How Google’s using humans as “computers”

Two days ago, my friend John sent me this Ajaxian article about Google Image Labeler. It’s a real-time collaborative game (using ajax), where you work with an online partner (random & hidden) to look at the same image and decide on some labels/tags together.

Note that this game is more productive than you might think. As you “play” this game, you and your partner are actually helping Google’s search engine make sense of images it encounters. If you find this idea facinating (just like me), remember these words: Human Computing.

To explain more about human computing, Google recently invited the best person in this field, Luis von Ahn, an assistant professor in the Computer Science Department at Carnegie Mellon University. Besides his string of academic credentials, he’s well-sought after by corporations worldwide for his groundbreaking reseach, patents and applications. This is the same person who helped invent the CAPTCHA system you see everywhere on various online forms as a means to circumvent spam (you know.

As the folks on Ajaxian noted, “in one year, 9 billion people-hours are spent playing Solitaire; it took 20 million human-hours to build the entire Panama Canal – no wonder the professor talks about finding ways to optimise human cycles. Also, note that a single 90-second game will probably yield somewhere between 50 and 200 labels – admittedly some of them are rushed, but how long would it take to gather that info in most web apps? The professor speculates that Google could label all of its images in two months.”

Aside: Even systems like CAPTCHA aren’t completely free from hacks (abit difficult ones): Optical character recognition (OCR) already works on some levels of captcha (See PWNtcha), and spam house have spawned captcha farms where lowly-paid people are basically made to fill captcha forms all day.

2 thoughts on “How Google’s using humans as “computers”

  1. This is ingenious! The problem with it, though, is that image indexing becomes contingent on level of popularity of the game. Good as a short-term method of indexing, though! I can imagine it could reach out to the bar-room crowd — imagine if those little touch-screen machines at bars were equipped with a keypad of some sort to enter terms, while pointing out on the screen where an item is!

    Maybe not. But fun to think about…imagine all the drunk would-be catalogers in the world, united in a common task…

  2. Here’s a Flickr multiplayer game called fastr where players compete to see who is fastest at guessing the “mystery tag” based on a series of images. It isn’t plugged into the “human computing” mechanism, but imagine if Flickr does keep track of submitted tags and through the use of probability, determine which tags are the relevant ones. Since not everyone bothers to tag their images (I tag everything!), this might help build a more semantic library of photos/images, which in turn lets people find images more accurately. Really, this idea can be employed for almost anything… just be first to do it!

Comments are closed.