As a species we have a curious habit of anthropomorphising almost everything (an example being the way people are internalising, in some way, the crop of beloved celebrities dying in 2016 as an act committed by the year itself on purpose). When it comes to artificial intelligence this is a little more difficult. At what point does an AI stop being anthropomorphised and start having an actual personality? And does this have extra bearing if you're running through a series of futuristic puzzles on one of Jupiter's moons, and you start to suspect your AI companion knows more than it's letting on and is also a bit smug?

This is the situation you find yourself in when playing through The Turing Test as Ava, an engineer for the International Space Agency woken unexpectedly from cryo sleep on a space station above the moon Europa. The mission's AI computer, Tom, informs you that he lost all contact with the ground crew roughly 20 days ago, and you need to go and see what's up. When you arrive at the research facility you discover that it's been completely rearranged into interlocked puzzle rooms. Together you and Tom work your way through them, whilst he tries to figure out what happened by piecing together his disparate, silicone memories, and describing the process out loud in a calm, Alan-Rickman-meets-Jeremy-Irons voice.

The isolated world you explore in The Turing Test is very well put together, especially combined with the music, which is cold, delicate, and barely noticeable. Some of the better parts of the game are when you're poking around in your crewmates' cabins and judging how poorly their bunk is made, or the weird way they line up HB pencils on their desk. The game is, in a way, about personality clashing with logic, and so you see little splashes of individuality hidden in a white, uniform research facility.

Solving the puzzles is the mechanical meat of the game. It's obvious The Turing Test has been heavily inspired by Portal and therefore manages to never appear quite as original in its puzzle format - although, of course, it never really could - and Valve veterans will feel a bit of deja vu tickling the back of their brain. The puzzles themselves involve moving around power sources to open doors, and are generally satisfying, but there are sections where you know what you have to do, but it relies on the game's physics to work as well, and it's slightly frustrating to have to do the same thing over and over again because a box won't fall in the right direction. The whole process becomes increasingly complex as more mechanism are introduced, and the game often tricks you into over thinking in a sly way. It even tells you not to right at the start, where Tom explains to Ava that the puzzles appear to be Turing Tests that only a genuine human intelligence can solve - they require creative thinking, and he's incapable of thinking outside the box.

Tom and Ava's conversations, ongoing as you enter each new test chamber, form the other part of the game. The mystery of what happened to the rest of the crew is tangled up in it, and you can discover fractions of audio logs and notes that fill in the shape of their story, which involves a few worn 'trapped in space for years' tropes (including, without being too specific and spoiling anything, surprise pregnancy, running out of food, discovering a strange lifeform, a loner going a bit weird, and so on). More interesting, however, are the logs you discover about Tom, a sentient AI struggling to find out what it is to be human and not understanding why he isn't. There are discussions of not only the Turing Test, but other famous thought experiments like the Chinese Room.

Ava and Tom are both voice acted fantastically, and their performances deliver a lot of nuance and quality to the game. They start by having quiet, practical conversations about how to find the crew and complete her mission, but end up exploring the theoretical limits of Tom's programming, of consciousness, and of free will. Eventually you discover a twist in their relationship that you almost certainly won't have seen coming.

Just because you think you are a real person, and you think you have free will, does that make it so? It's an unsettling proposition, especially since the end credits start by telling you that you have 'passed The Turing Test'. Taking the game as a whole it can be quite accurately described as Portal puzzles meets Soma, a game exploring similar ideas of AI and selfhood that came out last year. The trouble is, I suppose, that both those games did each aspect better.

Version Tested: PC