The Turing Test

The Turing Test is a puzzle game that is often at odds with itself: it is packed with design decisions that work against the discussions present in its story. It’s a game that contradicts itself at every turn.

You play as Ava Turing, an engineer on board a space station above Jupiter’s Europa, who one day is awakened from cryosleep by an AI called Tom. Tom immediately briefs her on the situation: after encountering a life form on the surface of Europa, the rest of her crew went dark. She’s to go to the station on the surface and find out what happened to them. When she arrives there, however, Ava discovers that the station has been modified by her colleagues, who transformed each and every room into a Turing Test: a test “designed to tell humans and machines apart. Typically problems only solvable by a human. A combination of logic and lateral thinking.”

It’s clear that something is off from the outset. The soundtrack is eerie. Tom’s explanations sound contrived and far-fetched. When talking about what could possibly have made the crew build such elaborate puzzle rooms, he simply ponders, “Perhaps they ran out of things to do here, the devil makes work of idle hands.” Since boredom is hardly a convincing explanation for those rooms, the player is immediately led to wonder if Tom is hiding things from Ava and maybe even using her, since the puzzles, by their very nature, can only be solved by a human being: without her help, Tom can’t explore the station.

The game’s story is surprisingly straightforward, however. There’s this suspenseful atmosphere, as if a plot twist is just waiting for the perfect time to appear, but it never comes. Early on, you find yourself in a room full of notes and audio logs that basically tell you everything there’s to know about the plot – and later events will just confirm this information, instead of flipping it on its head.

The puzzles themselves don’t make a great first impression either. They’re all built around manipulating energy orbs: you wield a gun that can pull them from certain devices and shoot them back at others, powering them up to open doors, activate bridges, or move some contraptions around the room. There are also power cells that serve the same function but require you to physically maneuver them around the room. Puzzles are structured along these lines: you must power the exit door, but if you take the power cell from the bridge next to you, the bridge will retract, stopping you from crossing to the other side. So you have to find energy orbs that you can take without penalty – usually from devices that you have already used or from doors you have already passed through – while pulling and shooting them through hidden gaps or unsuspecting windows.

You can’t pass through those same windows, however, no matter how big they are. You also can’t climb ledges or throw the power cells upwards: all actions that would simply trivialize the puzzles – and these actions are prohibited precisely because of that. In any other game, this wouldn’t be much of a problem, but here these puzzles are meant to prove that Ava is human, so the fact that they only keep reminding the player that she’s a videogame character is appalling, to say the least. It’s ironic, therefore, that a game about a character proving to be human by solving puzzles has those very same puzzles operate within a heavily gamified logic.

Things get even messier as the game introduces a new mechanic around the midway point. Suddenly, Tom is now actively helping you solve the puzzles, moving robots, and activating certain devices. There are two huge problems here: the first is that Tom isn’t supposed to be able to solve anything. That’s the whole point, it’s the reason he awoke Ava and enlisted her help. The second problem is that the crew had no reason to build a puzzle room that requires the cooperation of a machine: they are Turing Tests, after all – and the only machine that could help Ava is precisely the one they’re hiding from. In other words, the game’s story establishes not only that one of the characters solving those puzzles shouldn’t be able to solve them, but also that those puzzle rooms shouldn’t even exist in the first place.

It also doesn’t help that Ava never reacts to the documents and audio logs you find in some rooms. They show the Europa crew second-guessing Tom’s intentions, some of them even going to great lengths to prove that their paranoia regarding the AI is warranted. You hear and read all those things, but Ava remains silent, never asking Tom about anything: they have a brief conversation after entering each room and that’s it. This means that Ava will sound surprised by things that, theoretically, she should already know about, since you have read and listened to them. This artificial response to the events emphasizes once more how she is a videogame character, incapable of reacting to things she wasn’t programmed to react to. In a story that is all about Ava being a creative human and Tom being a programmed machine, that’s a huge contradiction.

In an early scene, you also get to a room with some screens that display the status of each member of the Europa crew and, while Ava Turing and Dr. Sarah Brook have clear videogame models displayed as their portraits on the screen, other characters, like the Russian Mikhail have a picture of a real person displayed on the screen, which once again pulls back the curtain to display how Ava is not actually a real person, undermining the main discussion in the game.

The Turing Test also fails at its climax, as it decides to focus on a certain dilemma but refuses to show its consequences. First, it makes a weirdly binary distinction between humans and machines: it defends that while the latter are entirely rational, the former are controlled by their emotions. Then, it presents a situation to which the human characters act instinctively, thinking only of self-preservation, while the machine thinks of the collective good. There are only extremes in The Turing Test: all human characters in the game are pure instinct and emotion, ultimately incapable of thinking about the repercussions of their actions. The fact that the game’s abrupt ending cuts to the credits without showing any of those repercussions, then, makes everything much worse – and anticlimactic.

The puzzles themselves are fine, even though they operate by very arbitrary rules. To switch to Tom, for example, you must have a robot or security camera in your sight, point at it, and press a button. This means that during puzzles the player will have to keep the positioning of security cameras and robots in mind. However, to switch back to Ava you don’t have to find her on the screen, you just need to press a button to return to her at any given time. Why? Because it would otherwise break some of the puzzles.

The Turing Test ultimately fails at what it sets out to do. It focuses on the difference between human beings and machines, but then proceeds to blur those differences at every turn, without realizing it: although the story paints Ava and Tom as opposites, to the player they will feel exactly the same.

December 09, 2020.