Under the Author Information section of an article in Nature last week (August 5), there was a highly irregular reference to an unusual number of extensive contributors: “Foldit players (more than 57,000) contributed extensively through their feedback and gameplay, which generated the data for this paper.” As you might imagine, this throws a small wrench into the intellectual property concept of scholarly publication, in which the publisher reaps the profits, while the honors of priority and attribution go to the identified authors. In this case, the 57,000 anonymous authors suggest something new is afoot.
In the article in question, “Predicting Protein Structures with a Multiplayer Online Game,” nine scientists at the University of Washington (who are listed as authors for this paper along with the Foldit players) describe how they have created a game that players download and play against others online in which they work with protein structures trying to predict how they will fold into their “native form.” It is no easy process to get right, and only the smallest of protein domains have been successfully analyzed to date, and yet a knowledge of these structures is proving vital for research on a number of diseases. Enough said, I trust, to suggest that this is rocket science. And yet with brief tutorials, the Nature article explains, interested game players, with little or no scientific background, are able to use various tools developed for manipulating the protein structures (made up of amino acids) in ways that allowed them to arrive at a stable 3-D model that represents the native state of a particular protein.
The players’ results were compared to a state-of-the-art automated methodology called Rossetta which applies various algorithms to predict the native structure. Rossetta is part of a distributed computing experiment in which hundreds of thousands of people allow the program to tackle protein folding on their computer whenever it is idling (see BOINC). This distributed cognition model takes this public contribution to a further level. And it works. Foldit players out-performed the Rosetta protocol on three out of ten puzzles, and did as well on five, while the two in which Rosetta did better were “basically incorrect,” the article reports. We’ll be seeing a lot more of this distributed cognition model of intellectual property creation, as the University of Washington researchers move forward, thanks to a 14-million-dollar grant from the Defense Advanced Research Projects Agency, in establishing this fall a Center for Game Science.
The success of Foldit speaks wonderfully well to Clay Shirky’s new book Cognitive Surplus: Creativity and Generosity in a Connected Age. In a nutshell, Shirky argues that if we just turn off the television and direct our mental gifts and free time (estimated to be a trillion hours annually) to creating something useful to others, some 2,000 Wikipedia-size projects might be realized. At the very least, those playing Foldit are contributing their own cognitive surplus to biomedical research, turning time, which might otherwise be wasted in playing World of Warcraft, into a public good.
Crowd-sourced intellectual property, from Wikipedia to Foldit, makes it clear that given the opportunity to do something of greater value to others – compared to watching television, updating Facebook, playing video games – people will direct their learning and skills, as well as their computer, to helping others work on large-scale problems.
The school teacher in me is greatly heartened by this “discovery” of people’s interest in continuing to learn, whether how to, in this case, fold proteins or write encyclopedia entries. In the process, they are stumbling on “the intellectual properties of learning.” Only the properties of learning are more like anti-properties, as they are decidedly not about exercising rights of exclusion to protect commercial value, as is typical with property. The intellectual properties that comes of learning are shared goods sponsored, in this case, by DARPA and the voluntary efforts of 57,000 gamers. The intellectual properties resulting from this learning serve others (such as medical researchers), even as their production may also be motivated by a top game score, an article published in Nature, or obscure reasons of national security (with DARPA sponsorship bringing us the Internet).
Now none of this is entirely new. Think of Isaac Newton relying on sailors’ reports about tidal activity around the globe, and Charles Darwin eagerly opening letters from amateur naturalists describing newly discovered species and specimens. What Foldit does is demonstrate an enormous increase in the scale of the collaboration between scientist and public in tackling a complex challenge for our health and well-being. It suggests wonderful possibilities for education. Through the Foldit model of distributed cognition, we might redirect some part of the unproductive cognitive drain of examinations and much homework toward far more socially useful ends. Think of what it might mean for students to realize that their learning leads to intellectual properties that serve the greater common good. It would provide a lovely turn to the double entendre: “Don’t waste your time in school.”