Free Will... Is it really Free?

I've always been a big follower of AI, and in fact as a computer scientist I did some small work in this area. What really catches my attention are the philosophical aspects related to the nature of cognitive processes and consciousness. In this regard I love John McCarthy's seminal papers on this area. I also agree with Marvin Minsky when he says that this aspect of AI has slowly been neglected, and instead modern AI has broken into a series of subfields, each one working on improving what are basically specialized optimization algorithms, losing the focus on what really started the whole thing: creating an intelligent machine. Maybe this is the way to go, and eventually we will be able to put everything back together... but somehow I don't think so.

Anyways, I recently ran into a thesis from a PhD student in Spain, that attempts to model self-awareness using neural networks. The work is very interesting, although not groundbreaking, but at the very least gives a very good survey into current ideas for modeling consciousness. All of this got me thinking back into a bunch of ideas I was having a while back regarding all this, and I've had conversations and discussions with several people about whether or not consciousness can really be replicated (to the point of self-awareness observed in humans), but then I got thinking into free will.

The thing is, if you start pondering about the idea of whether consciousness can be replicated by a machine, or whether strong AI is possible, you really have to get philosophical and start with a simple question: Regarding the philosophy of mind, do you stand on the side of materialism or dualism? If you adhere to the idea that the mind is a result of the physical processes occurring in the brain, then certainly the idea that these physical processes may be replicated by some sort of machinery is a plausible one. Let us say, for the sake of argument, that this is the case, and let us abide by the idea of a materialistic conception of the mind.

So, we have decided that the mind is the product of a machine (the brain), and therefore not separate from it. In this scenario, every cognitive process is the product of a computation, every thought, every decision, every idea. We could certainly view the brain as a machine exposed to a series of stimuli (the inputs), and based on those, producing a series of cognitive effects (the outputs: decisions, thoughts, ideas). If that is the case then... how do we interpret the concept of free will? First of all, let us try and be more specific about what kind of machine the brain would be: is it a deterministic machine or is it a non-deterministic machine?

Let us assume, again for the sake of argument, that the brain is a deterministic machine. This means that, given some pre-determined circumstances, the brain will always react the same. That is, on a given situation, if a person makes a particular decision, given the chance to do it again, if the circumstances are the same, he/she would make the same decision. If this is the case then... do we really have a free will? Under these circumstances, it would seem that free will would be nothing more than an illusion, a subjective perception or rather interpretation of our own congnitivity. Given a particular situation, we would be pre-determined to act in a certain way, no matter what, even if we thought that was not the case, that we somehow had the illusion that we had a choice.

All these ideas have already been tackled in areas like philosophical determinism, but it's interesting to see how they emerge by delving into the intricacies of another field, namely, Artificial Intelligence. Personally, I find this idea of mental determinism (to call it one way) very interesting, maybe a little scary, but that doesn't mean it's not true.

In the end, we would probably not be able to get away with a fully deterministic model for the brain, for we would have to account for the instances in which we just decide to flip a coin (both literally and metaphorically), so we would have to model some non-determinism into it, probably by adding some source of randomness, but the point would be that in that instance, we would always decide to flip that coin.

Hmmmm... I need to think more about this.