Ubiquitous Human Computing and the Future of the Internet
Two people on two different computers a world apart are shown an image of a boy on a bicycle. They are asked to label the image. One of them types in "child" and the other "bike" and so on until they agree on a tag and are awarded points.
The people are playing ESP, a computer game but also a way of harnessing the power of crowds to solve problems like the lack of labels for many images on the Internet. As the game's purveyors put it, "A database of human-provided descriptions of images can make searching for images on the web much easier than it is today. In addition, it can be used to help people with visual disabilities to have a better experience on the web."
ESP was just one example of "crowdsourcing" cited by Jonathan Zittrain in his Oct. 19 talk, "Minds for Sale: Ubiquitous Human Computing and the Future of the Internet." Zittrain showed his audience a variety of Web sites that publish open calls for work on everything from complicated problems like "Probabilistic Modeling of Spending Habits" (Innocentive) to mundane tasks like "Summarize a Web Site in a Sentence" (Amazon Mechanical Turk).
Zittrain, a Professor of Law at Harvard Law School, is a co-founder of the Berkman Center for Internet & Society. The author of The Future of the Internet and How to Stop It, Zittrain speculated on how crowdsourcing might change the nature of work and leisure and the ethical ramifications of those changes.
For example, he introduced the audience to LiveOps, a company that offers "an integrated, full-service virtual call center that leverages a workforce in the cloud." In other words, LiveOps marshals its independent agents to handle calls for any number of different companies. The model provides flexibility for LiveOps customers, who pay only when they use the service, and for employees, who can work when and wherever they want.
On the other hand, Zittrain said, LiveOps can monitor their agents "six ways from Sunday." Calls are recorded, response time is measured, and agents, who are not really employees, can be summarily dropped. "And that's it. It's not like you get any right to a performance review or even an ability to take your portfolio or reputation with you should you want to move to another company," he said.
Also, because agents are not traditional employees, their employment is not governed by current laws regarding child labor, minimum wage, or social security eligibility.
Zittrain was also concerned about alienation in such work environments. "You have no idea who's commissioning [the work]; you can't decide about the fruits of your labor and how they will be used." He worried that a more sophisticated site like Innocentive might post scientific problems whose solution might be put to nefarious uses without the people submitting their ideas realizing that they are contributing to a harmful project.
The same may apply to the mundane tasks or games offered at other sites. Zittrain described a "game" that asks players to decipher "captchas" -that is tests that humans can pass but computers can't. The most common captchas are the distorted letters visitors are required to copy when they register on a Web site. The purpose of these captchas is to discourage spam and other even more serious kinds of computer misuse. But the captcha "game" gets unsuspecting players to decipher these gate-keeping tests so that the answers can be used by spammers, the very people the captchas are meant to discourage.
"This has certainly led me to think that if there was a Nobel Prize for pure evil, this would be an extremely strong candidate," Zittrain commented.
Zittrain suggested several solutions to the problems he identified, including setting labor standards for contract agents, allowing unionization, and disclosing who is commissioning the various tasks that are being crowdsourced.
Jonathan Zittrain's talk was co-sponsored by the Markkula Center for Applied Ethics, the Center for Science, Technology, and Society, and the High Tech Law Institute.