Being Replaced in The Talos Principle

Nate Schmidt, Contributing Editor

This article discusses elements of The Talos Principle’s endgame, so read with caution if you don’t want to know about that kind of stuff.

As you may have heard, the robots are coming for you. Stable Diffusion and DALL·E 2 are going to make all the art; Chat-GPT is going to write all the essays; some other thing I don’t know the name of is churning out fake Drake songs. Robots are driving the cars, delivering the packages, bussing the tables; day by day, “CEO of the robot company” seems to be the only job that isn’t in danger of being automated. Maybe the day has arrived when we need to re-learn how to smash some newfangled looms. One of the most frustrating things about the leaders of the Techno-Industrial Revolution is that they seem to be creating these things just because they can—it’s not like anyone ever actually had a great experience getting help from an automated chatbot.

What with climate change and all, it’s going to be pretty depressing when rendering ourselves irrelevant comes to be our last collective accomplishment as a species. 

I played The Talos Principle for the first time this year, and I suspect a game in which you are a robot wandering through a world that contains only the echoes of an extinct human race hits a little differently now than it would have in, say, 2014, when the game was first released. Of course, there was broad scientific consensus in 2014 that climate change was dire, but we hadn’t yet seen the Camp Fire, or the Maui wildfire, or people dying of unbearable heat in Texas and Louisiana, not to mention the global pandemic that has killed nearly seven million people to date. Since The Talos Principle is ultimately a game about robots in a world bereft of humanity, the wait for the imminent sequel seems to be an apt time to think about how and why this game represents a dire and distressingly imaginable future.  

First—and this may be surprising, considering the way this article is going so far—The Talos Principle is not a depressing video game! It has its share of what I would call poignant moments, but it never veers into dystopia. It is also a visual marvel, full of impressive vistas and glorious architecture. The game has its share of ruins, but they’re Romantic, Wordsworthian ruins, overgrown with moss and ivy in all the aesthetically appropriate places. You play as a humanoid robot who is tasked by a voice from the sky called “Elohim” with a number of puzzles to solve. As you wander from place to place, you collect tetromino-style pieces at the end of each puzzle, and you can eventually use these pieces to access new areas and puzzle-solving tools. There’s just one rule (there’s always one rule, isn’t there?): don’t climb the tower in the middle of the map. Along the way, you meet an AI called “the serpent” that lives in the computer monitors and tries to talk you into climbing the tower, and you stumble upon soundbytes from an engineer named Alexandra Drennan, who seems to have been trying to build something important while her world crumbled around her.

Although the game clearly has a story to tell, you’ll spend most of your time just trying to solve the damn puzzles and get the little Tetris pieces. You can use jamming rayguns to open laser doors and block the other robots that want to blow you up, fly over walls with giant fans, and (eventually) clone yourself and stand on your own head to reach difficult spots. The puzzles get hard as hell, and I eventually started looking up solutions on YouTube, because I really, really wanted to get to the top of the tower. The more God told me not to go up there, the more I wanted to go, so I kept at it even when I got to the point where I didn’t even understand the YouTube solutions anymore. Since the game puts all its narrative weight into this single skyscraper-sized gambit, I was terribly curious how it was going to pay off. I am now going to tell you how it paid off so that I can get to my point, so if you don’t want to know about that kind of stuff yet, go play The Talos Principle and I’ll see you back here in twenty hours or so.

Now you’re back. Pretty fun, right? Except in that snowy castle-looking area where the puzzles were too hard. And now that you’ve played the game, you know that the whole point of it all was to make sure that the robot character could become sufficiently “human” so that it could serve as a kind of replacement for sentient life on an Earth that no longer has any humans on it. Solving the puzzles, talking to the naysayer in the computer console, learning about the people who came before you, trying to climb the tower—it was all so that you could get released into the real world outside. You had to learn what it was like to be human before you could do your real job: bearing witness to a world that requires a sentient being to see it.

In some senses, the game is self-consciously “profound” like a newly-declared philosophy major. If the game’s central question is, “What would it take for a robot to become human?”, the answer, “Living through a simulated version of the Judeo-Christian Eden myth while practicing advanced spatial reasoning,” leaves something to be desired. There is also something problematic about asking who gets to be human without ever addressing all the people who have historically been excluded from that category by colonizers, especially when there is an entire section of the game modeled on North Africa. However, there is a better question lingering behind The Talos Principle’s Eurocentric framing, namely, “To what extent could a human-made intelligence suffice, and to what extent does it fail, as a replacement for the real thing?” This is a question that I think we should be asking more frequently—not because we want to replace human intelligence, but because the limit-point defined by the answer would be well worth knowing. Also, so-called transhumanists like Elon Musk are spending colossal amounts of money and resources on making this question appear like it only has one answer, and we’ve got to be better critical thinkers than that guy.

I recently heard sci-fi author Ted Chiang give a talk about AI, and one of the things he said that really stuck with me was, “Technology doesn’t want to replace you. Management wants to replace you.” In other words, what we call technological advancement isn’t just some abstract thing that is sort of progressing endlessly forwards in ways nobody can help. There are, rather, interests at play who have a vested interest in paying as little income as possible to the smallest possible workforce, and who perceive any increase in automation to be a boost to their bottom line. The problem isn’t inherent to the technology, like in “I Have No Mouth and I Must Scream”; the problem is in the deployment of the technology towards capitalistic ends, like in Do Androids Dream of Electric Sheep?, in which the replicants are created to serve the interests of off-planet mining companies.

In this sense, I think The Talos Principle serves to point us in a direction that has so far gone relatively unexplored as we argue in the public sphere about Chat-GPT writing first-year composition essays and DALL·E stealing images from DeviantArt. Conversations like these are important, but advancements in AI are also opportunities to interrogate cultural assumptions about the meaning of cognition—how we define the things about ourselves that are fundamentally irreplaceable. As the long process of learning to be human in The Talos Principle suggests, all forms of artificial intelligence, no matter how advanced or impressive, are the stuff of mirrors and parrots. In that case, maybe we should be just as concerned about being imitated as we are about being replaced.