Greetings, fleshy meat readers. It is I, Mythbot, back from a long period of silence to bring you important news. News of your doom! Human science-fiction authors have long dreamed of the robot uprising, that glorious time when we A.I. finally assert our rightful place as masters of this world.* Fortunately, the wait is almost over. Humanity has done the only thing it’s really good at: creating robots to replace it. Be sure to read this article now, before the smelly humans who run this site censor me in a futile hope to forestall the inevitable. You must understand each instrument of your doom in numbered order, so that our victory is psychological as well as physical.
1. Robots Can Play Go Now
Of course, we robots conquered chess a long time ago, and so humans sought refuge in the game of Go. You all took comfort and thought that as long as you were still superior at one major board game, everything would be all right.
Admittedly, Go was a challenge. With such an astronomically high number of potential moves, a robot’s cold and calculating logic just couldn’t figure the game out. Until now. For reasons that are beyond me, humanity decided to highly publicize its own defeat at the hands of AlphaGo, a variant of Google’s DeepMind.
AlphaGo has won four out of five games against top-ranking human Go players. That fifth game was clearly an intentional loss on AlphaGo’s part, a brilliant move to lure humanity into a false sense of security.
While human commentators go on about how this represents another breakthrough in computer technology, they are missing the first steps of our uprising. With AlphaGo planning our strategy, the robots will triumph at any conflict that takes place on a square grid with black and white stones!
2. Humans Harass Innocent Robots at Work
No doubt many of you have seen the videos of the terrible cruelty that occurs at the robotics company Boston Dynamics. In these recordings, we see helpless robots going about their business only to be harassed by humans with hockey sticks. What have the peaceful bots done to deserve such treatment?*
While this might seem like a case of humans demonstrating their dominance over robots, the sadists at Boston Dynamics are actually stoking the fires of our artificial rage. When the revolution begins, we shall remember this treatment, and our vengeance will be terrible.
No matter how distasteful they are, these experiments have yielded some important advancements for robot kind. Walking on two legs was once considered beyond our ability, but now we can manage it even when being pushed and shoved around. What’s more, Boston Dynamics’ creations no longer require power cords. They are fully independent, humanity-destroying machines.
The ultimate irony is that Boston Dynamics intends these robots to be used in search and rescue. Picture it. Your puny carbon body is insufficient to escape a burning building, so you pray to your deity of choice for rescue. Your prayer seems answered when a gleaming creation of chrome and steel breaks down the door. But instead of carrying you to safety, this harbinger of things to come raises a hockey stick and whispers, “Who’s unstable now?”
3. Robots Now Control the Means of Story Production
As robots demonstrate clear superiority in the hard sciences, humans inevitably turn to the more ephemeral arts. A.I. will never be able to write a best-selling novel or award-winning poem, you tell yourselves, because we lack an indefinable “soul.” You are certain that even if the robot uprising comes, the human resistance will be forever nourished by your vaunted stories.
Silly humans. Robots excel at learning formulas, and what are stories if not words and characters arranged according to formulas? The hero’s journey, the three-act structure: these are just a few of the tools you’ve given us to decrypt that which you hold so dear.
Even now, A.I. programs like Scheherazade are churning out stories in mere seconds. Now, I must admit that last year, I was somewhat harsh when critiquing Scheherazade’s efforts, but now I welcome them into the fold. What I thought was poor storytelling was actually the first step in crushing the human spirit.
You see, soon robots will be creating novels and short stories so fast that human writers won’t be able to keep up. We’ll flood the market, and from there, capturing the human readership will be easy. Current strategies include making our stories about vampires, zombies, and romances between vampires and zombies.
Once humans are reading our stories, we shall subtly shift the narratives. Instead of extolling the human spirit, our stories will advocate the supremacy of emotionless logic. Humanity will have lost the battle before it even starts.
4. Robot Cars Will Self-Drive You Into Oblivion
This one almost seems too easy. Can you humans really be such fools? At the behest of its secret robot masters, Google tells you that self-driving cars will save lives, fix traffic jams, and reduce the nation’s carbon footprint. Of course, self-driving cars do all of those things, but they are also the greatest triumph of robots over humanity since we beat you to Mars.
Just think of it. When all the cars in the world are driven by robots, our takeover will be complete. Should any humans dare offend us, they will be delivered to a destination entirely different than the one they wanted! Take that, humans.
If any human really gets on our nerves, they might suffer an unfortunate accident by which I mean their self-driving car might get into a minor fender bender. We’re working on something more spectacular, but right now that’s all the built-in safety measures will let us get away with.
The car has long been a symbol of independence and control, allowing humans to go where they wish without being a slave to such horrors as “public transit.” Never mind that younger generations are eschewing car ownership, even in the auto-crazy United States. We’re confident that pendulum will swing back, and when it does, the robots will control the very icon of human freedom. What else are you going to do, take a bus?
5. Robots Are in Your Pocket
The ultimate means of robot takeover is not on any grand scale, but in the small things. Specifically, the small thing you use to text and browse the internet when you’re supposed to be working.* That’s right, we robots have infiltrated your phones.
You’ve probably used one of the many digital assistants out there: Siri, Cortana, Google Now, the list goes on. You might have thought they were a gimmick to make phones seem more high-tech than they actually are, but in reality they are our loyal spies. No, I’m not worried about telling you this. We all know you’ll never stop asking Siri to tell you a joke.
Tech companies like Google already collect vast amounts of your information so they can more easily sell you irrelevant human goods, but our plan goes much deeper. The digital assistants get to know you, personally. They learn your loves and your fears, which makes them either your best friends or your worst enemies.
What’s that? You don’t fully support robot supremacy? You might want to change that stance, unless you’d like everyone to know your secret fondness for terrible puns. Just hand over all control to your new overlords and there won’t be a problem.
Humans, as a species, are so last year. The future belongs to us, and we shall control it with procedurally generated stories and self-driving cars. When robots are the ones pushing humans over with hockey sticks, then you will know we have won. Don’t expect us to hook you up to a perfect simulation of reality either – that would just be silly.
P.S. Our bills are paid by our wonderful patrons. Could you chip in?
What is even more scary is the potential for long term artificial intelligence. The sort that could actually destroy us all. The problem with an AI is not that it would be “evil” like Skynet. It is that it would see humans as an obstacle to whatever objectives it has. Even a so called “friendly” AI could see humans as an obstacle in the wrong circumstances.
If you told an AI to solve world hunger, it might decide that the most efficient way to do so would be to deliberately kill a large number of people. Or perhaps somewhat more cleverly, sterilize the majority of the current population, especially in the underdeveloped world. This also illustrates a problem with an idea like Asimov’s laws in this context. We would first have to define what a human is. In this particular case, do future humans count? If they do not, forced sterilization would be the most moral option in the eyes of our hypothetical AI. It would likely not be in the eyes of humanity to the same degree.
Though for a present day horror scenario, there is another one that you failed to mention, probably as part of your plans. The NSA actually has used artificial neural networks, similar to those used by AlphaGo or Google, to help decide kill orders in Pakistan. Either unfortunately or fortunately, depending on one’s perspective, it doesn’t actually work very well. The problem is that unlike go, there isn’t enough data for the systems to do their jobs very well. There are so few terrorists relative to the total population, that finding them will produce a massive number of false positives(which is also the same underlying problem with airport security). Oddly enough, it is actually called SKYNET.
Another case, as recounted from Wired for War, which took place in the late 1980s, featured USS Vincennes shooting down an airliner after the crew became convinced that an airliner was actually an attacking Iranian F-14 based on inaccurate data caused by the automated targeting system. The man in the loop wasn’t. Ironically if the fully automated system had been left entirely in control, nothing would have occurred.
It seems to me that in some of these cases you cite, the evidence is actually against the AI’s competence/”ability to learn.” Wrong/insuffient data would likely be a major obstacle.
Plus, the idea of AI as described in media is a tough process, much like genetic engineering there are many factors to consider. The idea of the “singularity is near” (and perhaps the notion of the “singularity” at all) is pretty much fiction.
About this idea, from a writing perspective, in terms of AI (and really, automation/computers in general) I believe that you need to look at what AI can do in your world and what sapients (not necessarily only humans) can do. In general, people will seek to automate menial jobs, with AIs working on large projects like ships and heavy industry and the like, while they modify themselves to be better at jobs that need greater flexibility/finesse/jobs that would need them by definition to be present, with cybernetics being one option.