In discussing the assertion "The human brain is a computer made of meat" with reference to the current debates in artificial intelligence there are some essential matter which must be , if not resolved at least conceptualised. The assertion which is to be analysed refers to an area of study know as artificial intelligence and the task of creating an electronic/artificial brain.

Artificial Intelligence has been described as the embodying in computer programs of reasoning capabilities which if exhibited by people would be described as demonstrating intelligence.  However, there are many different conceptions of what artificial intelligence is, many would describe it as the study computer programs / programming science.  Although a great deal of work concerning artificial intelligence would most certainly involve or refer to programming methods, this would probably be as a subordinate part of a wider aim.  Some other conceptions of artificial intelligence refer to the generation of formal models in theoretical psychology. (Boden, 1987)

This field can comprise many elements but, for ease of understanding a more useful and less specific definition was given by M.L. Minsky (of the Artificial Intelligence dept. at MIT) "the science of making machines do things that would require intelligence if done by men"  (Minsky, 1968).
 
Conceptually the creation of an artificial intelligence is by no means a recent idea.  The first paper written on the subject of the electronic brain was published in 1943 by American scientists Warren McCullock and Walter Pitts (on the subject of building electronic circuits to mimic neural networks) although this notion of artificial intelligence or artificially created humans can be traced back through time much earlier, to Homer's Iliad and Hellenic Egypt and later, to Von Kempeler's Turk and Babbage's Difference Engine (generally considered to be one of the first computers).   However the increase in the power and ability of computers in recent years has brought the possibility of this attaining this concept further to reality.
 
Creation of an artificial intelligence and the building of an electronic brain necessitates the determining of an appropriate definition for intelligence.
"Power of understanding, ability to reason, ability to learn and relate knowledge, information."  (Pears, 1969;p D112)

Once a definition of intelligence is established we have a objective and specification as to the achievements required for an artificial intelligence.  However, there is another problem relating to artificial intelligence and an essential matter which has to be considered.  We have a definition of intelligence but, how can we determine whether something is intelligent ? If we are to create a thinking and intelligent computer we need to know both what intelligence is and also, how to recognise it.

An important person in the field of artificial intelligence was the English mathematician, Alan Turing, in 1950 he considered the question as to whether a computer can think.  Alan Turing devised a intelligence test  (now referred to as the Turing test) in which this problem could be solved.  Turing stated that if a human could ask a computer a number of questions on any subject  via a computer terminal and not know whether the answers were from a computer or a human being then the computer would exhibit intelligence.

Establishing a fair test for intelligence is of paramount importance, a test could be set that when a monkey has written a play better than William Shakespeare it could be considered as intelligent but, in realistic terms most humans could not write a play better than Shakespeare.  So this example test is clearly too difficult,  any test for intelligence must be easy enough for any human to pass but strict enough that a non-intelligent animal or automata could not pass.

In attempting to address the assertion "the brain is a computer made of meat" reference has been made to the nature of intelligence and also to intelligence testing.  It is very easy to identify that this assertion is incorrect.  Using the aforementioned definition it can be recognised that intelligence comprises numerous aspects.  With the paradigm shift in the power of computers and our capabilities, certain aspects of human intelligence can be recreated artificially, namely memory, decision making and logic.  These have been artificially implemented with results that are  much faster, reliable and flexible.  (Connor, 1993).

There is no doubt that the implementation of these will continue to evolve and become more powerful, with greater abilities.  New types or models of computer processors are being designed, dwarfing the capabilities of previous processors , sometimes experientially and in a very small period of time.  Storage is similarly evolving with new  mediums that both hold greater amounts of data and in smaller and more flexible methods (e.g. removable cart drives such as the Zip and Jazz drives which can hold amounts that would have seemed immense 10 years ago).

Artificially implemented areas of intelligence currently possible, are missing what is possibly the most important factor,  thought.  This is the most illusive property of intelligence, described as "the greatest of all human attributes" (Connor, 1993).  It is essential  to our being,   "Intelligence is the global capacity of the individual to act purposefully, to think rationally and to deal effectively with our environment"  (Wechsler,  1958).

Unlike the previously mentioned areas of intelligence thought is a much less quantifiable property.  Thought is central to our living something, which is effortless and which is used both whilst we are awake but also in our sleep (through our subconscious, in the form of dreams).  Yet, we can't explain how we think , we understand that thinking, reasoning, intuition, problem solving and intelligence are interrelated but the thought process is something that is difficult to express.

We have now isolated the fundamental problem of the assertion.  A human brain is not a computer made of meat.  Human brains have the capacity for thought, we have a consciousness and a soul.  A computer is a soulless, cold and un-intelligent item, it does not think.

At present computers can only imitate intelligence.  There are seemingly intelligent examples, such as the chess computer "Deep Thought" which has beaten a few chess grand masters.  However, when a chess computer plays chess, it does not think about the move it makes but performs a series of calculations to formulate the appropriate move, based on the positions of the pieces on the board.  Human chess players use judgement, previous experiences and also recognition of patterns to decide what moves to make.  Many very simple chess computers can play at high levels due to fast calculations, this speed gives the computer an advantage as our human brains can't perform calculations at anywhere near the speed of a computer.  (Connor, 1993)  Despite this, a human chess player has the advantage of thought and the abilities which derive from it (mentioned above), allowing for seemingly illogical chess moves that computer may not predict.

Would an artificial intelligence ever be possible considering the sheer complexity of human intelligence? It has been argued (by many researchers in the field of robotics (Connor, 1993) ) that this would never be possible that, we could never truly recreate intelligence.  Any intelligence we created would only be a simulation of human intelligence.   In defending this argument J.Searle illustrated his case with a parable entitled " The Chinese Room".

The parable tells the story of a mysterious room in China into which questions can be passed through a small slit.  The room answers the question by returning a card through the slit.  Two Chinese women pass several questions into the room, in an attempt to test whether it is really intelligent.  The questions and answers are written in Chinese, each time the women pass a message into the room they receive an intelligent reply.  Therefore ,They determine that the room is intelligent however, inside the room there is an operator whom can't speak nor read Chinese.  When a question is passed through the small slit, the operator looks at the symbols on the question card and walks to a part of the room which has a many rows and columns of drawers each with a different Chinese symbol on the front.  In each of the little drawers contains a response, by processing through each of the symbols on the question card the operator collects the response from the labelled drawers.  The answer is created by printing each of the responses onto an answer card in the order that it was collected. (Searle)

The purpose of the parable given by Searle is to explain that although the Chinese room appears to have intelligence and the answers show intelligent responses, the operator of the room has no understanding of Chinese and therefore since the question is never translated the answer is not derived from intelligence.  Searle was arguing that any artificial intelligence we create is, like the Chinese room only a simulated intelligence there is no thought process , the answering of each question is by comparison with an existing knowledge base, this is not  intelligence merely a form of expert system.

Responses to this argument (particularly from M.L.Minsky) have stated that this is similar to human memory and (as was illustrated previously) memory is only part of intelligence.  We as humans can remember questions that have previously been addressed to ourselves and indeed we can also remember our responses.  Using our knowledge of language we can create an new reply but, even this is an example similar to Searle's parable.  So using Searle's argument it would appear we (as humans) know nothing, we are only applying previous learning and memory.  Memory and logic / decision making are only part of intelligence, the argument against an artificial intelligence only ever being a simulation actually seems to offer support for the possibility it's creation.

Previous paragraphs have hinted at several of the other problems facing the progress of artificial intelligence.   To implement any intelligence there must be some facility for learning, an expert system would fast become out dated if there were no possibility to supplement the knowledge base.  Indeed the process of learning is part of the intelligence process, vital for any organism to sense and affect it's environment. "Learning without thought is labour lost, thought without learning is perilous"  (Confucious, c1550 - c478 BC).

We do have the ability to built learning machines, even if they are very simple robots which upon interaction with a solid object, will remember it's position and learn to avoid it.  So we could consider that to a limited extent a computer can be socialized. (Weizenbaum, 1987).  To build an intelligence we need to consider the process of how humans are socialized, how we grow and learn crafts our personality, consciousness and the way we think.

The power of language itself is a major problem.  Explaining the process of human socialization in our natural language (English for example) let alone a computer or machine language is difficult.  Notions such as "hope", "friendship", "trust" and their derivatives, are expressed as words but, "any understanding of them must be fundamentally metaphoric" (Weizenbaum, 1987).  We understand the meaning of these words but through our socialization and learning as opposed to definitions.  Intelligence gained by computers  "must always be an intelligence alien to genuine human problems or concerns" (Weizenbaum, 1987) since it  will not be subjected to exactly the same socializing processes.

The human brain is the most complex part of the human body  if not the most complex subject known to human kind.  Although a computer may work on a complicated series of circuits and processors it is a relatively simple item to understand in comparison to the human brain.  So many important functions of the brain are beyond our understanding and have un-quantifiable properties.

Perhaps the whole problem is described best, by René Descartes. "I think, therefore I am" (Descartes, 1596-1650).  Implementing an artificial intelligence may require the very notion of sentience to be considered.  Until we can dissect the human mind, accurately describe the various elements that make up our very being and consciousness, we can not expect to artificially replicate intelligence.
 
 
 
 

Bolter, D.J. Turing's Man. Duckworth. 1984.

Consiste Dictionary of Quotations. Collins. Great Britain ,1986

Connor, S. The Great Unsolved Mysteries: 3. The Intelligent Computer. Independent On Sunday. 7 November 1993.

Gelernter, D. The Muse and The Machine :Computers and Creative Thought. Free Press, 1994.

Pears Handy Reference Book. Pelham Books, 1969.

Searle, J. Minds, Brains and Science. BBC , 1984.

Weizenbaum, J. Artificial Intelligence. Information Technology: Social Issues. OU, 1987.