collaboration with Gmail, tells what he thinks of the Turing test. “I think the Turing test is a valid test for assessing the full range of human intelligence. You need all the flexibility of human intelligence to pass a Turing test. There is no simple natural language processing trick you can do to achieve this. If the human judge can’t tell the difference, then we consider AI to be human intelligence.”
Admittedly, the Turing test in its original version is no longer used today, but I agree with Isabelle Collet when she says that it’s “a good indication of what one of the founding fathers of computing considered intelligence. This intelligence was not in women.”
