An £825million project, estimated to take 12 years to finish, is currently underway to create a ‘human brain’ supercomputer. The purpose of the project is to simulate the working of the human brain right down to communication between individual cells, in order to gain better understanding of diseases such as Alzheimer’s and Parkinson’s. But with its billions of neurones, supporting cells and connections, is it really possible to create a computer as complex and intelligent as the human brain? Or maybe computers really are the answer to revealing some of sciences most important questions.
As of yet, there is no machine as complex and intelligent as the human brain. Billions of neurones interact with each other through synaptic connections, consisting of thousands of complex processes, which are often very far from being properly understood. Neurones themselves receive thousands of inputs at any given time and convert these messages into useful outputs, so how could a computer ever replicate this? Well, computer science is rapidly progressing, and with the development of adaptable ‘artificial neural networks’ the possibility of a simulated computer brain of some complexity certainly seems achievable. However, these artificial networks will always be limited to the level of understood electronics and therefore have limited available inputs and outputs. The brain on the other hand is able to form new networks and discard old ones via a process known as synaptic plasticity. The ‘plastic’ nature of the brain raises it’s adaptability far above that seen by any supercomputer, and certainly questions the possibility of a simulated ‘computer brain’, able to rival that of our own, being achievable.
In terms of processing power, the brain is bound to win. However it can be argued in terms of intelligence, that internal complexity doesn’t matter if the outcome is the same, giving the “Illusion” of intelligence. If you are having a conversation over the internet to someone you don’t know, as long as the replies to your questions make sense and are appropriate, as far as you will be able to tell you, will be conversing with a human. What if I were to say it was a computer program? You can’t tell the difference so is this in fact an acceptable level of artificial intelligence? Obviously conversation is only one aspect of artificial intelligence and as computers get more powerful, which is inevitable, the quality and scope of artificial intelligence will increase. For example, giving robots the ability to navigate their way through changing environments. This robot won’t use the same complexity of process as the brain but will accomplish the same task using sensors, as opposed to the complicated networks that process the visual data from the eyes. It would seem then, that an equivalent level of artificial intelligence can be achieved without matching the power of the brain.
Even though the true complexity of the brain may never fully be matched via artificial means, perhaps it is possible that its’ “intelligence” and “power” will be emulated to a level whereby no distinguishable difference could be observed in comparison to our own brains. In time this may well open the door to unravelling the mysteries of the most complex machine on the planet.
Tim Knight & Alex Smye