Patent Reform

Crowdsourcing of Prior Art – A New Road to Patent Quality?

201501 SocialistsThat Which Is Known –

“Prior Art” is one of the most important concepts in any statute-based system of granting a patent to an applicant.  Stated succinctly, “prior art” is “that which is known.”  In particular, as it relates to an invention and an application for a patent on that invention, prior art is information that demonstrates that the invention, or a part of the invention, is already known to the public.


As readers of this column may recall, when a patent application is filed in the USPTO, it undergoes what is known as prosecution.1  An Examiner studies the application to determine if the claims meet the requirements for patentability as set forth in certain statutes.  Two critical statues are 35 U.S.C.§102, which requires that the invention must be novel, i.e. new, and not previously known; and 35 U.S.C.§103, which requires that the invention must be unobvious, i.e., a person “of ordinary skill in the art” would not easily be able to combine knowledge from multiple sources to create the invention.

In examining an application against the novelty and obviousness statutes, the Examiner does a “prior art search” in the relevant classifications of patents and published patent applications in the U.S. and worldwide, as well as in relevant non-patent literature (academic and trade journals, Internet sources, etc.)  If the Examiner finds references that render the invention as defined in the claims of the application as lacking novelty or being obvious, they are rejected under “102” or “103” in an Office Action.  The Applicant then has the burden of responding with a rebuttal of the Examiner’s findings, and/or amending the claims so that they are found to be novel and unobvious as compared to the prior art.  This often takes more than one iteration of the Office Action/Response cycle, but at the conclusion, if the claims are not found to be novel and unobvious, the grant of a patent is denied.

Patent “Quality”

Once issued, a U.S. patent is presumed to be valid and enforceable, thus providing the patentee with the right to exclude any third party from making, using, or selling the claimed invention during the term of the patent.  Accordingly, a key aspect of the “quality” of an issued patent is its validity, which may be thought of as its ability to withstand any legal challenges asserted in court, or in further post-grant proceedings in the USPTO.

Thus it is easy to see that the “quality” of an issued patent is highly dependent upon the quality of the prior art search performed by the Examiner.  On any given application, an Examiner is given a limited time to perform examination, including searching for prior art that may be the basis for properly rejecting the claims as unpatentable.  Under such circumstances, and without enough time to do a more thorough search, an Examiner might miss a relevant piece of prior art that would have been the basis for a rejection, and instead allow the claims.  As a consequence of the lack of discovery of this critical piece of prior art, a “bad patent” is then issued by the USPTO.

Public Perception

There has been a perception in recent years, driven to some extent by parties with particular interests at stake, that the USPTO has been issuing too many “bad patents.”  Whether the perception is justified or not, Congress was duly persuaded when drafting the America Invents Act, which was enacted in September of 2012.  Under the AIA, new statute 35 U.S.C. §122(e) provides an opportunity for a third party to submit prior art to the USPTO for entry into the record and consideration by the Examiner during prosecution of a patent application.2,3

“Count us as skeptical that crowdsourcing of prior art will be effective in improving patent quality.”

The provisions of 122(e) are quite structured and specific, and are intended to conform to the ex parte proceedings of patent prosecution:

  • Submissions are limited to “any patent, published patent application, or other printed publication of potential relevance.” Thus collections of multiple documents are not permitted in a single submission.
  • There are very specific timing requirements, i.e. there is a specific “window” in prosecution during which the submission must be made.
  • The submission must be accompanied by “a concise description of the asserted relevance of each submitted document,” and a fee must be paid to the USPTO for entry and consideration of the submission.

If Some Prior Art Is Good…

More is better, apparently. The third party prior art submission provisions of the AIA notwithstanding, the perception that Examiners are still not accessing and considering the most relevant prior art remains.  On February 20, 2014, the White House announced4 a new executive action that directs the USPTO to develop and implement a program to “crowdsource” prior art.

Since that time, the USPTO has held two “roundtable” events for open discussion of the initiative, on April 10th and December 2nd.  The Office has also had two periods for the public to submit written comments on the initiative, the first running from March 19th to May 9th, and the second starting November 12th and recently extended to January 13, 2015.

In reviewing the online comments page for the March 19th comment request, we see that a total of 16 comments have been submitted. We were unable to locate the comments for the current period, and it appears that since the period is still open as of this writing, they are not yet available online. In any event, the total of 16 comments in the first comment period does not suggest a groundswell of support for crowdsourcing of prior art.

Not surprisingly, in the roundtables, much of the subject matter in various presenters’ slides, as well as additional crowdsourcing-related material that we found online, was directed to concerns with software patents.  These have long been controversial, and will continue to be argued in Congress, the courts, and the Patent Office.

In another development, the Office has recently posted an online solicitation of assistance in prior art crowdsourcing from the private sector, “… seeking information from qualified firms that are capable of electronically receiving and potentially hosting millions of structured/indexed and unstructured documents.  The firm shall index and/or add searchable metadata (e.g. keywords, author, date of public availability, etc.) to the received materials. The firm shall have the capability to provide access to a publically [sic] searchable database.”

Thanks For Asking, But…

The USPTO will have a difficult challenge implementing the crowdsourcing of prior art while still complying with 35 U.S.C. 122(c), which states. “…no protest or other form of pre-issuance opposition to the grant of a patent on an application may be initiated after publication of the application without the express written consent of the applicant.”  Any new crowdsourcing rules that the Office implements must comply with the ex parte structure of patent prosecution that is required by the statutes.  However the USPTO crowdsourcing initiative evolves, we are skeptical that it will be effective in improving patent quality for several reasons.

Examination of patent applications occurs on a case-by case basis, and patentability is determined by what is recited in the claims of an application.  Regardless of whether new prior art databases are made available from the private sector, they are still just that: databases that must be searched by an Examiner so that relevant information can be extracted.  Currently, this still requires human (the Examiner’s) intelligence to extract the information that is relevant to the claims of the particular application.  In engineering parlance, the Examiner must find the signal in a lot of noise.

Also, with regard to examination – if you were examining an application, which would you pay more attention to – a pile of crowdsourced information from a database, or a single document filed by one third party that includes “a concise description of the asserted relevance” of the document to the claims of the application?  Exactly.  The former is a shotgun blast of #8 at 50 yards, and the latter is a rifle shot with a scope that can be aimed right at the claims.  No serious opposers (e.g.,any rival firms) would submit prior art via the crowdsourcing route when, for the payment of a $180 fee, they can include that critical statement of relevance, which may well persuade the Examiner.

There is also the distinct possibility that if your rival has a key piece of prior art, it may hold back and not submit it during prosecution.  Submitting that prior art gives the Applicant a chance to address it, and potentially overcome it, resulting in a stronger patent.  Instead, by holding back until after the patent issues, the rival firm can then challenge the issued patent with the new prior art in new AIA post grant proceedings5 in the USPTO (which are turning out to be quite favorable to third parties), or through litigation in court.  Much as you don’t always lead with trump in Euchre, it may make sense to hold your prior art cards until after patent issuance.  Considering both of the above scenarios, the crowdsourced prior art option is not a useful method for a company that wants to influence its rivals’ patent applications.

So given that entities in competition with each other are not likely to crowdsource prior art, that leaves concerned citizens in general as potential crowdsourcers.  In FY2012, almost 543,000 utility patent applications were filed.  Of this total, by now, about 500,000 of these applications have been published, and are thus of public knowledge.  So if crowdsourcing of prior art is instituted, and present statistics hold, about a half million new patent applications will be available each year for crowdsourcing submissions.  That’s going to take a lot of concerned citizens to access and study those applications, and then search, find, and submit that relevant prior art –  especially if the USPTO maintains the requirement that a statement of the relevance of the prior art is necessary along with the submission.

Since most of the citizenry have better things to do, such as job and family obligations, there is potentially a small subset, which we refer to as “the angry mob.”  While it is possible that some activists could spam a company’s patent application with a pile of crowdsourced prior art, we expect that it would have little effect, for the reasons stated above.

Lastly, and ironically, we think that there is the opportunity for abuse of crowdsourcing by an unscrupulous Applicant.  Such an Applicant could submit one important prior art reference anonymously via the crowdsourcing route, camouflaged by another fifty or so additional crowdsourced references of no relevance.  In that context, the significance of the key reference could be overlooked by the Examiner, but it would still be checked off as having been considered in prosecution of the application.  That would make it much more difficult to assert the reference later in any challenge to the issued patent.

Next Move – the USPTO

It will be interesting to see how the prior art crowdsourcing initiative progresses, and if the USPTO issues new rules to implement it.  We will provide an update in a future column if that happens.  In the meantime, hopefully, this concept won’t be applied to health care.  We’d rather not have our doctor crowdsource the reading of our chest X-ray or MRI.

  1. See The Limited Monopoly® March 2009.
  2. See The Limited Monopoly® October 2012.
  3. See also 37 C.F.R.§1.290
  5. See The Limited Monopoly® September 2012.

PHOTO CREDIT:  “Socialists in Union Square, N.Y.C.,” 1 May 1912.  Courtesy of the Library of Congress, Prints and Photographs Division, Washington, D.C.

Authors John M. Hammond P.E. (Patent Innovations, LLC and Robert D. Gunderman P.E. (Patent Technologies, LLC are both registered patent agents and licensed professional engineers.  They offer several courses that qualify for PDH credits.  More information can be found at Copyright 2015 John Hammond and Robert Gunderman, Jr.

Note:  This short article is intended only to provide cursory background information, and is not intended to be legal advice.  No client relationship with the authors is in any way established by this article.PDF printer wo icon

          For a PDF download of this article or to print a copy , click here →