There’s Free Labor in Video Games

David Edery is the worldwide games portfolio manager for Xbox Live Arcade and a research affiliate of the M.I.T. Comparative Media Studies Program. He is also the author of the new book Changing the Game: How Video Games Are Transforming the Future of Business. This is his first of two guest posts.


There’s Free Labor in Video Games

By David Edery

A Guest Post

Consider the following: in 2003 alone, nine billion person-hours were spent playing the video game version of Solitaire — enough to create 500 Panama Canals. And as the popularity of highly “casual games” like Solitaire has grown in recent years, researchers have begun exploring ways to channel the enthusiasm of their players into experiences that solve very real problems.

One of those researchers is Professor Luis von Ahn, a winner of the MacArthur “genius grant.” Professor von Ahn created the ESP Game, which addresses the inability of today’s computers to identify random images.

The game works like this: two anonymous players are matched online without any means of communicating. Both players are shown an image (for example, a flowering plant) while a clock counts down. The players must then type words that describe the image, such as “plant.” When both players have typed at least one word in common, they both score points. More importantly, the players have also unintentionally taught the computer that the picture contains a plant!

More than 20 million labels have been harvested by the ESP Game in just a few years — the equivalent of several million dollars of free labor. Professor von Ahn estimates that just 5,000 people playing for a month could label every image on the web. Notably, Google has adopted the ESP Game and renamed it the Google Image Labeler, which anyone can play here.

Similarly, a team of researchers at the University of Washington have created a game called, which turns the process of determining the ideal shape of a protein into a grand puzzle game. Like the ESP Game, is useful because even very powerful computers have difficulty determining the shape of a protein. Human beings, with our spatial awareness and puzzle-solving skills, are better suited for the job. has attracted a wide variety of players, many of whom are not biologists but who simply enjoy the challenges inherent to the game.

INSERT screenshot

The ESP Game and are so novel that they’ve attracted a fair amount of attention in tech-friendly circles. In the future, however, figuring out how to build buzz will be a significantly more important challenge for the creators of games of this sort.

Fortunately, commercial game developers have been dealing with this challenge for a very long time, and they’ve refined a number of so-called “viral” mechanisms that increase the odds that a person will refer a game to a friend. We explore the challenge of making a video game more viral in Changing the Game; an excerpt taken from that section is freely available online.

The ESP Game and are just two examples of the ways that video games are now being used to solve otherwise difficult problems. In recent years, researchers and businesses have also used games to predict the future, to improve productivity in the workplace, and much more. You could say it beats working, but my point is that it shouldn’t have to.


If WoW could harness players power, what could bIngo players do to help?

Online bingoplayers spend hours chatting to each other, even though they've never met. Ifone could harness that brain power somehow, and use it to the better good, how great would that be?


Ah, the free market in labour.

Computer games are increasingly trading work done by costly developers for work done by users who'll actually pay for the privilege.

LittleBigPlanet, for example does provide a game but it isn't the main selling point, the tools are.

This is reflected in other high-demand professions too.

My father recently had a fence installed by a group of professional archaelogists. They have to do this as there's no decent paying archaeology around any more; it's mostly done by volunteers or even people who'll pay to go on archaelogical 'holidays'.

I also know several highly qualified ecologists who can't find employment because the area is dominated by volunteers and occasional holidaymakers.


if only there were finger-powered generators in those videogame controllers

A E Pfeiffer

Does open source software development qualify as one of these games? :)


This concept is hardly a new one. It's all in the presentation. See Exhibit A, Tom Sawyer and Fence Painting.

Dick King

Apropos of Ender's Game and the like, I recall that when the requirements of Reagen's so-called "star wars" software were being described, it amounted to a mediocre performance by a teen-ager in a video arcade. Choosing the correct missile to engage each of 1500 targets over a half hour period? One per SECOND?

Of course, the military gets it: the Predator, for example.



Fascinating piece. I watch my 6 and 9 year old click compulsively on webkinz or kongregate and i wonder if kid power can be harnessed.


No, no, no, no, no. 20 million labels is not the equivalent of several million dollars of labor. It is the equivalent of the incentives given to create the output. I've been trying to convince foster child teenagers that the amount of time they spend arguing easily equals the sum of all of their necessary chore time plus plenty of phone time with their friends. They don't buy it because they aren't interested in my problem set.

The premise is not completely wrong, thought, it is just badly stated. The premise gives the impressive that we have massive wasted resources that can be easily employed with incremental changes to game technology. What it does suggest is that there are new types of jobs that we now have a trained workforce to perform. Now if we can just give them an incentive to do so when it benefits US.


Have you seen ReCaptcha?

It's those obscured words you have to type in to prove you're not a spambot but what they've done is every time you get two words and one of them is scanned from a book which the ocr can't read. So to post on someone's blog you have to help archiving written material.


An (already) classic sci-fi book called "Ender's Game" by Orson Scott Card describes exactly how skills required by the game players can be (involuntarily) applied to more practical problems and achieve great results.

Thomas Brownback

Games that get players to do work are fine, but there's a larger untapped resource in distributed computing.

Some players play MMO's by leaving their computer on just running macros all day long, so they can level more quickly.

Most MMO's discourage this sort of activity, but I think they should encourage it, with a slight change. Imagine a game where you gained levels more quickly by contributing cycles to Folding@Home or some other public computing project.

Knowing the way power gamers work, they'd probably cure all disease in a few months. :)


Online games are already benefiting indirectly from free labor. Fans create interface modifications, maps, and content for a variety of games. The best ones usually end up distributed by the game company or absorbed into the game as features.


Darryl - you gave me a phrase that I would up using twenty minutes later to describe student behavior in regards to school property, thank you; and I think you're exactly right.

If WoW started building in "double-up" problems a lot of the playerbase would be turned off by the idea because the concept violates the core conceit of wasting time - that the time is being wasted. Destructive actions, from consumption to blowing up garbage cans, is its own reward structure. In order to avoid turning the audience off to the exercise you need to either hide it very, very well such that the user is never aware of what they're actually doing, or provide it as an opt-in program with its own incentivisation, such as a reduced monthly fee for those who participate.

You see this in the GoogleImages game, where participation is explicit in its efforts and those who think it's lame choose not to participate, or participate anyway because it's something to do. Retention of lame-sayers is probably really, really low.

Trying to sneak it in, though, is as bad as high school assemblies where the administration promises a movie and delivers a morality lesson.



Perhaps in the future MMOs like World of Warcraft will find ways to seamlessly work in problems that are hard to solve by computers into core gameplay. They already have millions of people grinding, who would do anything for XP. If they could sell those human-hours, it could up their income and possibly lower the price to subscribe to those games.

I remember reading on, a website about obscure gadgets and tech issues, an editorial on how video games are good at teaching people counter intuitive elements of math. Perhaps the next step is to get people to apply these tactics for real problems. Plus, real-world problems a computer can't solve but can check could make for good puzzle in a game since it challenges people in new ways.


Games have the ability to completely absorb people's attention. But right now they're only used mostly for recreation. There are a few very crude educational games like Reader Rabbit or Math Rescue that work for kids.

Imagine if the same information a university course provides could be slipped into a game. It isn't hard to imagine - a Matrix MMO game that slips in actual computer programming lessons, growing progressively more complex over years of playing. A cold war RPG that requires accurately identifying a variety of political figures and countries. An office version of The Sims designed to train human resource professionals.