Duke Ugrad Alum Profile: Luis von Ahn

From the Spring 2010 issue of Threads

Luis von Ahn

At the start of his college career in the late 1990s, it didn't take Luis von Ahn long to switch his focus from math to computer science. "It was really because of the people," he says, looking back. "Duke has such great instructors." The move to CS was a good one for von Ahn, who today is an assistant professor at Carnegie Mellon University, a MacArthur Fellow, one of Discover magazine's '50 Best Brains in Science', and creator of one of the most widely used anti-spam technologies on the Internet.

After graduating from Duke in 2000, von Ahn began his PhD work with Professor Manuel Blum at Carnegie Mellon. Interested in cryptography, von Ahn and his advisor developed a visual recognition test to prevent bots from spamming websites, a string of squiggly letters that humans can identify but computers can't. They called it the Completely Automated Public Turing Test to Tell Computers and Humans Apart, or CAPTCHA. "We knew it was going to be useful, but there were many uses we hadn't thought of," says von Ahn. Today almost every major site on the web uses CAPTCHA to distinguish humans from bots, including Yahoo, Ticketmaster, AOL, and Amazon.

But despite CAPTCHA's resounding success, von Ahn felt guilty. It takes about ten seconds to type in a CAPTCHA, and after a quick back-of-the-envelope calculation, von Ahn realized people waste roughly 500,000 hours every day on his tool. "I started thinking, is there any way we can make good use of this time?" he says. Shortly after, he debuted reCAPTCHA, a repurposed version in which humans help identify words from scanned books that were not recognized by optical character recognition. Facebook, craigslist, and many other sites now use reCAPTCHA.

Von Ahn continues to be fascinated by human computation - using human intelligence to make computers smarter - and he has found one of the best ways to encourage people to participate is through gaming. At www.gwap.com, "Games with a Purpose" which von Ahn premiered in 2008, visitors enjoy games like the ESP game, working with a partner to gain points by agreeing on words to describe a picture. Their answers help to build a database with accurate image labels-so accurate, in fact, that Google recently licensed the technology to improve their own image database.

As von Ahn continues to innovate new methods for human computation, he offers this advice for today's undergraduates: "In college, I thought my goal in life was to get a good GPA," he says, "but it's equally important to get involved with a good professor doing good research. Take advantage of what's going on around you."