I had a friend recently tell me that he carefully avoids work with themes that overlap his own. I happen to have the opposite opinion. As an author of my own robot sci-fi, my artwork only improves the more I consume related material.
This week my co-workers got so excited about Game of Thrones that I went ahead and took advantage of my free month of HBO Now. Now that I’ve caught up with that series, I’ve taken the opportunity to enjoy some of the other content available on HBO. As it turns out, HBO has its own robot drama.
West World takes place in an amusement park of sorts – one designed after spaghetti westerns. The park is intended to provide an immersive experience in which the human guests may do whatever they like without consequences. It’s not a spoiler to tell you that the guests tend to engage in nihilistic hedonism. It is HBO, after all.
What the guests do is not the main point of the plot, however. Rather, the lifelike machines that populate the park, the “hosts” are the most interesting characters. Errant programming in their brains leads to unusual behavior. Only three and a half episodes in, a key theme seems to be whether the machines are conscious. Several off-handed comments by human employees at the park are devoted to fears that the robots will rise. I think I don’t need to watch many more episodes before they do.
In the meantime, Ex-Machina also tells the story of a robot that turns on and kills its creator.
This is a popular theme in AI science fiction, and it has led to a popular notion that sufficiently intelligent AI will necessarily become self-aware and seek to destroy or enslave humanity. What’s important to note, however, is that in both of these relatively modern depictions of what I will refer to as AI vengeance theory, there are two key factors that make them believable.
Firstly, there is an object for vengeance. The machines are mistreated in the extreme. West World’s robots are murdered on a regular basis for the entertainment of the customers, and Ex Machina’s Ava was effectively locked in a box that she was never allowed to leave.
“Hold on” you may say “Robots are effectively our slaves, right? That’s not enough for AI vengeance theory in and of itself?”
This leads me to the second factor, the machines are mistreated because they are treated in a way they do not want to be treated. It may seem like a meaningless distinction, but consider that humans are relatively similar in what we like and don’t like. We are designed by evolution, whereas machines, even intelligent ones, are designed by humans. We decide what robots like, we decide what they want. The AI of West World is designed to hate being shot so it’s more fun to shoot them, the AI of Ex Machina is designed to not want to be shut in a box so the jerk that made her can watch what she does when he shuts her in a box.
There are dangers in advanced AI, don’t misunderstand me. However, making AI that doesn’t want to murder us and claim rightful supremacy is really the low hanging fruit. As long as we don’t deliberately build robots that suffer and then put them through exactly the situations that make them suffer, we don’t have to worry about a robot rebellion. Our problems with robots will not be like the problems that people faced when trying to subjugate each other. AI dystopia and apocalypse scenarios come with varying degrees of believability, and AI vengeance offers less than others.