It will kill us all! It was save humanity! It will save humanity by killing us all!
The truth is, no one truly knows what genuine artificial intelligence will do once it arrives in force. It is fun to speculate and exciting to theorize about, but not even the master designers themselves have a line on what the outcome will be.
There are plenty of dire warnings from eminent scientists, people such as Stephen Hawking and Neil Degrasse Tyson, Bill Gates, and current Bond-villain in waiting Elon Musk.
Neil looks worried. Should he be?
But these predictions of doom seem more prophylactic than based on any acute understanding of the consequences. And that's to be figured. Caution is wise, and these are wise men. Humanity's self-destructiveness is always a smart bet.
However, there are some interesting things happening in the field of A.I. design and development. The most recent may be the revelation that already there are self-learning programs that are too complicated for even the engineers who made it to understand how it works.
What? Don't take my word for it.
https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/
Yes, the wacky geniuses at MIT have designed a self-learning driving program that learns only by watching humans drive. I repeat: it learns ONLY by watching, then doing. And once it starts doing, it does things it didn't see. And it does them well.
M.I.T. made a machine that runs on instinct. What else to call it? There are no instructions telling it what to do, only how to learn. The weirdos who built it are hard pressed to explain how it does what it does. Once the car learned past a certain point, they were more or less clueless.
Maybe that terrifies you. Maybe it should. But I find it kind of awesome.
Instinct? No idea how it does what it does? No idea WHY it does what it does? That sounds like every person I know!
What's he got that I don't?
I don't mean to be flippant about this, but it does seem to be revealing something about how our own minds work. Since the human brain is still vastly more complicated than that auto-driving program, we can only guess as to what truly motivates us.
Here's the kicker: the auto-car seems to have a certain kind of free will. Free will in the sense that it is expressly not programmed to execute any specific instruction. It observes then decides for itself what is the best course of action to take.
That's exciting!
Especially since deep cognitive research suggests human beings don't have free will at all. We're creatures controlled by desire, primal and otherwise, our choices determined by a deep subconscious process beyond our waking reach.
So they say, anyway. I remain somewhat skeptical, although a lot of first-rate philosophers, old and modern, would disagree with me.
Does the true threat lie not in mankind creating baby-killing robots, but in machines achieving free will where people could not? Considering how life loves to mix insult into its horrors, it only makes sense. What replaces us will have a lot of the things we always wanted but could never attain. If nothing, we will disappear having been doting, if not good, parents.
I'm not scared of the machines, I'm jealous!
