Tomorrow is January 1, 2020 and we do not just start a new year but a new decade. In the world of artificial intelligence, some believe in this decade we will pass the Turing test. What is the Turing test and why is it important?
The Turing test is a measure of the power of Artificial Intelligence. When a computer passes the Turing test, it means it will be equal to humans in every way. The test was developed by the pioneering computer scientist Alan Turing in 1950 to determine whether a machine has human like intelligence. The machine passes the test if a interviewer cannot tell whether a response is coming from the machine or a human. If someone can’t tell the difference then we consider AI to be of human intelligence. In 1999 the futurist Ray Kurzweil predicted machines will pass the Turing test by 2029. That is this decade.
So how does the Turing test work? In his paper entitled Computing Machinery and Intelligence, Turing outlined a method for answering the question “can machines think.” He proposed a hypothetical game with 3 players. One player is a interviewer separated from the other 2 players, one of which is a computer and the other a human. So you have a human interviewer asking questions of a computer and another human. Through this process the interviewer tries to figure out which is the computer and which is the human. Ultimately if it can’t be determined which one is a computer, then maybe they’re dealing with a thinking machine that has passed the Turing test.
Is it realistic to anticipate human level machine intelligence by 2029? AI researchers believe we have the computational power to build Turing’s thinking machine but a major problem is that computers still struggle with routine small talk and are even much worse than me at telling jokes. Language is widely seen as humankind’s most distinguishing trait. And for a machine to have a conversation with a person takes more than increasing memory and processing power. It requires understanding the meaning of language and all the implications in speech.
Kurzweil, head of natural language at Google, is more concerned that when the machine passes the Turing test, we’ll have to be careful about what kind of feelings that computer has, about it’s emotions and consciousness and we’ll have to care about what its’ thoughts are. He thinks future AI is emotion and will come with the Turing test being passed. And although consciousness is a philosophical question not a scientific question, because you can’t test for it, he believes computers will be conscious and have all the secondary features we associate with consciousness such as having an opinion, an ego, and desires. And that raises questions about what it means to be human.
As an aside, in my reading for this flash talk, I came across a comment which raises a subject I want to discuss in 2020, something I call dystopian AI and encompasses the ethics of artificial intelligence. The comment was: “ I’m not scared of a computer passing the Turing test. I’m terrified of one that intentionally fails it.”
From Short and Sweet AI, I’m Dr. Peper