How does it Know?
There was an old Mel Brooks and Carl Reiner comedy sketch, “The Thousand-Year Old Man” where Brooks plays the role of the old man and Reiner plays a reporter interviewing him. Reiner poses the question,”In all of your experience, what do you consider to be mankind’s greatest invention?” Without skipping a beat, Brooks replies, “The Thermos Bottle”. Astounded, Reiner says, “The Thermos Bottle? (disappointed by this mundane choice) How do you figure that?” “Well”, replies Brooks, “In the winter, it keeps things warm... In the summer it keeps things cool. .... How does it know?”
The joke in “How does it know” is based on the assumption that our technological tools
have the capability to “know” something in the same way we humans know. But,in fact, the Thermos bottle is not sentient, it is just obeying the laws of physics whereby the incorporation of a barrier within the thermos bottle retards the heat flow in either direction. This seems obvious for low-tech machines like Thermos bottles and cash registers and we are fast approaching a society that treats the newest computing technology as if it had a sort of intelligence.
But we are also willing to ascribe a certain amount of intelligence to many other animals..such as apes, dogs and cats, birds of all sorts (especially crows) and even sea creatures like dolphins and octopuses (octupi?) because we believe that intelligence requires a thinking process. But one must be careful about defining “intelligence” because it’s a moving target. Just a generation or two ago the ability to add up a long list of numbers in your head was considered to be a sign of intelligence but when adding machines entered the picture, that ability was withdrawn as a criterion for measuring intelligence. As soon as it can be done by a machine, it does not qualify as an intelligent action.
Roughly 40 years ago, some researchers were beginning to describe computers as “Thinking Machines” and in 1983 a company was formed with just that name. Its motto was: “Some day we will build a thinking machine. It will be a truly intelligent machine. One that can see and hear and speak. A machine that will be proud of us.” This bold claim was to be achieved using Artificial Intelligence or AI.
Be that as it may, advances in AI seem to be integral to the creation of an actual thinking machine. The reasoning is as follows: if the thinking process occurs in the mind and the mind exists in the brain then a computer could become a Thinking Machine if the hardware and the software in a computer corresponded to the brain and the mind in intelligent creatures. As computer technology progresses, they more and more can act as if they possess intelligence --- thinking machines as it were.
But we have a long way to go. In 1950, Alan Turing proposed a scenario to determine if a computing machine could act intelligently: The simplest form of what has become known as the “Turing Test” is that of a Human in a room A typing questions to an entity in a separate room B
that might be either a computer or another human. They can communicate only by keyboard. The human in room A can ask any questions of what’s in room B. If, based on the responses from room B, the human in room A can not determine if Room B contains a computer or a human being or guesses wrong, then Turing claims that we can say that the computer has demonstrated intelligence and the Turing test has been passed.
One trick that humans have used to distinguish between a computer or a human is to give it two large numbers to multiply; if the answer comes back too quickly, that’s evidence that it’s a computer and not a human being. And one way the computer can fool the human into thinking it’s not a computer is to make occasional spelling errors in response to questions.
Since then there have been many Turing tests resulting in contrary claims of success --- including that it is not even a valid test. One of the arguments goes: if a computer can fool you into thinking it’s a human being , is that enough to draw the conclusion that it is intelligent? Somehow an intelligence test that depends on the gullibility of a person doesn’t ring true. The counter is that the computer is not trying to fool you into thinking it’s human, it’s only trying to imitate a human. And the back and forth continues. (see Wikipedia, “Turing test”)
Some believe the ultimate goal of AI is to construct an intelligence greater than our own. If we succeed, what is to stop these AIs from creating machines smarter that they are? Despite the contentious history of the human race, let us hope they will be proud of us and that we can live together in peace.
