“Patentability statutes do not, and cannot, address actual products—it is the claim that must be enabled by the patent specification; and the invention as claimed must meet the utility requirement.”
Last week, the House Judiciary Subcommittee on Courts, Intellectual Property, and the Internet, held an oversight hearing on the U.S. Patent and Trademark Office (USPTO) with Director Andrei Iancu as the sole witness. A particular inquiry from Rep. Zoe Lofgren (D-CA) regarding the USPTO’s allegedly lax examination quality under 35 U.S.C. § 112 caught my attention. She remarked [at 1:33:30]: “Theranos, the blood testing company whose founder is being investigated for fraud, was granted nearly 100 patents based on an invention that didn’t work; and it concerns me that a patent application for an invention that doesn’t work gets approved.” She generally questioned examiners’ attention to Section 112 requirements. Rep. Lofgren’s statement was no doubt primed by information from the Electronic Frontier Foundation (EFF) in the Ars Technica blog post titled “Theranos: How a broken patent system sustained its decade-long deception.” In this article, the author, who was introduced as holding the “Mark Cuban Chair to Eliminate Stupid Patents” at EFF, declares with no evidence or proof, that the “USPTO generally does a terrible job of ensuring that applications meet the utility and enablement standards.” The article cited no study, identified no patent, nor any claim to any “invention that didn’t work.” This outrageous, baseless allegation is outright reckless and irresponsible.
- A Decade-Old Assault
We have seen this movie before. Anti-patent narratives are disseminated through media and congressional committees in a campaign to impugn the examination quality of the USPTO with anecdotal innuendos, mythical premises, distorted logic, and flawed evidence. This latest EFF allegation continues the decade-old, systematic, unjustified assault on the reputation of the USPTO’s examination process by using bad science in search of “bad patents.” This assault is continued today to ridicule USPTO allowances, to undermine patents’ presumption of validity, and to pressure the USPTO to return to the mid-2000s culture of “reject, reject, and reject.” The article on Theranos alleging a “broken patent system” proved nothing of the kind. Contrary to the premise of the article, the statutory utility and enablement standards are neither designed to avoid issuing “stupid patents” (which harm no one), nor can they ensure that actual products work. And these statutes certainly do not empower or authorize the USPTO to substitute its judgment for the Security and Exchange Commission’s (SEC’s) in policing or eliminating fraud on investors. Furthermore, the notion that “100 patents” were issued “based on an invention that didn’t work” is fictional because many of those issued Theranos patents cover subject matter that is peripheral, secondary, and unrelated to the alleged “invention that didn’t work.”
To be sure, as Director Iancu himself acknowledged, examination procedures can and should be improved. But that does not mean that EFF’s allegations have merit. That examiners seriously attend to Section 112 requirements is evident from the relatively high percentage of such rejections in USPTO office actions, particularly in initial rejections (see accompanying figure). More specifically, other empirical evidence suggests that the USPTO’s allowance error rate on claim enablement is the lowest among its other major allowance error categories. Data on patent litigation published by the Institute for Intellectual Property & Information Law (IPIL) at the University of Houston Law Center, shows that patent challenges for lack of enablement were the least successful in federal courts among challenges under major statutory patentability criteria. For example, patentees were 62% more likely to prevail on Section 112 enablement than they were on anticipation over published prior art under Section 102(b). In any event, the basic predicates underlying the EFF allegation and Rep. Lofgren’s remarks are flawed, as explained below.
- Mischaracterization of the invention “working” requirements
First, patentability statutes do not, and cannot, address actual products—it is the claim that must be enabled by the patent specification; and the invention as claimed must meet the utility requirement. Thus, the examiner’s focus must be on the claim, ensuring that each limitation and interconnections of the claim is enabled by the specification, as the statute in Section 112 requires. The USPTO does not examine claims made in press releases. Enablement of a patent claim cannot guarantee that any product feature that the applicant promises in a press release to deliver to market would actually work. Second, an applicant need not have actually reduced the invention to practice prior to filing. Gould v. Quigg, 822 F.2d 1074, 1078 (Fed. Cir. 1987)(“The mere fact that something has not previously been done clearly is not, in itself, a sufficient basis for rejecting all applications purporting to disclose how to do it.”) Indeed, we will have a lot less disclosure and investments in new inventions if one were to erect “working” requirement barriers for applicants. This is particularly true now that the first-to-file provisions in the America Invents Act (AIA) shift the risks onto those who take more time to experiment and vet their invention prior to filing; and completing this “Catch-22” jeopardy, the countervailing risk of not verifying the “working” of the invention is also borne by the applicant.
Constructive reduction to practice is recognized no later than the date the patent specification compliant with Section 112 is filed. Hyatt v. Boone, 146 F.3d 1348, 1352 (Fed. Cir. 1998). However, the scope of the claims is limited to that which is enabled in the specification, AK Steel Corp. v. Sollac, 344 F.3d 1234, 1244 (Fed. Cir. 2003); and thus there is little harm to the public if an inventor gains exclusive right to an invention that does not work, or is “stupid.” If anything, the disclosure is beneficial as it can teach others how to overcome its faults and invent an improvement that actually does work.
Similarly, the public suffers little harm if the examiner, despite correctly verifying that all claim limitations are properly enabled by the specification, cannot mentally construct an actual product to recognize that the invention as claimed may not work. This recognition is required in order to reject for lack of utility. While examiners do reject claims if specific utility is not disclosed in the specification, the IPIL data shows that failure to meet the utility requirement was never challenged in patent infringement suits. This is likely because any alleged infringer who would argue that the claims cover an invention lacking utility supplies his own proof to the contrary by his infringement—the infringer is usefully “utilizing” the claimed invention. In other words, there could be no risk of actual infringement of a claimed invention that does not work. This counterfactual is a bit like Woody Allen’s complaint that the food in summer camp was “terrible” and that the “portions they served were too small.”
Until the Patent Act of 1870, applicants were generally required to submit a “model” of the patented invention to the USPTO, to help examiners and the public (who upon request could gain access to the model room) learn how the claimed invention operates. But Congress recognized that the costs imposed on applicants for making models and paying fees for the Office to house and preserve them was often prohibitive for ordinary citizens and had a chilling effect on incentives to disclose new inventions. Congress also recognized that, with the advent of the electrical and chemical arts, many inventions do not admit to the use of “models,” and thus repealed the model requirement and relied on the language in what is now Sections 112 and 113, the latter requiring that “[t]he applicant shall furnish a drawing where necessary for the understanding of the subject matter sought to be patented.” The USPTO regulations in place since before the 1952 Patent Act go even further than the statute, requiring in 37 C.F.R. § 1.83(a) that the “drawing … must show every feature of the invention specified in the claims.” Examiners often require compliance with this rule, as those who frequently receive “objections to the drawings” in office actions can attest.
The test for determining whether the specification meets the enablement requirement was judicially set a century ago by the Supreme Court in Minerals Separation Ltd. v. Hyde, 242 U.S. 261, 270 (1916), which created the question of whether the experimentation needed to practice the invention is undue or unreasonable. This standard is still the one applied by the Federal Circuit today. Trustees of Boston University v. Everlight Electronics Co., Ltd., 896 F.3d 1357, 1362 (Fed. Cir. 2018) (“[T]o be enabling, the specification of a patent must teach those skilled in the art how to make and use the full scope of the claimed invention without undue experimentation.”) (My emphasis). The test of enablement is not whether any experimentation is necessary, but rather “the issue is whether the amount of experimentation required is ‘undue.’” In re Vaeck, 947 F.2d 488, 495 (Fed. Cir. 1991). Technically, the information in the claim and the specification should permit the examiner to determine which claimed elements are adequately enabled and which require a degree of experimentation. However, the examiner may not be able to show that the experimentation would be unsuccessful—a showing that must be made as part of the examiner’s initial burden to establish a reasonable basis to question the enablement provided for the claimed invention. In re Wright, 999 F.2d 1557, 1562 (Fed. Cir. 1993). Many meritorious and useful inventions would lose patent protection if examiners would be permitted to speculate inoperability and reject a claim with no proof by shifting the burden onto the applicant, contrary to the requirement of MPEP § 2164.04.
- The self-regulating features of U.S. patent law improve compliance
Ultimately, the legal system and the nature of patent enforcement provided self-regulating forces for applicants’ compliance with enablement requirements that supplemented the USPTO examination process, by independently creating strong incentives and public protections as follows:
- As discussed above, a patent holder is not entitled to a claim interpretation that is not enabled by the specification. Applicants have incentives to provide such claim support in the first place. Unfortunately, the AIA’s first-to-file provisions necessitate races to file first, sometimes with half-baked inventions, undermining the quality of patent disclosures.
- A patent need not teach, and preferably omits, what is well known in the art. In re Buchner, 929 F.2d 660, 661 (Fed. Cir. 1991). However, applicants often err on the side of further disclosure because arguing otherwise—that an element in the claim is enabled as being well-known in the art—essentially constitutes an admission that an examiner’s rejection of that claim for anticipation and/or obviousness may be justified.
- The “utility” prong is essentially self-executing because an applicant would seldom spend resources to protect that which has no utility.
- The “Best mode” requirement coupled with the “grace period” was also self-enforcing: “The specification. . . shall set forth the best mode contemplated by the inventor of carrying out his invention.” 35 U.S.C. § 112(¶2) (pre-AIA). Applicants were highly incentivized to disclose the best way to implement the invention in the first place because a discovery and allegation that applicant concealed from the USPTO the best mode known to the inventor at the time of filing is an affirmative defense of inequitable conduct, which could render the patent unenforceable. This is a significant and nontrivial matter: owing to the doctrine of equivalents, applicants may obtain strong claims based on disclosure of suboptimal embodiments that conceal the inventors’ competitive advantages. The grace period permitting the inventor more time to perfect the invention and disclose the best mode on the one hand, and the requirement for such best mode disclosure on the other hand, can be seen as the American second patent bargain—one-year grace period for optimizing, reducing to practice, and perfecting inventions in exchange for the inventor’s disclosure of the results of this optimization—the best mode for practicing the invention. Unfortunately, the AIA did away with this second patent bargain by eliminating the best mode defense, 35 USC § 282(b)(3)(A), and the Supreme Court decision in Helsinn v. Teva confirmed that it also eliminated the “on-sale”/“public use” grace period, undermining the quality of patent disclosures.
The USPTO must examine claims for patentability under a limited time per application (about 26 hours on average); one cannot expect the Office to be the custodian and public trustee of “peer review” services on specific designs claimed in patent applications, rendering determinations on the technical feasibility of implementing the claimed inventions. While USPTO examination quality should be generally improved, there is no evidence of a trend suggesting that the USPTO “does a terrible job of ensuring that applications meet the utility and enablement standards.” In fact, there is empirical evidence to the contrary. Pinning unjustified blame on the USPTO is a familiar tactic of the “efficient infringement” cabal. If any blame for a declining quality of enabling disclosures should be placed, it is squarely on the enactment of the AIA, which repealed important self-regulating disclosure safeguards that were established over more than two centuries.