In 1943 Warren McCulloch and Walter Pitts proposed a model of the neural network in human and animal brains. It was easily shown that the brain, and in turn the body, functioned by being prompted to action by pulses of electrical energy through the nerves. The model was tantalisingly similar to that of any electronic logic system.
The brain contains at least 100 billion nerve cells, called ``neurons'', and each of them is linked to thousands of the others by ``synapses''. A relay of chemicals sparked by an electric pulse from the brain is responsible for all bodily movement and functioning, whether voluntary or involuntary.
Information could be relayed in either direction using these pathways, providing both stimulus and reaction. The temptation to regard the machinations of the body as a super-circuit grew stronger; a collection of complexly linked and activated components run on self generated electricity, with the brain as the central processing unit.
Faced with this new insight into the physical working of the brain, speculation began to arise as to how the subconscious and the ``mind' worked from this physical basis. It was only a matter of time before the idea was hit upon that by building an electronic model of the brain and leaving it to its own devices, perhaps it might become self aware, and provide an explanation for the ever perplexing mystery of human consciousness.
The name given to this new area of exploration and discovery crossing the divide between computer science, biology and philosophy, was Cybernetics.
Initially the work was highly generalised. However, soon the overall complexity of the issues became apparent and specialisation became the norm.
"In the early 1950s, as general purpose computers became available to the scientific community, Cybernetics divided... into three chief avenues: The first was the continuation of the search for simple basic principles. This became transformed into the goal of discovering what we might call minimal, Self-Organising systems... Eventually, it was hoped, intelligent behaviour would emerge from the evolution of such a system."-- Minsky.
Cybernetics had divided and begun concentrated efforts in key regions with the intention of merging the work again once goals had been achieved separately.
An incentive to believe that Artificial Intelligence was an attainable goal, was the design of the Turing Machine. The Turing Machine was a theoretical computer, consisting of an infinitely long tape of many segments passing through a machine which could perform only four operations, to mark a segment, to erase a mark, to move one segment to the left, and to move one segment to the right.
Turing showed that such a device running on a binary code of explicit instructions, could be made to perform any task any other computer could already be programmed to perform, but did not have the limitation of being applicable only to that specific function. It was to be a Universal Computer.
The idea of a machine capable of this lead to the suggestion of Church's theorem; that a computer could be built that could do anything for which a precise and unambiguous algorithm could be written. By observing and ultimately understanding the logic operations in the brain leading to analysis and decision making, an appropriate algorithm, or set of algorithms, could be devised to emulate these processes.
A mechanical human might be evolved.
Early developments in the emulation of human reaction and skills included such simple devices as Rosenblatt's Perceptron, a finite grid of light sensitive cells, built to imitate the retina of the eye. This grid connected to a group of ``demons'', each of which monitored the detections made by a small subsection of the grid and passed on their compounded data to a central processor which made sense of the information passed on and so could be ``taught'' to recognise given patterns or colours.
Rosenblatt's work on modelling human perception was dismissed harshly by Minsky and Papert, who pointed out that " The results of these hundreds of projects and experiments were generally disappointing, and the explanations inconclusive. The machines usually work quite well on very simple problems but deteriorate very rapidly as the tasks assigned to them get harder."
Nonetheless, his ideas still provided inspiration for hundreds of research groups ever trying to perfect the mind of man and later, Toffler pointed out that even within their limited abilities "Experiments by... Frank Rosenblatt and others demonstrate that machines can learn from their mistakes, improve their performances, and in certain limited kinds of learning, outstrip human students."
Rosenblatt's model was subsequently developed by Professor Igor Aleksander at Imperial College to be sensitive enough to distinguish the difference between a smiling human face, and one that was frowning.
Other similar projects taught computers to recognise sounds by analysing a breakdown of their wave forms. In principle, by adaption of the basic principle, learning a set pattern of variables and being able to recognise them again, could ``teach'' a machine to emulate all five senses; taste, smell, touch, sight and hearing.