Posts Tagged ‘crowdsourcing’

Crowdsourcing the Prior Art Search Process

January 31, 2010

Searching for better ways to find relevant prior art to determine whether a patent is truly novel and nonobvious has been a constant challenge, both for the PTO when examining patent applications and for courts in patent litigation. Simply searching through databases of patents and relevant literature captures only a portion of the available prior art that may be relevant. Sometimes, the most relevant prior art is tucked away in a library somewhere, or an uncatalogued journal, clearly published but not easily found. Only a handful of experts may know of its existence. Unless that art is found, many patents that do not teach novel or nonobvious inventions will be issued by the PTO and validated in court . The problem, of course, is connecting those who are looking for such art with those who might know, or motivating experts to come forward with relevant art.

Efforts to solve this collective action problem have been successfully solved in certain contexts, such as open source software development communities or peer-produced website references such as Wikipedia. The PTO’s Peer-to-Patent pilot is probably the best known attempt to apply crowdsourcing techniques to prior art searches and examinations. Private non-profits and businesses have engaged in similar¬† crowdsourcing and peer production efforts.

Article One Partners applies crowdsourcing as a business model- offering an award for the best prior art submitted by selef-selected participants. It was started by Chery Milone, who worked with BountyQuest in 1998, an earlier iteration. Article One offers rewards to self-identified experts who find what they deem to be the best prior art. To start building a community engaged in finding prior art, theyoffered bounties to invalidate patents that aren’t necessarily requested by third parties.¬† They have reported successful prior art searches for litigants as well.

The Electronic Frontier Foundation launched its Patent Busting Project in 2004, aimed at what it calls “bogus software patents”. It is currently targeting Volomedia’s podcasting patent, which I described earlier here.

EFF solicited nominations for the what it views were the 10 worst patents, and it began reexaminations of them.¬† The project has led to the rejection of several patents, such as the PTO’s rejection in reexam of Hoshiko’s internet subdomain patent, or in the case of Neomedia’s database access patent, a re-issuing that the EFFargues has narrowed the claims. Neomedia’s counsel has responded that the patent has been strengthened, and that additional limitations added to clarify the invention have no practical impact on the patent’s coverage.