The Myth of Artificial Intelligence

In the present age of fascination with computer technology, there is an inordinate amount of attention being paid to the phenomenon of "artificial intelligence" - and the misinformed belief that computers are capable of thinking.   The faith that uninformed individuals place in computers' intelligence is in many ways similar to a belief in magic - and its truly amazing how many people who seem otherwise intelligent subscribe to this belief.

The irony is that there was a time, a few centuries ago (and in certain circles of developed nations even today) in which people strongly believed in the supernatural, and many very important decisions were made based on the belief in otherworldly forces that control anything that cannot be adequately explained.   Diseases were attributed to evil spirits in the blood and human behavior was guided by angels and demons whispering in our ears - and such nonsense was proffered and accepted with just as much earnestness and conviction as people believe in the inner workings of computer technology today.

Individuals who are knowledgeable about technology recognize that computers do not think - they merely execute programs, which is to say that they follow processes that implement the intelligence of the human being who wrote the programs.   In effect, they are mindlessly following their instructions, much in the way that a person might blindly follow orders.  Any intelligence a computer system seems to exhibit was instilled in it by the person (or persons) who wrote the instructions that the machine executes - and much in the way that a person who merely does as he is told is not regarded as a thinker or decision-maker, the machine does not think or make decisions because it is merely doing as it is programmed.

By its nature, any computer program is designed to solve a problem.  When a problem is narrowly defined (identify which number is largest or add two numbers together and report the sum) the program is not regarded as intelligent.     There may once have been a time when people regarded a simple calculator as a mathematical genius, but most people are now sophisticated enough to recognize it is dumbly performing specific functions and is remarkable only for the speed and precision with which it does so - which is valuable and a bit amazing, but nothing like "intelligence."

As the program gains complexity (subtract the smallest prime number in a set from the square root of the largest number in the set and report the difference as an absolute value) it becomes difficult for human beings replicate the process of logic that has been performed.   We appreciate that the computer is more accurate or faster at preforming a task than is a human being, but still do not consider it to be intelligent because we can perceive the very simple logic by which it operates even if we cannot match its speed or accuracy

However, when relationships among a large set of data are evaluated in more complicated ways (consider the return and risk of various investment options along with the resources and risk tolerance of an individual to provide him with an investment plan for retirement) if is difficult for human beings to conceptualize how the programming works - and they regard it as being intelligent simply because they do not understand how it does what it does.   It would require reading tens of thousands of lines of code in order to comprehend the way in which it evaluates input and arrives at a conclusion - but ultimately, the instructions are individually quite simple, though tedious and convoluted in aggregate.

So in the end, the belief in artificial intelligence reflects a failing in human intelligence: if we do not understand the way in which a computer is able to evaluate data and arrive at a solution, we proclaim it to be intelligent - much in the same way that a person in the Medieval era would have proclaimed to to be the result of magic or the will of the gods.   The two are essentially no different.

It also doesn't help that the definition of "intelligent" is problematic.  Cognitive psychologists have argued for decades and have no firm agreement of what constitutes intelligence in a human being, and some of the definitions are loose enough that a computer (or even a calculator) might satisfy the criteria.   And the best that has been offered for assessing computer intelligence is Alan Turing's principle of AI, which loosely suggests that a computer can be considered intelligent if it can trick a person into thinking they are interacting with another human being- a gut-check as to whether seems to have a quality we are unable to clearly define.

A Turing test measures the perception of the person who is making the assessment as well as the predictive ability of of the person who programmed the system - and as such the machine is merely a channel.  We do not think a telephone or a typewriter is intelligent because it has said something clever or profound, but recognize that it is a (dumb) device by which we are receiving the message created by a human being - yet we fail to apply the same logic to realize that a computer or digital device is merely a delivery mechanism for a message created by its human programmers.   And just as a typewriter cannot deliver a message unless a human being works the keys, so does a computer fail to deliver a response unless a human being has provided instructions.

If we consider "intelligence" to be the ability to follow instructions to the letter, machines may seem to match the definition so long as the user limits himself to acting within certain parameters - that is to say that if you ask a machine that is programmed to calculate income taxes a question about income taxes, it will provide an answer that seems intelligent.  Ask it a question about gardening, and it will fail to provide any response or even an intelligible excuse for lacking that knowledge, but spit out an error message that vaguely indicates that it doesn't understand the question.  

The irony of this is that many would regard such a user as stupid to preserve the reputation of the machine, which cannot think outside of its narrow programming.   When we begin to believe in the power of a physical artifact so strongly that we make excuses, exceptions, and find some other source to carry the blame for its failings, we have clearly departed from the realm of intelligence and have returned once again to religion - and as is all too evident in the present day, the proponents of dogma will stubbornly maintain a failed argument in spite of all evidence to the contrary.