Science21 Highlights: Peer to Patent and Government 2.0

(Over the course of this week, I’m going to post a handful of things about talks that struck me as particularly interesting at last week’s conference. The order will be chosen based on how much time I think I will have to write them up, given SteelyKid’s demands for attention…)

On Thursday, the Science in the 21st Century conference featured a lawyer. Well, a law professor, anyway– Beth Noveck of the NYU Law School gave a talk on Science in Government 2.0 (FriendFeed microblogging). This wasn’t as odd as you might think at a Web 2.0 kind of meeting, because she talked about her involvement with a project called Peer to Patent, which uses the Web to help provide expert assistance to patent examiners.

As Noveck noted, one of the largest problems facing the patent system in the US is that it takes forever to check for “prior art,” previous examples of the same idea that might indicate that a proposed invention is not actually original enough to deserve patent protection. This information can be hidden in unlikely places– she gave the example of an idea that turned out to have been mentioned in the owner’s manual for some Norwegian product, that was found in a single library in Scandanavia somewhere. As a result, patent examiners are faced with a backlog of something like a million applications, and occasionally just punt, granting dubious patents and letting the courts sort it out at great expense to everyone involved.

The idea of Peer to Patent is to get the public involved, as an aid to the overworked examiners.

In many cases, it turns out that while the information they need is difficult to track down through library searches and the like, specific individuals may be able to come up with it very quickly, if they know that it’s needed. Peer to Patent is a trial project that takes selected patent applications and, with the permission of the applicant, makes them available for public review. People who come to the site are asked to provide any knowledge they have of prior art relating to the applications, including references and sources, and that information is forwarded on to the Patent Office.

The idea here is simple but powerful– individual people who have worked in relevant fields, or who just happen to have an interest in some area, may be able to readily recall information that a patent examiner with limited resources and experience would be unlikely to unearth. Using the web, we can reach out to those people, and get them involved in the review directly, and make the whole process more efficient.

(The concept is somewhat similar to InnoCentive, in which people are invited to help solve problems for major companies and win substantial sums of money. The big difference is that there’s no money involved in Peer to Patent…)

So far, this is a very small project– they’ve been permitted to do this on a trial basis, with something like 400 patent applications, out of about 400,000 per year. They’ve completed review of 80 applications, and the Patent Office has finished with 40 of those. The numbers are still far too small to give much of a breakdown of how this is working, but Noveck was very positive about it. The review of those 40 applications has involved about 400 comments on the site, and turned up 176 pieces of prior art, which suggests that this sort of distributed data mining might be a really useful tool.

Later that evening, there was a discussion (billed as a “Design Exercise”), in which we kicked around ideas for ways to apply the sort of Web 2.0 tools the meeting was focussed on to problems in policymaking (it wasn’t live-blogged in detail, but Greg Wilson typed up what were basically minutes in a text file). This went around and around about ways to distribute decision making or attempt to influence opinion, but I wonder if that wasn’t misguided. It seems to me that projects like Peer to Patent and the Galaxy Zoo suggest that the real power of “crowdsourcing” and the like is in sifting through vast amounts of information.

During my talk, Jacques Distler noted that the appropriation for ITER funding was cut from the omnibus budget bill through the action of a single legislative aide, and nobody else noticed it, given the extreme length and density of the bill itself. This seems to me like the sort of thing where the web could be put to good use, by farming giant pieces of legislation out to be read by huge numbers of people, each taking on a manageable chunk. You could get a good idea of the real contents of things like budget bills much more quickly this way, and possibly quickly enough to allow members of Congress to make informed decisions about their votes.

(This would be similar to the Dover Panda Trial, in which (as i understand it) the mountains of paper turned over in discovery were sifted through by an army of volunteers to find evidence of the creationist origins of “Intelligent Design” texts.)

I mentioned this during the discussion, but it didn’t really go anywhere. So I’ll throw it out there again, just in case somebody else comes up with it later, and tries to patent the idea…