It’s the crossover event that this whole website has been leading up to. Four years, over 350 articles, and we’re finally talking about religion in video games. Fuck yeah, buddies. The Talos Principle is a 2014 game made by Croteam, who made Serious Sam. It’s about puzzles and AI and God, and I fucking hate it. I know I’m supposed to like it, it’s very on-brand for me – it’s thinking seriously about religion and making big claims about consciousness and faith and whatever – but fuck, I just hate it.

Let’s start with the basic set-up, anyway. I’ll try and give it a fair shake as far as possible. You play as an AI in a computer simulation. You have to run around and solve a bunch of puzzles, and God sits upstairs doing voice-over. It’s not actually God – it’s like a creator program slash creator figure – within the structure of everything it’s basically God, and it calls itself Elohim, which, like, congratulations on having someone in the dev team who’s heard of Ancient Hebrew, I guess. The whole thing is structured like a Garden of Eden scenario – so Elohim introduces itself as your creator, it talks about a covenant, you’re not allowed to go up the secret tower (“For in the day that you do, you shall surely die”), and then there’s a deceiver in the computer terminals called Milton – which, again, credit where credit’s due, they’ve definitely heard of Paradise Lost. Great job.

So it’s Portal set in the Garden of Eden with an AI twist. Not the worst premise in the world. I think the Portal thing is where my problems start. A while back, I wrote about the Enlightenment heritage of games like QUBE 2. It’s a heritage shared by all those puzzle games, Portal and so on. There’s this basic idea that if you just sit down and think hard, the best thinking will inevitably rise to the top, and we’ll be able to improve and progress and eventually we’ll fix society with new technology and rational, logical thinking, and ah shit the planet’s on fire and it didn’t work. The Enlightenment belief in progress has taken a battering. It doesn’t seem plausible, for a bunch of reasons. And yet all these puzzle games are still going at it. That was my basic criticism of the QUBE games, which essentially say ‘no no, just a few more contrived logic problems about moving blocks and redirecting beams and opening doors and we’ll definitely solve this impending world-shattering event.’ Portal at least embraces the meaninglessness of it all as part of the fiction. In Portal, you do the puzzles, but there’s no meaning to it. There might have been, once, but now all the scientists are gone. The observation rooms are empty, and you just have to do the puzzles because the computer doesn’t know how to let you do anything else. It’s later additions to the genre – QUBE, Talos Project – that tack their big implications onto things. Do logic puzzles, save the world. Or, in Talos Project, do logic puzzles because God told you to.

The religion thing is where I really get my skates on. I’ll say this in favour of The Talos Principle: the narrative structure of creature and creator is a good vehicle for talking about artificial intelligence and machine consciousness. The Garden of Eden is a useful metaphor for having that conversation. The way they use the dynamic of sin and obedience is smart. So, ah, specifics – the whole thing here is that you’re an AI and you’re in this program and the God program is testing you to find out if you’re able to think for yourself. It’s looking for machine consciousness by testing for disobedience – you’re told not to go up the tower, and if you go up anyway, you win! You’re capable of being disobedient, and you’ve therefore proved your own sentience through your use of free will to disobey God’s direct instruction. The religious framework is a clever vehicle for framing that dynamic. I appreciate it as a way to think about sentient AI. I don’t appreciate it as a way of thinking about religion.

And maybe I’m reading against the grain a little here. Arguably The Talos Principle isn’t really interested in religion per se. It’s interested in AI, and it uses religion as a vehicle for talking about AI, but it’s not really directly interested in religion. So in that sense, maybe it’s a little unfair to be so hostile to it. On the other hand, the way that The Talos Principle characterises religion serves as a good case study for some of the wider issues that I have with this underlying Enlightenment attitude in video games. It crystallizes the problem even though it’s not the worst offender.

First problem then, broadly speaking, is the mechanistic vision. When you’re working with this strict scientific framework, where everything is quantifiable and containable, all the interesting little un-scientific un-quantifiable bits of life fall out the bottom of the picture. Games like The Talos Principle are always going to be restricted to levers and buttons and pushing around boxes. They’re logic games. Move A to B, move C to B, move B back to A – they deal with things that are quantifiable. You’re not going to see a puzzle game where a central mechanic is love, for instance. The closest you get are the dating games, which, of course, are criticised for reducing love to a series of systems. Insert enough friendship coins and sex falls out. Something un-mechanical reduced to a set of mechanics.

This problem is specially thorny in a puzzle game touching on religion. Where we might commonly think of God as infinite and incomprehensible, the only God accessible here is bound by the terms of the machine, by the parameters of the code. How do you simulate the supernatural within a deterministic system? The whole idea of the supernatural is to be beyond the natural, beyond the laws of physics or any other governing principles of the universe. Unsurprisingly, then, the god in Talos Principle ends up being not real. It’s a computer program – God subordinated to the machine by a game that believes that everything is ultimately subordinate to logical, rational rules. You win the game by disobeying God and climbing the forbidden tower. You repeat the original sin, and with your roving, inquisitive eye, you realise that religion was a fiction all along. It was never real, and by refusing to follow its precepts, you surpass it. The fiction falls away, and you are released into the real world. Of course, if you choose to obey God, you have failed to prove your independence, and are terminated so that the next iteration of tests can begin. Even though it’s meant as a comment on AI, not religion, if you read it from a religious perspective the implications are pretty awkward. It’s basically a game where believers get killed for quote-unquote failing to think for themselves.

As a final note, we might consider the reality that the AI is sent out into. In The Talos Principle, humanity has destroyed itself through global warming, and the world is empty. The AI emerges in that context as the legacy of humankind, carrying something of their history and imprint. Despite its Enlightenment vision, The Talos Principle suggests that ultimately the scientists weren’t able to save us. Rationality and logic were insufficient. Their entire fictional world is premised on the failure of the Enlightenment project – and yet they’re still very keen to give it another go. Is that weird, or encouraging? Is it an appropriate acknowledgement of the problem, or is it just repeating a bad method and hoping for different results? I don’t know. I don’t like it.