While considering a patent application, a patent examiner at the United States Patent and Trademark Office (USPTO) is expected to search prior art to determine whether an application is novel and non-obvious.  Unfortunately, prior art can come in a variety of forms.  For example, prior art can be patents, scientific publications or presentations, trade journals, Ph.D. theses, or even evidence of use of the technology (just to name a few sources).  Prior art can come from anywhere in the world and can be in any language.  With examiners spending only about 22.5 hours on each patent, a lot of prior art has been missed.

On September 16, a new provision of the America Invents Act (AIA), which will have a major impact on patent examination, was implemented.  Title 35 U.S.C. 122(e) allows third parties to submit prior art to the USPTO during the application process.  “The submission by third parties of prior art – the library of published patents, applications, or other publications in a specific technology area – allows the USPTO to tap directly into the U.S. innovation community,” according to a USPTO press release.  “Submissions provide a fuller, more exhaustive scope of materials for examiners to review in determining the novelty of a given application.  This new mechanism will help ensure that truly novel, useful, and non-obvious innovations obtain the intellectual property protection they deserve.”

But simply all0wing submissions is not enough for the USPTO; they’re looking to encourage widespread public involvement in prior art searching.  As a result, they have partnered with the people at Google and Stack Exchange to crowdsource prior art searching.  According to the USPTO, at the new website, Ask Patents, “subject-matter experts volunteer to suggest prior art for given applications, as well as offer their input on the proposed value of those suggestions from others.”  Meanwhile, at Google Patent Search, links will be provided from relevant patent applications to discussion pages on Ask Patents.  David Kappos, the director of the USPTO, said “[w]e encourage our nation’s innovators to follow Stack Exchange’s example and assist us as we improve the examination process and resulting patent quality that will drive our economy and create jobs and exports.”

Do you think crowdsearching is the answer?  What effect do you think this will have on the patent application process?

–Mike Ritter

Image Source


Tagged with:

4 Responses to Crowdsearching: the Crowdsourcing of Prior Art Searching

  1. I work for a crowdsourced prior art search firm (Patexia, http://www.patexia.com, disclaimer?) so I thought I’d add something.

    Note (for those who aren’t familiar): our model is basically that we’ll offer a prize to whoever can submit the best invalidating prior art for a patent, usually one involved in litigation.

    It’s been over 6 months since the launch of stack exchange ask patents and, if you look at the activity on the site, it just doesn’t seem to have took off/gotten great traction.

    I think that the major reason is probably lack of incentives.

    Interestingly though, I don’t think it’s just the lack of monetary incentives (though that is likely part of it as well). I think there’s probably a lack of satisfaction from the work.

    1) The patents they target are kind of all over the place (vs. one’s that have come up in litigation, say), so it’s harder to know if you’re working on patents that are going to make a difference.
    2) There’s not a great feedback loop to find out the results of your labor, so you don’t find out if your art does a good job of invalidating/blocking the patent.

    So obviously that’s a plug for the kind of approach we (and similar crowdsourced search firms, e.g. Article One Partners, and BluePatent) are using, but we seem to be getting more traction for the kind of endeavor.

    Anyway, just wanted to contribute my thoughts. Would be interested to hear yours.

  2. Will P. says:

    It seems like you need to give the “crowd” some incentive to search to really reap the benefits of having a system like this work. I just don’t, without knowing more, see this yielding great results except in the case of very high profile patents as Mike’s comment above noted. Although it may be worth it even if it only helps in those cases too.

  3. Niels Melius says:

    To echo Mike’s comment, this is an interesting post. However, I do think the comparison to wikipedia may be inapposite given the type of credibility we’re seeking from issued patents versus your average wikipedia article. Minor flaws in wikipedia articles can be corrected by actual research, whereas flawed patents have important economic effects. I worry about the overwhelmed PTO examiner, who only has 20 hours to spend on a patent application, and now has to add to his to-do list a review of all citizen-submitted potential prior art. It seems cumbersome, but if there is a way to make it more targeted, I think it could be very powerful.

  4. Mike Silliman says:

    Really interesting post Mike. While I applaud the PTO’s attempts to innovate and find creative solutions to the inadequacy of the current examination process, I am not convinced that crowdsourcing will provide much relief. First, other than a small amount of high-profile patents (i.e., anything Apple files), I believe there will be little participation or interest in prior art searches. Furthermore, I am concerned that the complexity of the standards surrounding obviousness and novelty in patent law will lead to many prior art references being submitted that do not raise patentability issues.

    However, Wikipedia is a strong counterexample. Maybe subject-matter experts will dedicate themselves to assisting the patent system with the same fervor they have applied to creating a free encyclopedia. We can only hope.