New to my blog? You may want to think about subscribing to be notified of new posts. Thanks for visiting!
This is an extract from the book I am writing called 'Living in the Cloud'. It is presented as a draft - comments and criticism are welcome.
I Think, Therefore I Am
It is evening and a group of men sit around a table deep in discussion. They are debating whether a machine can be made to act like the human brain. After sitting quietly for a while, a young man at the end of the table interrupts:
"The answer is to make a universal machine. One that is capable of turning itself into any other machine."
The rest of the men shuffle uncomfortably. They always felt a little uneasy around the young man. Someone asks:
"How would it know what to do? Is it possible to give a machine purpose?"
The young man responded, "It can be done with trial and error. Purpose is the use of previous combinations plus trial and error."
"But how would it learn, how would it choose what to do?"
"Random operation can be made to become regular after a certain prevailing tendency has shown itself." There were a few murmurs among those present. Who invited him anyway? But the young man continued:
"I am thinking of the kind of machine which takes problems as objectives, and the rules by which it deals with the problems are different from the objective. For instance, if I do an addition sum on the blackboard I can do it either by consciously working towards the solution or by following a routine habitual method. I get the same result either way, but they are two different things and they should be kept separate."
At the other end of the table, one of the older men leaned forward and said:
"The vital difference seems to be that a machine is not conscious. Every aspect of a machine can be fully specified, while the mind cannot."
The young man replied immediately, "The mind is only said to be unspecifiable because it has not yet been specified. A machine may be full of incompatible choices, but when it gets a contradictory result, there is then a mechanism to go back and look at things which led to the contradiction."
"But surely this is an argument against the machine - humans don't do this kind of thing do they?"
"Yes, mathematicians do."
The men whisper among themselves again. A murmur can be overheard: "Are mathematicians human?"
This discussion took place in 1949 in Manchester, England. The young man was Alan Turing, already well-known for the code-breaking machines and formulas he developed during World War II. At the time, any machine designed to do computing had to be wired and set up in a certain way. If you wanted the computer to perform a different task, it had to be painstakingly redesigned and then physically rebuilt, something which could take weeks. Turing continually emphasised the importance of what he called a 'universal computing machine', a computer that could be used for anything. He had written about it as early as 1936 and had done much research and work towards the concept, including the making of a prototype. The universal computing machine wouldn't need to be physically rebuilt to change tasks, it would simply rearrange itself internally.
Alan Turing is today widely considered to be the father of modern computing, as well as a pioneer in the field of artificial intelligence. In 1948 he began writing a chess program, but by 1952 there was still no computer in existence that was powerful enough to run it, so Turing set up a game where he played as the computer himself, following the program's logic. It took about half an hour for each move. The program lost to its human opponent, but it inspired many in the field of computer science to further the concept.
Despite Turing's many achievements however, it is another man who is given credit for designing the computing systems we use today. Hungarian-American mathematician John von Neumann took the universal Turing machine and made it a reality. Not only was he responsible for this major technological breakthrough, but John von Neumann also made significant contributions to quantum physics, economic game theory and was a member of the Manhattan Project that developed the atomic bomb.
The von Neumann architecture, as it is known, is fundamentally simple. Instead of a computer processor that follows a fixed set of instructions, the processor fetches instructions and data held in a separate store which can be reprogrammed easily as required. This store of information is referred to as the computer's memory.
When we begin to use terms like memory in relation to machines, we unavoidably draw comparisons between humans and computers. Can a computer think? Can it ever be conscious? What will happen if it does? These questions have been explored time and again throughout our culture in books, movies and philosophical debates.
Alan Turing foresaw many of the questions and concerns arising from artificial intelligence and published his thoughts on them in his now famous 1950 paper 'Computing Machinery and Intelligence'. In doing so, he chose not to follow the question 'Do machines think?' with the traditional idea of first defining what a machine is and what intelligence is. Instead, Turing posed a different question - 'Can machines do what we (as thinking entities) can do?' In this approach, the debate about consciousness is secondary, the question is whether computers can fulfil the role of a human in a given situation.
To demonstrate his way of thinking, Turing established what is now perhaps his most famous legacy - the Turing Test. The test is a variation on a popular party game in which a man and a woman would go into separate rooms and guests would have to try to guess which was which by asking and receiving a series of typed questions. Both the man and the woman would pretend to be each other and it was up to the guests to gauge from the responses which was the real thing.
In the Turing Test, one of the participants is human and the other is a computer. The interrogator is tasked with deciding which is which by asking each a series of questions of his or her own choosing. The questions and the responses must be typed. If the interrogator can't pick the computer's answers, the computer is assumed to be as intelligent as a human, at least for the purposes of the test.
To date, no computers have passed the Turing Test, though a couple have come close. However, computers have become much better at chess. In 1997 IBM's Deep Blue chess computer beat reigning world champion Gary Kasparov in a well-documented tournament match. My computer gives me a solid thrashing at chess whenever I play it, even when I follow all of its hints, suggesting at least to me that computers are capable not only of intelligence but of malevolence as well.
Computers are dangerous. They have their own languages and they talk among themselves. They take our jobs and undermine our society. So goes one theory anyway. As with most prejudices it's due to ignorance, which in turn can be combated with understanding and familiarity. Computers can certainly adapt, interact, speak and interpret, which makes them intelligent, at least for practical purposes. We know they think. But do we understand how they think?
{ 0 comments… add one now }