New technology is in the works right now that will impact gaming in big ways. I’m not talking about cutting edge graphics or online infrastructure (although that’s changing too – but I’ll talk about that stuff some other time). I’m talking about shit that could possibly shake the very foundations of our interaction with games in general.
Now, would you kindly let me tell you about the future?
I’ve written before about how games have changed since I was a kid, and I’ve hinted at how Web 2.0 connectivity has changed our expectations, but I’ve been working lately with some academics and theoreticians who are out there on the edge, waiting for the rest of us to catch up. Today, I’m going to talk about three things and how they could impact gaming: software development, language, and AI.
One of the big things about software that some gamers are not aware of is how businesses want it to be made. Software used to be thought of as mostly a product to be created and sold. Not so much anymore. In the 80’s, Johnk Krafcik, working and studying at MIT under the tutelage of Michael Cusumano , looked at Japanese auto manufacturing, encapsulating the process with the term “lean” manufacturing.
Since that time, Cunsumano and others noted that similarities between some of the principles of lean manufacturing and Microsoft’s daily build, as well as “agile” development processes, where the concepts of assembly-line style of production weren’t as efficient as fixing problems as they arose (as opposed to building the product first and then fixing all of the problems). Lean Thinking (by Joames Womack and Daniel Jones) made some of those concepts more accessible for software developers.
The take-away here? Instead of software being thought of as a product, bigger ventures are beginning to mimic services, where the software is never really “finished” per se; instead it becomes a living, breathing service, able to be constantly updated and fixed. This has been the case for some time with MMOs, where the launch is followed by patches and updates as the developers deal with bugs and the increasingly finicky tastes of gamers. You think that Diablo III’s always-connected model is going away? Hardly. We almost have that with the Call of Duty series (I mean, you *could* play it offline, but why bother?), but the structure of the business (and gamers’ expectations) still creates a sense that games are products that are finished when they are bought. Always-connected backbones are the first step towards a lean video game development infrastructure (and it presents an interesting “solution” to DRM).
As these processes become even more mainstream (and they will, because they are deemed as more efficient), gaming will begin to look more like cable subscriptions, where you pay monthly fees for access to living, breathing game-worlds as opposed to single-chunk experiences. Persistent worlds, social connections, and decreased reliance on quality-control departments will all create a much different way of accessing content.
Another thing happening on the edge is the incorporation of automated word sense disambiguation (WSD). What the shit is that? Well, it’s basically the ability of an AI or script to understand language (think about the multiple meanings of a word like bass. Do I mean fish or the thumping from a subwoofer?). And when I say AI, I’m not talking about HAL, I’m talking about the algorithms and codes that serve as the foundation for a piece of software that interact with the user.
Right now, teams of researchers compete, using different WSD algorithms to decode language meaning. The goal is to create a process of disseminating meaning, based on access to dictionaries, in-text inference, and updated prior-knowledge banks. While I can infer from a paragraph of text what most of the words mean, a computer has a much harder time. The ultimate goal is to create a program that can do it without any human interference or coaching.
Some of the craziest work (PDF) is coming out of the University of Houston-Downtown. Shit’s about to get really nutty. Dr. Ping Chen and his colleagues have started to have some successful trials of an unsupervised WSD algorithm that can interpret meanings better than some supervised ones.
The take-away here is that as this technology gets more mainstream, not only could we have Scribblenauts 5.0, where the game could interpret complex and compound sentences instead of just nouns and adjectives, but with the increasingly commonplace use of voice commands (with Siri and Kinect-type tech), a gaming environment could not only understand what you are saying, but it could update its own dictionary the more you talk to it. By connecting to a user-created wiki, it could also update its own lexicon from other users, while learning your specific vernacular.
When you scream “get that bastard right there!” it’ll know who you’re talking about and how you should, in fact, get him (also, with big LED screens and eye-tracking cameras, it’ll know which bastard you’re referring to, judging by where you’re looking).
The last bit I want to talk about is probably the scariest. I tend to be an optimistic futurist, one that thinks that technology has been, and will keep, helping humanity, but there’s something that scares the shit out of me. From a conference on Ambient Intelligence, some researchers have been looking at the relationship between computer and user; specifically, they have been looking at “how computers can be used to change human behavior in order to promote individual, organizational and societal values.”
Let that sink in for a second.
How many times have you had an emotional reaction when playing a game? For me, it’s not simply a question of how many times; it’s a question of how deep that reaction was. Vodka Fremen has talked about some of the possible foundational problems with games as high art, and I think that the issue of fear vs. being scared and loss vs. inconvenience is huge here. Real emotion doesn’t come out of some ones and zeros; it’s brought about by the story and interaction with the characters that my mind chooses to associate and empathize with. I’m using the game for my own reasons; the game is not using me. Right?
Perhaps I’m being used to increase market share for some company, but that’s about it. Go watch Indie Game: The Movie, though (or listen to The Gamer’s Garage talk about the movie), and you’ll see that some developers’ intentions are not just to increase their wallets. Jonathan Blow (love him or hate him) didn’t create Braid to make himself rich; he made it to communicate and change the player (and getting rich wasn’t bad, either, I would guess). So some of the games are more than games; they are avenues of communication, and they are possibly catalysts for change (just like any other artform).
More importantly, the previously-mentioned researchers are looking at ways that human-computer-interaction can be graphed and analyzed to help improve the ability to change human behavior. I’ve mentioned before how developers are using vast amounts of user data to change the ways that they create and maintain game worlds, but the Chomskyan implications here for changing gamers’ behavior should be frightening. Not because of some sinister plot to turn us all into a bunch of mindless communists, but to turn us into a bunch of mindless button-mashers. But, the good thing is that indie developers tend to be a little less scary than a corporation like EA or Activision.
But what happens when even the small developers have an itch to make people more, I don’t know, egotistical or shitty? If they have access to the data (Column 1 shows that when faced with choice A, 45% of the users did action X, but when faced with choice B first, 67% did action X. Column 2 shows the percentage of users’ final choice when confronted with different orders of the choices, confirming that choice-order C was the most effective at making the user choose to kill their companion in the final scene, etc.), will that impact the way in which they create the game worlds? As that process becomes more automated, will the developer simply input a desired outcome, letting an AI create a game world based on existing statistics and data? Is the AI an artist at that point? If the AI learns my language and also has access to all of that other data, will it subtly start to control my ability to communicate? Could it change my language, my outlook, and my opinions?
I like games for their combination of simplicity and complexity (complex to make, simple to enjoy), but my tastes have grown pretty discriminating over the last few years (mainly because of the thousands of choices out there), so I welcome most changes to the industry, as long as they are for the benefit of the gamer. Yes, I get scared for the future. I worry that gaming development will become an automated service, used by ultra-corporations to turn us all into mindless consumers, content with near-permanent mental apathy to keep us from rising up and demanding more out of life. I worry that gear-grinding will replace competition, using data to recommend design and infrastructure, relegating competitive play to he-who-has-more-disposable-time. I worry that someone whose “societal values” are horrible will be in charge of my favorite intellectual properties. I worry that I worry too much.
Do I think too much about the future of games, instead of just playing games? Probably. But I do that so you don’t have to. Go play your favorite game, and bask in the ability to be in charge of the experience. It might not be like that for long.