It’s been a while since I did a technology related post, and in fact this one is going to fall into the category of complete geekery, it’s also going to fill certain aspects of philosophy and ethics. If your not interested in any of these things, then it’s time to find something else to do. Try Nedroid.com for some random comics of a less serious nature, for those of you who have had their curiosity glands pipe up in the last few lines, read on, but beware, this is going to be a long read.
There is one, seemingly simple question that has always brought up masses of curiosity in me, one that I have never found the answer to. It is in fact the most complex of all questions surrounding AI in the world. It’s not whether it would be ethically right to command something that is intelligent or conscious to be a slave, and nor is it a question of how we should limit information or derestrict information sources to an AI. It’s far more complex, and far more simple than that.
Does an AI have a soul?
This question is one that divides many, confuses others, disgusts some and evokes strong emotional opinions in almost every single person who has ever been faced with the question.
Let’s start at the basics shall we?
A computer, to a certain extent, is a unique individual. It has it’s own characteristics thanks to imperfections in manufacturing processes, and whilst manufacturers will have you think otherwise, no two computers are exactly alike. Memory is slightly slower or faster due to tolerances, hard drives performance waivers to within microns of every other hard drive in the world, even something as simple as your computers cooling, two CPU coolers sit on the processors of two different computers, and whilst they are manufacturers to be the same, the tolerances, and performance reveal minute, almost undetectable instances, so one computer might read 37º and the other might read 37.4º.
These minute differences are almost never detectable by the end user, they are so minuscule that in basic terms, the computers are exactly the same, but in more advanced terms, each and every one of them is just a tiny bit different. So in actual fact, even without AI, your computer is already a unique individual.
In very basic terms, your computer is also alive, it breathes air, excretes hot gasses and technically excretes electricity, since it’s not 100% efficient. It consumes electricity and protects itself from danger. Granted the last one is due to a network of sensors and interlocks, but it does technically protect. About the only two requirements of life that it doesn’t fulfil, is that it doesn’t reproduce or have the ability to consciously think.
Of course with AI, it may not be able to reproduce, but a true AI will mean it can consciously think, so in technical terms if it wanted to reproduce it could command software to build another AI program, or even control a robotic production line to build another. It would in essential steps, be alive in basic terms.
Ok, so being alive is still up for debate, although one theory is a random collection of parts with a simple ‘go’ command could cause a computer to write it’s own code, essentially in the same way a foetus, gets a sudden electrical charge to start beating, that same electrical charge jumpstarts the brain. The jury is out, and probably always will be, but it’s interesting to know.
Of course that brings us to the next part, is an AI actually conscious?
Farthing 1982 defines consciousness as “…. subjective experience, awareness, the ability to experience “feeling”, wakefulness, or the executive control system of the mind”
If computer code essentially writes itself it could be considered aware, since it would need to adapt it’s code to the environment around it. It would need to be able to react to a given situation and be aware of it. Some might argue that the sensors sending signals back, mean it’s simply processing what it reads, but then if you think about it, humans have sensors called nerves. These nerves simply respond to a given situation and send a signal back to the brain for it to interpret. Biology and Psychology teaches us that certain responses are ingrained in our very makeup, but the human minds ability to respond to a different situation has often been described as a computer program, one that can be changed.
The gag reflex can be changed and taught to work in a different way, pain can sometimes become pleasure, and in the case of illness, our responses can be mixed up and turned around. A condition called Synesthesia, means that in some cases a person can ‘hear’ or even ‘smell’ colour. Dissociative identity disorder means that a number of different personalities, each with their own individual reactions to something, are displayed through a persons individuality. This last one isn’t unlike a computer program either, a program can respond to one input, and another program can respond to the same input in a different way.
So, if the code writes itself and we haven’t given it a set of parameters to tell it how to respond the computer is essentially aware of what is going on, of course it may react differently, but a true AI may feel fear because of consequence, e.g. it does something bad and it has one of it’s sensors triggered, or even the power taken out. Essentially the computer could be considered to be aware, and of course because of past experience it could learn to react in a different way. This fulfils the requirements of subjective experience, awareness and it even fulfils the requirement of ‘executive control system of the mind’ since it would have the choice to perform an action, knowing that there could be a consequence.
It could also be described as being capable of ‘feeling’ since sensors would be added to allow it to react to a certain event, since the code could in theory write itself, it could program itself to feel fear, even malice as well as consequence. Remember where we said about consequence? It does something bad and gets the power turned off? It could also feel fear, even become repressed and unwilling to act.
So what about wakefulness? How do we define that in terms of a computer. It’s on, so it’s aware of it’s surroundings, it’s awake, but a computer never really sleeps, it could however choose to shutdown sensors so that it enters a resting or idle state to allow for a processor to cool, this one is a tricky one, and may even become an exception.
Of course the other school of thought is that we would program the AI with an initial set of commands, parameters that tell it what is right, wrong, good, bad, happiness and sadness. The consensus however is that eventually the system would add to it’s own program, even replace the code that is already there. A true AI is widely believed to be one that can exist without input from other users or assistance.
Of course we are at the stage now where we have effectively defined that a computer is technically alive, and could technically be conscious. Now comes the burning question, the one that this entry is all about. Can a computer have a soul?
This perhaps is one of the hardest questions to answer. We’ve shown that a computer with an AI could potentially be it’s own entity, with AI you could be creating a race. But since it is something we built, and in theory, could terminate, is it a soul?
There’s no real medical definition of a soul, there isn’t really a spiritual definition. Do we even know what a soul actually is?
Rather than try and answer the question, I will leave you to ponder your own possible answers, but let’s remember a few things.
A Child, is technically something we create, in theory as with any human, we could terminate it. Although it would not be ethically correct, it’s the same as a computer, we could create it, and terminate it.
A human will inevitably die, and so a computer will through component failure. Even the oldest computers can suffer electrical degradation, albeit over a much longer period.
Humans can be injured, so can a computer. An electrical component could be subjected to an electrical shock, it could still run, although at a reduced capacity.
However, a human in most cases can heal, this is the big thing with a computer. It cannot repair itself unless it has access to a robotic arm and the ability to command them to identify and replace a component. It would however have the ability to run a diagnostic, and work out that something isn’t right, the same way a human could be injured and just know that something isn’t right with themselves.
One thing that is difficult, is identity. A child can usually identify itself or at least what it is within a few months. A child, can usually identify that it is male or female thanks to it’s organs. A computer on the other hand may not be able to do this. Can a computer identify if it is male or female? Could it even choose to be male or female?
The question isn’t so much whether an AI could be alive, but more if it has a soul. In some peoples opinions a soul is defined by so many things, including gender, a soul however to some could be nothing more than a personality which an AI would undoubtedly develop.
Would an AI have a soul? You decide.
This post was originally posted on thelongmile.net - Click here to see the original
.sketch.png)
Comments [2]