by David Alan Grier
Stars and Patents. The two don’t seem to have much in common but they are things that offer the same kind of opportunity to crowdsourcing. Both astronomy and patent law need people to keep track of events in a large and complex entity. Astronomers follow the activity of the night sky, while patent examiners have to compare the new applications for patents to the ever-growing descriptions of technology that have been used in the United States. While astronomy has had a long history of using crowdsourcing, a history longer than the term crowdsourcing itself, the U.S. Patent Office is moving only slowly towards crowdsourced patent search. The difference between the two applications of crowdsourcing reflects the different social structures that we have built in the two fields and the changing nature of search technology.
The difference between searching the night sky and searching the U.S. Patent Office records can be summarized in a single phrase: “There is no money to be made in discovering a new comet.” Astronomers have long relied on the crowd, on legions of amateur astronomers, to monitor the sky. The major professional observatories simply cannot keep track of every event in every sector of the universe. As a result, it has relied on amateurs to gather and analyze data from the different sectors of the sky.
While there is no money in astronomy, there is a glory in discovering a new object or a new event. At times, the professional and amateur astronomers have fought over credit for finding a new comet or new supernova. Even with such conflict, the professionals have usually acknowledged the contributions of the amateurs.
Recognizing that amateur search has played a role in other fields, the U.S. Patent office did an initial test of crowdsourcing in 2008 and 2009 with the Peer to Patent experiment. This effort attempted to determine if a crowd of amateurs might improve the search for prior art among the patent office records. Overseen by New York University, the Peer to Patent program used a group of inventors and patent attorneys to search for prior art. It was based on the idea, similar to that used in astronomy, that inventors and patent attorneys would be familiar with the patents in the part of the field in which they worked and hence would be more easily able to identify prior art than professional examiners.
The Peer to Patent Program was a limited success. In a way that can only be communicated through a double negative, the Patent Office found that their ideas about crowdsourcing were “not untrue.” The crowd was able to find some information that the patent examiners could not find and sometimes this information was of use to the patent examiners. “Peer-to-Patent is growing,” noted the report on the organization. “This is apparent in the increase of involvement and discussion across the board, ranging from government recognition to new applicants wishing to become involved in the program.”
However, the Peer to Patent test discovered some issues that might limit the application of crowdsourcing to patent search. They discovered that patent applications were highly technical documents and that most inventors could not easily understand them. They discovered that the crowd often needed information that often did not accompany a patent application. Finally, they discovered that some of the people who were in the best position to identify prior art for a patent were those with an interest in the field and that these people often did not have a financial interest in disclosing what they knew about any patent applications.
The Patent Office is in the process of reviving its program in crowdsourcing patent search. They held a seminar on the subject in December 2014 and have been soliciting comments from interested individuals about their views on crowdsourcing. Surprisingly, many of these comments have focused more on the information that is available to amateur searches than about the process of amateur search. They were concerned that key information was not accessible to support amateur searchers. Such information “contain numerous clues about claim construction,” remarked one of the sent comments. “These clues must be available to effective crowdsourcing efforts” as “claim language, standing alone, is often vague and impenetrable.“
Government offices rarely move quickly. Even though patent search has a certain resemblance to astronomical search, it will probably require at least another year before we see another test of crowdsourced process for discovering prior art. Even if we get such a process, it will likely operate in an environment that is much more structured and much more organized than the one that has supported astronomy for nearly two centuries.
About David Alan Grier
David Alan Grier is the principle technologist for Djaghe LLC and an associate professor at the Center for International Science & Technology Policy at George Washington University. He is the author of Crowdsourcing for Dummies.