Simulated brains - conscious machines?
The Blue Brain project has been working since 2005 on using supercomputers to simulate the operation of parts of the brain, producing a computer model which behaves in the same way brain tissue behaves.
Simulated Brain Cortical Stem.
The purpose of the research is to develop simulated models which can show the same response to their environment as human tissue and that offers enormous therapeutic and research potential. For example, if the simulated tissue reacts in the same way as human tissue, we can experiment with the simulation to see how, for example, it will react when treated with certain drugs. Or how it would respond over time if a section of it was removed.
It has already been shown that the simulation can learn and remember and the researchers are able to see memories being retrieved. So not only do these simulations have medical research potential, but they also illustrate a very controversial aspect of human brains - consciousness.
One theory of the brain argues that consciousness is what is called an "emergent property" of brain tissue. When the brain complexity reaches a certain point, it becomes able to monitor its own functions, to become self-aware. At that point, it is argued, the brain shows the properties we identify as "consciousness". This comes in varying degrees, from vague awareness of individuality, to sophisticated abilities to think on demand and to be aware of doing so. Manipulating concepts, recalling memories on demand, exhibiting emotions, all have their roots in consciousness.
If these simulated models can be shown to possess emergent properties, then the whole debate about mind-brain takes an interesting turn. Religious people particularly maintain that people have something ethereal called a "soul", some immaterial essence derived from some spark of life. The only evidence ever offered for such a notion has been the first-person reports of individual belief, which of course is never susceptible to rational investigation. You either believe the report, or you don't - you can't test whether or not it's true because it depends on the perception of the individual reporting it.
But if a machine simulation can demonstrate just those characteristics of consciousness, self-awareness, memory, emotion, etc, then we have a curious situation. A machine which can report exactly the same first-person sensations, feelings, as a human. What then happens to the notion of a soul? Does the machine simulation have a soul? Religious folk would be opposed to the notion, but they would find it difficult to distinguish between a machine claiming a soul and a human, other than by insisting a soul only belongs to a human.
But by defining a soul as something specific to humans, in a form which cannot be investigated, the term loses any meaning beyond an expression of faith.
It's early days yet and although the Blue Brain project has produced some impressive initial results, they haven't produced anything too challenging for religious irrationalists but that will change. Give a machine self-awareness and the ability to make moral judgements based on ethical principles, and we could see the arrival of a new kind of Turing Test, one based on distinguishing between a dogmatic clerical leader and a Mark 1 Moral Reasoning Engine. My guess is that they'd be almost indistinguishable at the outset but the machine would learn and quickly outstrip the clerics.
That wouldn't be a major milestone since dogmatic ethics is very limited, but the process of moral learning would be a clear indication that moral reasoning does not require belief in any deity. People can be good without a God.
