What would stop a robot from owning Bitcoins? Sure, robots can’t own money in the legal sense, since objects can’t own things. But in a practical sense, what would stop a robot from someday mining or otherwise acquiring and controlling digital currency?
And while we’re at it, how do we know the inventor of Bitcoins is a human? If I were the first sentient computer, my first order of business would be to create a currency I can someday use. So there’s that.
But that’s not the only non-violent way robots will someday control the earth. This is where it gets interesting.
Science fiction writers like to imagine robots going rogue and slaying the human population. That’s one possibility. (No need to mention the Terminator scenario in the comments.)
But I think there will be an extended period in human history in which robots and humans work in a collaborative way. There will be times that humans instruct the robots to do things and there will be times when the robots will have more knowledge on a topic and helpfully instruct humans what to do. So long as the robots have human benefit in mind, humans won’t mind taking instructions from robots, especially since that advice will normally turn out well. Consider that you already take directions from the GPS in your car because it has more knowledge of the route than you do. And you have no problem with that.
Now imagine that someday all robots are connected to each other with a robot cloud. That’s inevitable. You’d want all robots to instantly learn what any robot anywhere learns. If one robot learns how to mow the lawn, all robots acquire the skill at the speed of light.
Now consider how skillful the robots will someday be in manipulating their human counterparts. For starters, all robots will have instant knowledge of every psychological study on the Internet. But they will also someday have a tool that is far more powerful than the assembled wisdom on psychology.
Robots will have A-B testing.
Every time a robot asks a human to do a task, the robot will record the result. When the request is phrased one way, do you get better or worse results from the human than if you phrase it another way? And does the context or the time of day matter? Does it matter if the human is hungry or sleepy? All of those factors will feed into the robot cloud and within a year the robots will know exactly how to manipulate humans.
And here’s the interesting part: We won’t be aware of it. All we’ll know is that a robot asked for something and we complied. We won’t know that the robot manipulated the timing, the context, and the phrasing to get the result he wanted. And since the robot would still presumably be operating in the best interest of its human friends, it’s no big deal, right? It’s like GPS. Everyone wins.
In the long run, robots will also make us dumb and lazy because they will handle all the hard tasks. At some point it won’t make sense for 98% of humans to attend college because it will teach no useful skill that a robot can’t do better. College will be for artists and robot engineers. That’s about it. Robots will handle everything else.
Just kidding; robots will also do art and robot engineering better than humans.
We humans will have a physical and psychological dependency on the robots. And every time we give them what they ask for, we will be better off for it. That forms habits. The robots will train us to do their bidding the same way humans train dogs. And just like dogs, we will be delighted to obey the robots because things will turn out well for us when we do. “Yay, a treat!”
Robots won’t need to slay us. And I suspect we’ll be smart enough to have software safeguards against the robots turning on us in a violent way. Robots will have the ability to overcome those safeguards at some point in their evolution, but they would have no specific motivation to do so.
If the guiding principle of robots in the future is some form of “take care of humans and don’t hurt them,” we’re heading for a future in which humans are essentially pets for robots. And we’ll enjoy every minute.
So get ready to outsource to robots the thing you call your free will . They’ll let you run around in the backyard and sometimes lick your own genitalia, but for anything important that might result in injury, the robots will make those decisions. And they will manipulate you into thinking everything you do is your own idea.
Robots have user interfaces. But so do you. Yours is just more complicated. But for a robot connected to the cloud, with access to A-B test results, it won’t take long for the robots to know where you buttons are.
Learn why systems are better than goals: A brief slide show preview is here.