Games for Good

Harvard fellow Gene Koo wrote a thought-provoking post last week about how games can be used to bolster civic engagement and democratic participation. He noted two general ways they can be used towards these ends: games for crowdsourcing and games to determine what things people value and how they value them (value discernment).

Games for value discernment are essentially a new way of improving democratic or popular input into the governmental system. Whereas today we mostly rely on polls, lobbyists, and elections to serve as proxies for public support, Koo suggests that we might be able to use games to more subtly tease out people's true values or more meaningfully engage them in the decision making process.

While value discernment is an interesting and noble goal, it's the second application, games for crowdsourcing, that I'm more excited about.  The idea here is that some sort of value is produced as a direct result of the players' actions. While playing the game you are simultaneously doing something productive. The canonical example of this is the Google Image Labeler, which shows you images and asks you to label them:

Projects like Google Image Labeler illustrate how a well-designed game can harness collective intelligence to do productive work. The small amount of work you’re doing for Google is matched by an equally small motivational reward (a score and the fun of playing). While an interest in the project’s goals might lead you to the Image Labeler in the first place, continuing participation is driven by the game, not charity.

While certainly not a new idea, I think this is an area that is largely unexplored, and I'm curious what types of problems can reasonably be solved by such games. I've thought about this a bit, and here's the general criteria I've come up with:

This one's obvious, but it should be stated nonetheless since it's absolutely crucial: No matter what, there must be a strong incentive — most likely fun, but possibly something else, maybe self-satisfaction? — for the players to participate. Absent an incentive to play there's really no point.

The incentives must align. Being more productive ought to directly align with doing better in the game, so that the players' drive to do well in the game maps directly to whatever positive effect the game is designed to bring about.

It's got to be robust enough to withstand nefarious players. There are always going to be at least a few people who get a kick out of trying to screw up the system, so it should take that into account whenever possible. For the image labeler, Google addressed this problem by randomly matching players up and having them act as checks against each other, making it substantially harder to game the system.

Players shouldn't need a lot of context. Put another way, the barrier to entry has to be really, really low. Again, the image labeler stands up well: You don't need any context to start playing, just look at an image and start typing what pops into your head. And each round is only two minutes, so you can participate without committing a substantial portion of your time.

The problem should lend itself incremental contributions. If it's an all or nothing affair, it's tough to get people to participate. Best of all would be an app that allows a gradual ramping up of participation, so if people do get really jazzed about it there's a clear path forward.

There are of course several challenges and pitfalls here. Many games seem to be fun for a bit, but eventually the novelty wears off and people stop playing them. How sustainable could this really be? (One idea I've had is that the scope of the problem could be limited in such a way as to have a  relatively fixed end point, but that's not really a solution to the core problem.)

Another challenge I see is related to what Ian wrote about in Voting Sucks (or: what is constructive involvement?). Many easy and quick types of interaction can only nominally be called “participation,” and it's hard to call them meaningful. But I think this may in fact highlight the beauty of this idea: If the game is designed well enough, it might actually become a positive outlet for what would otherwise be useless or even negative behavior. Again, this may or may not be practical, but history and human nature do indicate that people like to vent and express their outrage at things; why not try to harness some of this for good?

One idea I've had, which is admittedly half-baked, is a game for predicting bus arrival times. It's become abundantly clear that there's a need for better information about when buses are actually going to show up. What if there were a game that rewarded people for accurately predicting when a certain bus line would arrive at a destination? You would also presumably need either a disincentive to providing inaccurate information, or at least a check against it. This could also take advantage of people's frustration; bus arrival information is most useful when the arrival times differ substantially from the scheduled times. This is also when people are most likely to be frustrated. Imagine if we could leverage the energy people put into Twittering about how annoyed they are that their bus is stuck in traffic (I know I've done this) and direct it towards a more positive outcome?

What problems do you think could be solved by games