Search the archive.

IN SUPPORT OF PATENT ELIGIBILITY JURISPRUDENCE STUDY” 10/8, //DH

B. Eligibility Uncertainty Incentivizes Litigation

The current unreliability of patent-eligibility law, documented thus far here, here and here, has also created undue burdens on litigants and the courts. In this final installment, we detail how the current unreliability burdens litigants and the courts and how it is a fundamental threat to U.S. competitiveness and national security.

Patent infringers now routinely raise Section 101 as a defense, often merely as a strategy to complicate and prolong the litigation, rather than as a good-faith defense with a likelihood of success. For example, one analysis found that, from 2012 to 2014 (when Alice was decided), Section 101 was raised in just two Rule 12(b)(6) motions across the country each year. In the year after Alice, that number rose to 36 motions, and by 2019, accused infringers were filing over 100 such motions each year.

The vagueness and randomness of the Alice/Mayo framework have also enabled patent infringers to exploit Section 101 as a litigation weapon exacting unnecessary burdens and costs on good-faith patent holders and the courts, further disincentivizing investment and innovation. Patent infringers now routinely raise Section 101 as a defense, often merely as a strategy to complicate and prolong litigation, rather than as a good-faith defense. One analysis found that from 2012 to 2014 (when Alice was decided), Section 101 was raised in just two Rule 12(b)(6) motions across the country each year. In the year after Alice, that number rose to 36 motions, and by 2019, accused infringers were filing nearly 100 such motions each year.8

A natural candidate for reducing industry consolidation, and thereby potentially increasing levels of AI innovation, is antitrust law. In general, antitrust laws are meant to "promote vigorous competition and protect consumers from anticompetitive mergers and business practices." 398To that end, both the Federal Trade Commission and the Department of Justice enforce federal antitrust laws in an effort to protect consumer welfare, including by way of limiting industry consolidation that is deemed to undermine socially beneficial innovation. 399Hence, antitrust is one possible means by [*1251] which to thwart AI industry consolidation if that consolidation's effect is to impede valuable AI innovation.

Yet relying on antitrust law to upend artificial stupidity seems dubious for several reasons. For starters, in general leading scholars often view U.S. antitrust law as rather anemic, 400 particularly when compared to more robust conceptions of antitrust law found in Europe and elsewhere. 401Hence, U.S. antitrust authorities have typically been reluctant to intervene unless a clear harm to consumer welfare is evident. 402

Of course, copyright is generally a weaker form of protection than patent protection. 260 For instance, independent creation is a viable defense to claims of infringement under copyright law, 261 whereas patent law provides no equivalent defense. 262 Hence, AI developers wishing to protect functional elements of their AI-related software via copyright would have to demonstrate, among other things, that the purported infringer actually copied the software from them (rather than coming up with it on their own). 263 But if the AI developer is also maintaining the copyrighted software as a trade secret, then in many cases proving that the alleged infringer had access to the software may be impossible. 264 With a patent in place, on the other hand, the AI developer would merely need to show that the patent covers the allegedly infringing product, regardless of whether the alleged infringer had access to the patent owner's AI innovations. 265

2. In response to a March 2021 letter from myself and Senator Cotton, the USPTO launched the “Deferred Subject Matter Eligibility Response Pilot Program,” which invited selected patent applicants to defer consideration of subject–matter eligibility issues until other patentability issues are resolved.

What are your thoughts on deferring consideration of subject–matter eligibility issues during patent examination?

Finally, with respect to his second proposal that patent protections be limited to “advances in technology,” this fails the purported empirical standard proposed by Mr. Jones in his first proposal. On the one hand, it is never clear if new innovations, such as the biotech and computer innovations secured as patent eligible by the Supreme Court in the early 1980s under the broad terms of § 101,12 represent actual “advances in technology.” What counts as an “advance in technology” is the question at the heart of the reform effort represented by PERA, in which real-world technologies in industrial and consumer machines, drugs, medical tests, and other innovations are now rejected by courts and the USPTO as ineligible from protection by the patent system as such. On the other hand, Mr. Jones appears in his proposal to limit “technology” to an excessively narrow conception of “technology,” such as machines, computer hardware, industrial products, etc. But the U.S. patent system has never limited itself in such ways. From the very beginning, Congress adopted a very expansive conception of the “useful Arts” that is authorized in the Constitution to promote with a patent system, and thus the U.S. protected processes, including even business methods from the first decades of the country through the nineteenth and twentieth centuries, as I explained in my written testimony submitted for the hearing on January 23,

2024. Thus, Mr. Jones’s second proposal is a radical change to the patent system that would codify an extremely narrow conception of patent eligibility—the same narrow conception of patent eligibility that has plagued innovators since the Supreme Court created the Mayo-Alice inquiry as a judicial gloss on § 101.

The Supreme Court destroyed patent eligibility for entire swaths of tech innovation critical to our national security and economic growth in a 2014 decision called Alice Corp. v. CLS Bank International (Alice).

Alice created an unintelligible and undefined exception to patent eligibility called the “abstract idea.” Because nobody knows what qualifies as an abstract idea, Alice gutted the patent protection for most inventions in artificial intelligence, cybersecurity, quantum computing, blockchain, 5G/6G, enterprise systems, and much more. These technologies are critical to our national security and economic growth and are the bulk of the 37 of 44 technologies in which China now leads the world.

Government subsidies are not close substitutes for market incentives for innovation. The National Science Board estimates that federal agencies provided over $132 billion for R&D.87 Wright et al. find that even though corporations provide just over 5% of U.S. university research funding, "corporate-sponsored research is surprisingly valuable for further innovation."88 As they observe: "Data collected over 20 years at nine campuses and three national laboratories administered by the University of California show that corporate-sponsored inventions are licensed and cited more often than federally sponsored ones."89

Government agencies would replace decentralized markets with centralized allocation. Government agencies are not well equipped to run market places. Health insurance exchanges are an example of government market place design.90 Government allocation of goods and services often entails high administrative costs, significant transaction costs, and allocative inefficiencies. 91 Historical evidence suggests that state and federal regulation led to considerable allocative and dynamic inefficiencies in markets for electric power, natural gas, telecommunications, railroads, trucking, and airlines. 92 Increased regulation in health care is expected to generate similar allocative and dynamic inefficiencies. 93

Public commercialization and innovation would generate insurmountable problems of economic planning. Exercising command and control over economic activity has led to significant problems in socialist economies. The collapse of socialist economies such as the Soviet Union provides sufficient evidence that even the operation of traditional technologies by governments is not feasible. The diffusion of new technologies and management of innovation generally are far beyond the scope of government administration.

Governments are not very good at commercializing and diffusing inventions. The failure of government diffusion of technology is illustrated by the major benefits that resulted from the Bayh-Dole Act. The effects of the Bayh- Dole Act demonstrate how markets diffuse innovations and governments do not. The U.S. government provides research grants to universities, non-profit organizations, and small businesses. However, the incentives for invention provided by such grants are accompanied by markets for invention, which provide additional incentives for commercialization and innovation. This was recognized by the Bayh-Dole Act, which permits organizations that receive federal government funding to own inventions rather than transferring ownership to the government.82

The existing layers of bureaucracy have obvious costs in speed. They also have subtle costs in creativity. The NIH’s pre-grant peer-review process requires that many reviewers approve an application. This consensus-oriented style can be a check against novelty—what if one scientist sees extraordinary promise in a wacky idea but the rest of the board sees only its wackiness? The sheer amount of work required to get a grant also penalizes radical creativity. Many scientists, anticipating the turgidity and conservatism of the NIH’s approval system, apply for projects that they anticipate will appeal to the board rather than pour their energies into a truly new idea that, after a 500-day waiting period, might get rejected. This is happening in an academic industry where securing NIH funding can be make-or-break: Since the 1960s, doctoral programs have gotten longer and longer, while the share of Ph.D. holders getting tenure has declined by 40 percent.

Second, there is a specialization paradox. Despite considerable domain specialization in the sciences, individual scientists cannot focus enough on doing hard research in their chosen field.

Since 1970, the number of years the average Ph.D. student in the biosciences spends in graduate school has grown from a little more than five years to almost eight years. Producing experts is taking longer, and those experts are getting less productive. In the famous paper “Are Ideas Getting Harder to Find?,” the Stanford University economist Nicholas Bloom and his colleagues found that research productivity has declined sharply across the board since the 1970s. Research from the University of Chicago scholar James Evans has found that as the number of researchers has grown, progress has slowed down in some fields, perhaps because scientists are so overwhelmed by the glut of information they have to process that they’re clustering around the same safe subjects and citing the same few papers.

Most scientific funding in the United States flows from federal agencies such as the National Institutes of Health and the National Science Foundation. This funding is famously luxurious; the NIH and NSF allocate about $50 billion a year. It is also infamously laborious and slow. Scientists spend up to 40 percent of their time working on research grants rather than on research. And funding agencies sometimes take seven months (or longer) to review an application, respond, or request a resubmission. Anything we can do to accelerate the grant-application process could hugely increase the productivity of science.

The major concern with this solution, however, is the uncertainty of sustained funding into such research, given political conditions and changes in the control of government. While federal funding for both R&D and basic research has been increasing in recent years,272 the bulk of U.S. R&D spending now comes from industry.273 Such spending is, of course, subject to incentives. This uncertainty as to sustained funding may limit the investment universities and companies are willing to make in researching genetic associations.

The NotPetya case is just one example. Nevertheless, the overall absence of any meaningful instances of cyber escalation raises questions about the extent to which the cyber domain is truly escalatory, let alone more dangerous than other domains of warfare or strategic competition. In other words, there is a significant mismatch between the cyber escalation views held by many prominent experts and the available evidence. Three factors could account for this gap.

First, differences between theoretical expectations and empirical reality may stem from definitional issues—specifically, how those who assert that the cyber domain is inherently escalatory define what "escalation" means. As we discuss in detail later in this chapter, we conceptualize escalation as something that changes the nature of a strategic interaction between parties in a way that makes war more likely or, if conflict is already taking place, more severe. According to this definition, neither the sheer volume of cyber incidents in the international system nor evidence of any type of response to a cyber incident necessarily constitutes escalation.2, In contrast, some may point to the observation of a high volume of cyber incidents or any form of tit-for-tat dynamic (irrespective of the relative level of cost and risk they may generate) as forms of escalation. However, the mere presence of behavior or activity per se is not sufficient evidence of escalation.

The market foundation role of patents implies that antitrust policy should support rather than weaken IP protections. When there is a market for inventions, competitive pressures, both among inventors and among producers who apply inventions, increase incentives to invent and to innovate.26 On the supply side of the market for inventions, competitive pressures increase incentives to invent and decrease royalties for inventions. On the demand side of the market for inventions, competition drives adopters’ incentives to purchase or license discoveries, introduce economic innovations based on those discoveries, and invest in complementary activities. Weakening of patents reduces incentives to participate in the market for inventions, leading inventors to protect their IP through other means such as trade secrets, R&D alliances, and vertical integration. In the absence of a market for inventions, competitive pressures tend to reduce incentives to invent and to innovate.

b) Socially wasteful R & D and duplicative innovation. Prize advocates cite another problem with using intellectual property to determine the incentives for innovation: the potential for socially wasteful R & D and duplicative innovation. One social cost of offering greater incentives than are necessary to motivate development is a phenomenon known as “patent racing.” Notwithstanding the name, the phenomenon exists in a prize system as well as a patent system.147 It occurs when an excessive incentive leads firms to engage in socially wasteful behavior to capture those excess profits. Prizes and patents are both a reward for successful innovation. When a larger incentive is given than is necessary to motivate an invention, the excess reward creates unnecessary competition between firms to expend resources in ways that, although justifiable for the winning firm, reduce the benefit of the innovation to society as a whole.

In a prize system, an excessive incentive can stifle future innovation because subsequent innovators will likely be seeking compensation from the same system. It is true that the stifling effects of licensing are avoided in a prize system by placing the innovation immediately into the public domain.158 However, since government resources are finite, and the government must divide profits among sequential innovators, an excessive incentive for an early innovation will subtract too much profit from the pool for later innovators. To the extent that subsequent innovators anticipate smaller prizes for this reason, an excessive prize could have the same effect on cumulative innovation as an excessive patent (assuming that subsequent innovators will be depending in part on the prize system for their profits).

In this article, I explain why arguments for replacing market prices with public prizes are incorrect. First, I show that the standard deadweight welfare loss argument is a fallacy because it assumes government prizes create no economic distortions - a classic case of a "free lunch." Yet, the government must collect taxes to pay for the prizes and such taxes create substantial deadweight welfare losses." The bureaucratic and compliance costs of the taxation system are substantial as well. The standard argument for public prizes assumes that the government would incur no costs of administering the prize system or managing subsequent technology diffusion and commercialization - another free lunch. Yet, the administrative costs of operating a government prize system and commercializing technologies are significant.

The standard argument for public prizes further assumes that market prices create monopoly deadweight welfare losses. This assumption is inaccurate because extensive competition in the markets for inventions and products reduces prices and mitigates deadweight welfare loss. Also, market competition promotes efficiency because firms have incentives to reduce the transaction costs of commercialization and innovation. I conclude that the deadweight welfare losses resulting from a government prize system are likely to substantially exceed any such losses from competitive markets - replacing prices with prizes would lower social welfare.