Are sentient computers such a good idea?
Recently reading a blog post on the Guardian (A UK newspaper) website, I stumbled across a thought I've been holding for some time: is it really such a good idea to create sentient computers?The reason the blog post brought this recurring thought back to the front of my mind was that it was about genetic screening and the sadly intertwined ideas of ethics and ethic debates - an admittedly unrelated topic for the most part.Humans, as sentient beings, are capable of knowing right from wrong (basically what goes with social norms and what goes against) and making decisions based on this. We also form opinions about things, based on the evidence we recieve aboout it. The genetic screening debate is a brilliant example: some of us think that if we use genetic screening we start down the road to "designer babies" and some of us think that by using genetic screening we can eliminate the suffering of those born with debilitating conditions and their families. Which side of the fence we fall on is determined by many factors: religious background, previous experience, evidence, upbringing etc. With us having the ability to form opinions comes hand-in-hand with disagreement. Hundreds of other factors mean some people have more power and others. Sometimes, opinions + differences in power = conflict. All of this is derived from the fact that we are sentient (The fact we know we are is also derived from the fact that we are). So what happens if we make sentient computers, and by extension, probably robots too? We give them the ability to tell right from wrong and the ability to develop opinions. Simple enough idea. Consider all the conflict humans are in because of a simple differing of opinions. We would be giving all our computers the ability to run in to these troubles. If a computer is sentient, it will probably be able to learn (in the way humans learn) - and learn quickly. If one computer learned how to gain power over another computer, we'd have an inequality. What does inequality lead to? Conflict.If you think about how much of our world is controlled by simple computers right now and if you consider that even more things will be controlled by computers in the future, is it such a good idea that we make a computer that can learn? If we give an artifical intellgience, like a sentient computer, the ability to develop an opinon then it will eventually come in to conflict with our own . If a computer could learn then surely it wold be able to learn to take control of non-sentient computers and, with a huge number of things being controlled by "simple" computers, an AI could easily shutdown production lines, means of communications and military defenses - crippling its human makers. If it got so bad, I don't doubt at all that a smart computer would learn to kill, if need be, and if it could take control of simple computers it could probably fight a better war than any humans could fight, given the amount of things cotrolled by, or relying on, computers in modern day warfare: planes, tanks, intelligence, communications etc. It could even use production lines to assemble its own troops.Despite a sentient computer probably being one of my top ten list of the most awesome ideas in the world, a sentient AI is just a flawed idea. Sadly, but also gladly, I doubt we'll see one in the next 100 years - not without a huge leap in the understanding of our own minds and our ability to build and program computers. Plus I don't think we want to have to draft in Will Smith to shoot a thousand pissed off robots any time soon.


