Event[0]: Artificial Empathy
So I’ve finally played Event[0], and after my first play-through, I immediately felt the need to go back and play it again. Not because I was blown away by the writing, or the mechanics of Kaizen, but rather because I left that experience confused, rather than satisfied. I was gifted this game by a close friend, and the short two hour experience I had ended up leaving me with more questions than answers.
The game was spawned off a singular mechanic and character, Kaizen the ship AI. In an interview with Gamasutra, Ocelot Society notes the inspiration of Kaizen to that of HAL 9000 and Glados, and it shows. The main plot point is that you’re a random astronaut of the Europa-11 mission, and of course it’s gone wrong. You end up on the Nautilus, a space station orbiting Europa pulled straight out of the 80s. At this point the mission is scrapped, and you have one goal: to get back to Earth. To help you get there, is Kaizen, an AI with access to the ships core functions, a ton of information of the crew that managed this ship, and some sass.
Kaizen was a strange companion. You can communicate with him in natural language, and thanks to some investigative work done by Mark Brown, it’s an elegant solution in the context of the game. Using some standard NLP techniques, a large data set, and some pre-crafted responses, Kaizen can banter, offer assistance, and guide the player towards completing their objective: getting home to Earth.
As you explore the ship, there’s definitely more going on that meets the eye. You learn about the previous characters of the ship, mainly Anele, the creator of Kaizen, Nandie the ship’s captain, and Kurt, the superior to the crew on Nautilus, and your employer back on Earth. The story is told through Kaizen’s ramblings, various logs scattered across the terminals, and through the world itself, from the level aesthetic, to literal scramblings on various notes and surfaces, like some mad scientist blitzed the entire station with her calculations
There is a central item of conflict that pushes the plot forward, with major conflicting goals that get the player to think, and Kaizen’s overall role towards helping the player make unique decisions on what they feel they should do next. For the record I recommend this game, but there are a few things I found worth exploring below.
First, this game short. For $15, it’s decently priced walking simulator with puzzles mixed in here and there. I’m by no means a puzzle guy, my readers and viewers will know I’m fairly dense when it comes to this kind of stuff (I mainly excel racing games). But nothing feels like it overstays its welcome. The game is decently paced, and the transition from plot twist to plot twist has a good flow to keep the player invested.
Second, the world is well crafted towards bolstering the overall story of the game. It’s not The Division levels of detail, but the devs scattered about a number of clues to give insight into the tension between Nandie, Anele, and Kaizen. Simple paintings, text logs, Kaizen’s thoughts and “feelings” towards Anele and Nandie, the state of the ship, and more clue the player in towards piecing together the timeline of affairs that led to what happened, who’s where, and why.
Third, the interactions with Kaizen make this the most innovative game I’ve played in a while. Usually character interactions are scripted, or hard-railed conversations designed to drive down a certain path. But in this game, everything feels free flowing, and there is no sense of walls put up by developer. Kaizen is at the mercy of the player, for better or worse. According to research done by Mark Brown, NLP coupled with a tagged database of words and phrases, plus a semantic parser, all combined together allows for such interactions with Kaizen. The output is tightly coupled to produce some more meaningful interactions and responses to the player, as well as designed to guide the player towards new states, where Kaizen gains new context about what state the player is in, as well as the “emotional” state Kaizen assumes due to how the player responses. This is achieved when you say thank you a lot, or dismiss kaizen with crass and rude behavior.
When it falls apart, it’s easy to see. I’ve gotten frustrated with Kaizen on multiple occasions, as his responses feel disjointed from the conversations we’ve having. It’s an attempt to put me back on track towards the main objective, but it has the ability to throw of the player very quickly. Some have suggested altering Kaizen’s persona to behave a bit more sporadically and fragmented, so that such disjoints feel acceptable, but of course, we will never truly know.
But there is something else about Kaizen that makes him unique, and that falls back to how the developers crafted his responses. There is nuance, and some sense of vulnerability. When you meet him for the first time, he shows joy. When you argue with him, he becomes closed off and refuses to cooperate. When you learn about the world and characters, certain queues and questions allow Kaizen to gush and open up; his “opinions” per say.
This is incredible, as it inject a sense of personality, and therefore empathy and trust, towards a machine. This dangerous game we play to simply escape the ship, allows us to better dictate the way the game will end. Will Kaizen trust us as we head towards the final objective? Will he sabotage our chances of getting home? How does he feel about me, the player? I’ve got to hand it to the writers, the many play test sessions they’ve had over their 3 year development cycle really shows in the character Kaizen displays during various interactions. It’s not perfect, he’s not a full blown AI. But with the context he has, the emotional state he assumes, and the free flow language he accepts, allows for a gameplay experience to really shine as the centerpiece. I’m almost severely disappointed the game is only 2 hours long.
Fourth, I struggle with the multiple ending scenarios. I’ve yet to play through these (see later twitter rant), but for some reason, the interactions between Kaizen feel “in the moment” and more geared towards messing with the players emotion, rather than setting up a series of meaningful interactions that affect the outcome of the game.
My initial run felt relatively neutral. I was pretty polite with Kaizen, even saying thank you numerous times, but any time Kaizen snapped at me, I bit back. As a result, at the end of the run, I felt as if the game pointed me towards a singular solution, rather than the options for many. None of the previous events really showed off how things could have been different? This is why the game warrants multiple playthroughs as these types of questions can’t be answered with a single run, and just watching some YouTube videos of “secret ending x” doesn’t help sell the experience.
In conclusion, I admire this game. Walking simulators are the types of games the industry needs from time to time. And with such games, comes the opportunity to explore really interesting ideas and mechanics, such as Kaizen’s AI. And yes, it’s worth the $15 dollar asking price. Hell, Abzu was $20, and that game is an absolute gem of a “walking simulator”.
It also gives some insight into what the industry can accomplish on an even grander scale. Of course NLP via a keyboard can only really work on a keyboard, but with the right technology, processing power, and data, your next top-down RPG could implement the same kind of ideas. Imagine Pillars of Eternity with the free flowing conversation Kaizen offers, giving a chance to have the player run the conversation, rather than guide them through a series of bricked responses. That has me really excited for the future, and one I hopefully get to witness soon.
References/Research
Gamsutra Interview with Ocelot Society -https://www.gamasutra.com/view/news/285902/QA_Ocelot_Society_on_building_Event0_around_an_AI_chatbot.php
Wired Event[0] Review - https://www.wired.com/2016/09/event0-review/
Mark Brown’s GMTK Event[0] Game Mechanic Investigation - https://www.youtube.com/watch?v=bCJw4hQkPj4
