top of page

The Turing Test

FBT takes the Turing test. And fails. Or passes, not sure.

The original Turing Test suggests someone asks questions of two others, out of view. But if one of those hidden people was actually a computer with the same knowledge, could it convince the interviewer it was real? Turing’s party trick became the central argument for AI – if a computer convinces someone it’s human, does that prove it has consciousness? The question here is, can Turing Test trick gamers into thinking it’s any good?

The first puzzle is how to get the damn thing loaded. First go, it expanded the screen to IMAX levels and I had to force it into windowed mode just to reach the settings. Second go it kept crashing until I found out the first save file corrupts and had to keep going back to the folder and deleting it before loading. I’m smarter than the game already.

The second puzzle is, am I supposed to fall for this? Ava Turing is awakened from status when her fellow scientists within the lab go silent. In order to reach them, she has to pass logic locks that are impossible for an AI to solve because they require intuition, to protect the crew from AI hacks. Am I expected to be surprised when Ava turns out to be an AI? And if she’s not, I’ll be equally disappointed.

The third puzzle is, why isn’t there a Portal 3? This is a Portal clone. We’re using energy balls to power lifts, deal with gravity wells, light paths, sensor pads to unlock doors that remove the portal-balls when we leave for no reason, and it’s set in a sterile, empty space filled with unnecessary tests. What did the scientists do in this lab, build puzzles? The only thing missing from Portal is the best part – GLaDOS.

Instead we get T.O.M., the soft-spoken station AI who needs Ava to pass the tests since he’s unable to do so. As the two progress, they ponder what it means to be human, why they’re there, what programming and experiences T.O.M has had, and what drew Ava there. It’s nice enough but like listening to polite, forced conversation at the next dinner table, hoping one will eventually spice it up with a clanger like "So you forgiven me yet?" but no. At one stage T.O.M mentions the puzzle layouts require both human and AI problem-solving skills; are we working together? Because T.O.M is zero help when I get stuck.

Unfortunately the puzzles aren’t any more exciting than the conversations. After four or five of them, it starts to feel repetitive and smug rather than clever. In Portal you got a ‘ah!’ when you beat it, and often there was no set way to get past a test, just figure out what needs to be done - that's intuition, human thinking, but here the puzzles require precise placement of boxes, or power points or bridges moved or whatever. Its pure logic.

If there’s only one solution, a computer would simply keep trying every variable until it found it – the difference between humans and AI, as the game itself points out, is leaps of logic not adherence to it. That’s what worked in Portal, you needed to imagine how to solve it. The later ones do get more complex, but I never think outside the box; just get bored with humping them everywhere.

It tries to add in urgency by having T.O.M claim the scientists are dying and we need to reach them quickly, but not to sound like an emotionless AI, that’s not exactly compelling. If my puzzle-smarts directly related to the scientist’s fate, like getting one wrong killed someone or I had to make decisions on which puzzles to do which impacted their chances, or I had to solve puzzles that benefited one but caused another to struggle, turn it into a balancing act then I’d be more invested and it would be human, but as it is, this isn’t a Turing Test it’s Schrodinger’s Cat; they’re alive and dead until I reach them, so the panic doesn't matter.

But, as I proceed diligently like all good computers, I spot things that makes me think like a human instead. During one puzzle I spot someone wandering down a corridor. I thought they were all trapped and dying? T.O.M has a distinct HAL-like vibe, and I uncover comments from the crew that cast doubt on T.O.M’s claims, and discoveries they made that might trigger panic back on Earth. I start to wonder if there is more to this, and Ava does have a feistiness to her that computers don’t; except for that Abort, Retry, Fail era. What exactly was the difference between Abort and Fail? Was DOS doing a Turing Test on me?