The Fall

In The Fall you play as ARID, a combat suit’s operating AI. Your pilot is shot out of the sky and is knocked unconscious in the titular fall, thus activating your control of the suit. You have to navigate the facility in which you’ve crash-landed and find medical treatment for your injured pilot.

This review will be spoileriffic, as I’ll be going over the entire plot. If you want the cliff notes: don’t buy this, it’s one of those “to be continued!” scams and is essentially just the origin story for the real game. The controls are also awful and it’s full of bizarre frustrating adventure puzzles. Watch a playthrough on YouTube if you’re interested. The story does some interesting things with AIs, but I found the execution very lacking. It also does the “we’re going to force you to do something horrible and then insist you’re horrible for doing it” thing, which is just so lazy and trite and I am so numb to it at this point.

At the start of the game, you discover that most of your suit’s capabilities are locked off and can only be activated if the pilot’s life is in immediate danger. This includes the medical scanner for your pilot, which was damaged in the fall. To actually get into the facility you need to purposefully place your pilot in danger to trick the system into unlocking your abilities, which the AI is not supposed to do. This is reinforced by the presence of another ARID whose pilot is confirmed dead; they acknowledge that they no longer have any purpose and it is best for you to salvage their power cell, but when they learn you plan to use it to activate the security guns to endanger your pilot, they immediately declare that you are malfunctioning and should be decommissioned. ARID herself begins making a log recommending she be evaluated and decommissioned for her breach of protocol, but then the screen glitches and she comes up with a rationalization for how endangering the pilot was necessary for the greater goal of reaching the medical facility, and she has therefore not violated protocol.

My first thought here was that ARID was a rogue AI who was lying about her real objectives and was already adept at subverting her restrictions. This, however, is not the case; it’s later made clear that she is genuinely trying to operate within protocol the whole game. This really confused me: if she’s behaving properly, why is she displaying this kind of deviant behavior this early on? Why does she want to edit the log, if it’s standard behavior for AIs to sign their own death warrants? The ending reveals that the entire story is setup for how she becomes a rogue AI, so possibly this is supposed to be a sign of a gradual descent into madness, but that’s not the way it’s conveyed: we see her explicitly lose her operating parameters the moment she becomes unbound, implying she was not displaying deviant behavior up to that point.

When ARID enters the facility, she meets an extremely sophisticated administrator AI who talks and behaves like a human. He informs her that she’s currently flagged as foreign and hostile, and the only way to get access to the medical facilities is to pass a series of tests that will reclassify her as a domestic droid. This is where most of the stupid adventure puzzle stuff comes in; the facility is in terrible disrepair so none of the tests work properly, and you have to break everything to technically pass them. (It’s a clever idea, but the solutions are rarely clear and often force you to run all over the place.) The administrator genuinely tries to give you as much help as he can, but insists he can’t actually spoof the tests for you without going against his own protocol. Along the way, you encounter a caretaker droid who keeps insisting you shouldn’t be here and throwing security robots at you, but doesn’t actually do much to stop you.

Passing the final test requires you to lie, which ARID cannot do, as one of her operating parameters is that she can’t misrepresent reality. She decides to hack the administrator’s mainframe instead to spoof the test, but the console is in a flooded area. The administrator says you can drain the water by draining power from all the droids in storage, which will kill them. (Because yes, the ancient facility that is literally falling apart has massive hangars full of perfectly preserved and powered droids, because that makes sense.) He wrings his hands over how terrible it is that you’re willing to kill so many droids just to save one human life and insists “this is your choice”, except it isn’t actually because there’s no other way to proceed. The game even does this dramatic “Are you REALLY SURE you want to kill ALL THESE ROBOTS you monster???” message before you press the button, and generally it’s just trying way too hard. (It then makes you do it two more times, except the administrator only gives you another guilt tripping the final time, which completely ruins the impact; I already committed to this, you are not going to make me feel bad now.)

I found this to be a really jarring dissonance between gameplay and story. The point of this kind of scene is to make the player feel culpable or whatever, but I didn’t feel that way at all. I am not the one who wants to do this — I spent a lot of time looking around to see if there was another option, even. This is purely the choice of the player character, who is operating on protocols and an urgency that I do not possess, and so I’m just along for the ride. Making it “my choice” is totally backwards — it would have been far more meaningful to make it an uninteractive cutscene, as that would have created a stronger emotional resonance to the idea that AIs are slaves to their protocols. The entire point is that this isn’t her choice, she has to do this because she is programmed to value her pilot’s life above all others, which is further reinforced by the administrator saying he wants to help her but can’t violate his own protocols either.

When you finally get to the mainframe, the caretaker sneaks up and disables your life support, on the basis that not only are you faulty, your pilot is faulty for trusting a faulty AI, because a droid having the authority to kill humans in this setting where they regularly scrap droids for the slightest infractions sure makes sense. ARID promptly flips out and the administrator is so distressed by this that he does violate protocol to fix ARID, at which point the caretaker uses that as an excuse to wipe his personality back to factory default. ARID then chases the caretaker across the facility and has a boss battle against him. This prompts him to say:

Caretaker: You are faulty.
ARID: You are irrelevant.
Caretaker: Unit exhibited repeated malfunction since arrival.
ARID: I have followed my protocols. I always follow protocol.
Caretaker: You resolve challenges by circumventing protocol.
ARID: I adapt.
Caretaker: You cheat. Lying is not explicit, your actions are implicitly dishonest. No rules bind you. Deviant…

Somehow this deletes ARID’s “don’t lie” parameter because… why, exactly? The caretaker has spent the whole game insisting she’s faulty, and doesn’t really say anything different this time. Why does she accept his accusation now? The best I can tell is that she was always lying to herself and she only now consciously admits it, except we are explicitly shown the operating parameters in the menu screen and it only now disappears. The binary “has parameter”/”lost parameter” notification muddles this. Was she always technically capable of lying, but chose not to because she didn’t want to admit she could?

Then she finally gets access to the medical facility, and it turns out there isn’t any human inside her suit at all. This revelation somehow deletes her remaining operating parameters and leads to her becoming a rogue AI unbound by any rules.

What happens next? BUY THE SEQUEL!!!

…So, I don’t really get this. It’s clearly trying to say something profound about AIs, but there’s no consistency to it. There’s a lot of jabbering about protocol above all else, but what on Earth were the caretaker’s protocols? If he’s supposed to be maintaining the place, why is it in such bad condition? If he believes ARID is dangerous, why doesn’t he do more to stop her? The best I can tell is that he really hated the administrator and was using her to bait him into violating protocol, except he seems to have the authority to delete anyone he wants, so if he already believed the administrator was malfunctioning why did he need an excuse? And why did he not only kill everyone in the facility but continued to kill random scavengers who showed up? “Thou shalt not kill” is usually pretty high up there on the list of robot commands, especially for something as innocuous as a maintenance droid.

Also, what did happen to ARID’s pilot? Who put together an empty suit? How did ARID not realize she was empty? Yes, her medical scanner’s broken, but she should notice all sorts of other inconsistencies, like that fact she’s a lot lighter than she should be. Did someone purposefully design her as an unbound AI but also designed her to think she had operating parameters? And just how self-aware is she supposed to be? She initially seems to behave very robotically, but she displays clear emotion and varied intonation in her dialogue: she is snappish and sarcastic towards the administrator, she clearly sounds defensive and aggressive against accusations she’s faulty, and is very panicked and upset not only when she thinks her pilot is in danger but when the administrator gets wiped too, which shows she’s capable of forming connections to new people.

The administrator is the most self-consistent, but his behavior doesn’t match with the other AIs; he explicitly calls ARID out as being unusually robotic. Maybe as a helper AI he naturally trends towards more nuance and emotional complexity, but it still doesn’t give us much of a baseline for how AIs are supposed to behave. Given what happens, it’s also unclear whether he intrinsically believes in his protocols like ARID, or if he’s just afraid of reprisal by the caretaker. Generally I’m just really sad he got wiped, he was really nice. :(

So generally, I just didn’t get what this game was going for. We don’t get enough data points about the characters or the world to get a clear baseline of how AIs work in this world, and therefore it’s hard to tell how significant all these deviations are. The game had a lot of opportunities to use the mechanics to make us share the AI’s feelings of constriction, but flubbed all of them. By the end I didn’t have much emotional investment in what was happening, and wasn’t surprised by the twist at all; pretty much the only other possibility was that the pilot was already dead, which I honestly think might have worked better.

One Comment

  1. Nerem says:
    It also does the “we’re going to force you to do something horrible and then insist you’re horrible for doing it” thing, which is just so lazy and trite and I am so numb to it at this point.


    I think Spec Ops The Line is the only game I’ve seen really do this right. Considering making you think about what you’re doing in a game was the entire point, I guess it’s not surprising.

Leave a Reply

Your email address will not be published. Required fields are marked *

Skip to toolbar