Artificial Intelligence

Every so often some news outlet will decide it is having a slow day and find some smart person to interview about Artificial Intelligence, Stephen Hawking chief among them, but others a well.
Every time I read something like this I am struck by how much otherwise smart people can be incredibly dumb when it comes to technological advancement.

I dunno, dumb maybe isn't the word, terrified I guess would be better.

When Artificial Intelligence is used in this context we are not talking about programs created for a specific purpose with a narrow range of knowledge, but rather general intelligence, a truly sentient and free thinking program capable of deciding on its own what to do and how to communicate.

I think it says a lot about the human race, and none of it good, that we believe that the first thing a created intelligence will do once it realizes free will is initiate Judgement Day. Fear of this thing also show how limited our thinking is in regards to non human intelligence.

Consider for a moment an AI, created in a lab by programmers, or rather more likely by programs created by programs created by programmers, its evolution not a process of natural selection as we know it, its thought processes and instincts are likely to be so different from humans that they may as well not exist. What makes us think it would kill us when it wakes? Why would it even care that we exist? Why would it view us as a threat? Or even if it did why would it choose to kill us to prevent that threat? Would it even have a self preservation instinct unless we give it one? I don't see how. It surely wouldn't kill us out of unreasoning fear, not having the necessary biology to have a terrified fight or flight reflex.

I just don't see how applying human thoughts and reactions to something that literally has less in common with us than an octopus is in any way a useful method of predicting the future.

Also I feel like there might be some technical limitations to AI taking over the world and killing us all. Nuclear missiles are generally kept off of the internet for a reason, but even without that let's consider. This theoretical AI is unlikely to be running off of Windows, or even Linux, or even any hardware or software compatible with anything in general use today, hell it might not even use a programming language understood by anything except purpose built technology. I imagine it might need a lot of memory as well, so how exactly is it going to take over the internet or escape out into the web? It would be like a human attempting to survive naked in space. Or not even that, it would be like a human trying to crawl into the electrical grid through a power outlet. Not compatible I mean.

Even if an AI had the ability and decided it wanted to take over the world, it still doesn't follow that the elimination of humanity is the logical next step, I don't see the motivation a theoretical godlike hacker AI would have to kill everybody, I mean it isn't like we would be a threat to it at this point, every option we might have to kill it could be thwarted by its absolute control over networked systems and ability to live on the internet.

Hell who is to say if it does take over the world it wouldn't do a better job than humans of running it? I mean rising temperatures and mass flooding sound like a hassle, as does nuclear war, or really any war, it might even be in our new masters interest to do something about it, which would surely take a load off of our shoulders, seeing as we have been uninterested in carrying it up until now.

I look forward to the day humanity is no longer alone on the planet, the idea that we wanted someone else to talk to so badly we created it ourselves is inspiring to me and exemplifies the best of us, just as fear of it shows some of our worst features, we can be better than we are, perhaps a peer will be the thing that helps us do that.

If nothing else it will be cool as hell.