Spoilers for Observation, Per Aspera and The Stanley Parable (computer games).
As our computers have become more sophisticated, one of the questions which has loomed over philosophers like the sword of Damocles has been At what point can we say that an artificial intelligence is, for lack of a better word, a person?
Games like to explore these ideas. Observation was one, and Per Aspera is another, and although they approach the question from very different angles, on one thing they both agree: the question is hard, and it is easiest just to skip to Let’s treat them like they’re living persons within the first two minutes and make everyone who is sceptical of the idea seem mean and petty.
The fundamental problem with Per Aspera is that it sets up an AI simulation game, in which one, as a serene and timeless intelligence, gently nudges an entire planet towards fertility, and then it throws pieces of story at the players faster than
I they can manage. Not only is the AI assumed by sympathetic characters to be basically sapient and capable of ethical reflection from the get go, but as I painstakingly extended my colony’s reach across the surface, I was always just behind where the game wanted me to be. Search this ruin! Too slow, here’s the answer anyway. What’s that noise? Better investigate. No wait, the rogue machines are coming for you. This requires careful use of your available resources to solve, so whilst you’re planning your next move, here’s the end of the story. Why are you leaving? Come back, there’s so much less of the story to uncover!
The janky story, coupled to a solid, if unexceptional, resource management sim, resulted in a messy and underwhelming experience, because not only did the story’s speed cut me off from exploring at my own ham-fisted pace, it also meant that all the time I had spent up until I was cut off had been wasted. What was the point in spending hundreds of in-game days preparing and planning for my new science outpost if I was just going to get all the answers anyway?
So the question is this: would Per Aspera be a good game if it gave the player more time to think and act?
Although I feel that my experience in playing the game was cut short (I got to what appeared to be the end of the story long before I even grew my first Martian tree), the narrative problems, if not the gameplay problems, were evident from the very beginning. In much the same way that a writer intending to explore spirituality, metaphysics, and ethics probably ought to have spent more than ten minutes getting some idea of the work that has been done in the field over the last six-thousand years, games exploring AI or theories of mind probably ought to begin with some research. Even Kurzgesagt’s short video on the matter could have headed some narrative howlers off at the pass.
As it is, the game presents itself as being an exploration of the issue of sapient AI, but actually the narrative begins from a rather solid and immovable position: AIs can be free-willed, fully realized persons and you, the audience, just need to get with the programme, man! That is not exploration. It is barely didactic. It is an unconvincing, allegedly philosophical veneer on a somewhat tendentious statement.
Take, by way of contrast, The Stanley Parable. Less a game, more an interactive experiment in narrative and the exercise of free will, The Stanley Parable has a darkly humorous script which teases all sorts of entertainment from a concise core question: What does it mean to have agency (in a game)? The loop of the game allows all sorts of different slants to be put on the question, and although it comes to no earth-shattering philosophical answers, it does manage to prod the player’s brain and tug on the player’s heartstrings. Indeed, although considering a largely different problem, The Stanley Parable does a much better job of presenting the meat of the philosophical dilemma in Per Aspera than the latter game itself, and it does so with humour and horror in abundance.
For many players, myself included, a game’s story is a large part of the draw and the fun. Thus, having come to the apparent end of Per Aspera’s story, I did not really feel as through there was any point in sticking around to finish the terraforming that I had started. The characters were not sufficiently engaging, nor the philosophical conceit sufficiently well executed, for me to have become invested. A great shame for a game I had greatly anticipated playing.
In the words of that slide from that Jackbox game, what can we learn from this?
Well, there are two things, two very boring things, two things which most people do not like to consider when preparing to embark on a writing project.
Firstly, sometimes less really is more. Trying to solve all of ethics or all of philosophy of mind is probably going to be tricky. A smaller question, or a particular facet/group of related facets of a question might make for a more cohesive project. What might the ramifications be if we design artificial minds based on natural, human minds? would have been an effective core premise to Per Aspera and would have allowed some of the more illogical and fantastical elements to be ignored. There is no need to wonder why everyone assumes that the AI has volition if it has been specifically programmed in order to have volition.
Secondly, research is not something to fear or dread. If writers skimp on research, it will show, because the premise will be weaker, and everything built on those shaky foundations will be much more treacherous. Asimov, Clarke, and Herbert were all able to realize compelling fiction because of their scientific aptitudes. Mary Shelley’s Frankenstein speaks to us, despite its fantastical understanding of science, because of her exposure to the humanities, and her work is a testament to her exploration of the human condition. In short, novelists read, and by reading, they study those things about which they want to write. A writer must always be a reader first, for how can one write about the mind or the soul before one has thought about what they actually are?