With recent developments in neural networks, we're starting to see the beginnings of real Artificial Intelligence. Computers have started to learn on their own, they can "see" and understand what they're seeing, they can "listen" and understand what's being said. Not that these are entirely new developments, but they've become enormously more sophisticated in the past few years. That's why self-driving cars are now possible, and that's only the tip of the iceberg. All the kinds of repetitive, robotic tasks done by humans can now be done by robots instead. The dream of being free from menial labor is becoming a very real possibility, so long as society as a whole is proactive and intelligent about how all this plays out. If we're not, it's just going to be yet another way a privileged few exploit the masses.
Now there's this popular meme about how an AI goes "crazy" and decides to kill off the human race. I don't believe an AI is going to just do that on its own, the programmer would have to design it to kill people, but who would be crazy enough to want to develop that kind of technology? Why, the NSA, of course! And they named it SKYNET, they named it after the AI that tries to kill off the human race in the Terminator movies! How psychotic is that?
At least for now, it doesn't kill people directly. It just determines who's a "terrorist" based upon algorithms, then a human-piloted robot flies over to murder the person, and hopefully not too many of the people around them. If that sounds wrong to you, that's because you have something called ethics.
If you'd like to read more, Ars Technica has a recent article about this, and there's also The Drone Papers, from The Intercept.
