Imagine That IP Law Blog

Uncertainty surrounding patent eligibility jurisprudence has been a thorn in the side of many patent applicants, patent holders, scholars, and judges for years, leading voices in the patent sector to issue repeated, though unavailing, calls for legislative action. Recent activity by lawmakers, however, suggests that rising innovation in the artificial intelligence (AI) sector may be viewed as a catalyst for change.

Central to the issue of patent eligibility is the framework set forth in Mayo Collaborative Services. v. Prometheus Labs, 132 S.Ct. 1289 (2012) and Alice v. CLS Bank International, 134 S. Ct. 2347 (2014). The Patent Act permits patents for “any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof . . . .” But “laws of nature, natural phenomena, and abstract ideas are not patentable.” 35 U.S.C. § 101.

According to the Mayo/Alice framework, a claim is unpatentable under section 101 if (1) it is “directed to” a patent-ineligible concept, i.e., a law of nature, natural phenomenon, or abstract idea, and (2) the elements of the claim, considered “both individually and as an ordered combination,” do not add enough to “transform the nature of the claim into a patent-eligible application.” Alice, 134 S. Ct. at 2355.

A lack of clarity in how to apply the Mayo/Alice two-step test and the resulting uncertainty over which inventions are patentable and which are not has proved frustrating for those who apply for or attempt to assert patents in certain fields

In July, the United States Patent and Trademark Office (USPTO) published a notice requesting comments on the current state of patent eligibility jurisprudence in the U.S. and on how such jurisprudence has impacted investment and innovation. Patent Eligibility Jurisprudence Study, 86 Fed. Reg. 49,521 (Sept. 3, 2021). The notice, published at the request of a bipartisan group of senators including Thom Tillis (R-NC), Mazie Hirono (D-HI), Tom Cotton (R-AR), and Christopher Coons (D-DE), is the latest step taken by Senators Tillis and Coons to effect legislative patent eligibility reform.

Artificial intelligence (AI) is one technological area specifically identified within the questions for public comment along with quantum computing, precision medicine, diagnostic methods, pharmaceutical treatments, and other computer-implemented inventions. AI may be defined as “software and/or hardware that can learn to solve complex problems, make predictions or undertake tasks that require human-like sensing (such as vision, speech, and touch), perception, cognition, planning, learning, communication, or physical action.” USPTO, OFFICE OF THE CHIEF ECONOMIST, IP DATA HIGHLIGHTS NO. 5, INVENTING AI: TRACING THE DIFFUSION OF ARTIFICIAL INTELLIGENCE WITH U.S. PATENTS (2020) (restating U.S. National Institute of Standards and Technology (NIST) definition of AI). While a relatively new area of innovation in comparison to the others listed, AI stands out by virtue of its broad applicability across scientific and social realms and its high growth potential.

Recent spikes in inventive activity within the AI space have piqued the interest of both academic researchers and the U.S. government. The Department of Commerce recently established the National Artificial Intelligence Advisory Committee tasked with “ensuring continued United States leadership in artificial intelligence research and development” among other duties. Call for Nominations To Serve on the National Artificial Intelligence Advisory Committee, 171 Fed. Reg. 50,326 (Sept. 8, 2021). The USPTO also recently launched an initiative for evaluating innovative activity in the AI space. In collaboration with the Office of the Chief Economist (OCE), the USPTO released a report entitled “Benchmark Study on the Artificial Intelligence Patent Landscape”, which considers issued U.S. patents and pre-grant publications dating back to 1978 that contain components related to AI. USPTO, OFFICE OF THE CHIEF ECONOMIST, IP DATA HIGHLIGHTS NO. 5, INVENTING AI: TRACING THE DIFFUSION OF ARTIFICIAL INTELLIGENCE WITH U.S. PATENTS (2020). The study was spurred by similar surveying efforts by the European Patent Office (EPO) and the World Intellectual Property Organization (WIPO). In August, the USPTO publicly released the dataset underlying the study.

The study concluded from recent literature “AI invention is growing rapidly and diffusing broadly.” Between 2002 and 2018, the number of AI patent applications filed annually more than doubled from 30,000 to over 60,000, and the percentage of all patent applications that contained AI increased from 9% to roughly 16%. The study also suggested employment growth is 25 percent higher for companies holding AI patents than for companies without AI patents, and revenue growth is 40 percent higher. U.S.P.T.O., OFFICE OF THE CHIEF ECONOMIST, USPTO ECONOMIC WORKING PAPER NO. 2021-2, IDENTIFYING ARTIFICIAL INTELLIGENCE (AI) INVENTION: A NOVEL AI PATENT DATASET (2020).

Further research using the underlying data is needed to derive forecasts of the economic impact that AI will have. But in keeping with the USPTO’s mission to promote “industrial and technological progress in the United States and strengthen the national economy,” surely some patent policy makers will view the growing of AI as an indication that more should be done to ensure that AI inventions are protectable. The question remains, though—how much does AI stand to gain from patent eligibility reform, specifically in the U.S.?

Opinions offered by several of the top AI patent holders in response to the USPTO’s request in July appear split on these questions. IBM, standing in favor of eligibility reform, voiced concerns over the difficulty of “adequately claim[ing] an AI invention without focusing the claim on the desired result”:

Given that AI techniques are intended to cause a computer to learn and respond as a human would, patent applications for such techniques face the illogical situation where the more broadly an invention achieves these goals, the more difficult it is to obtain meaningful patent protection directed at the broader scope of the invention . . . . navigating the current patent eligibility jurisprudence may require unduly narrowing AI claims away from the more valuable aspects of the innovation.

Elizabeth Shaw, Request for Information: Patent Eligibility Jurisprudence Study, Comments submitted by International Business Machines, USPTO, Oct. 8, 2021, at 4.

Ericsson, also in favor of eligibility reform, pointed to examples in patent prosecution where the current eligibility doctrines either significantly delayed the grant of an application or proved fatal to the application, despite corresponding applications in foreign jurisdictions facing no eligibility rejections. Ericsson also emphasized that even a promising telecommunications patent that is granted by USPTO as novel, non-obvious patentable subject matter may still face a significant likelihood of being declared ineligible if the issue is raised during litigation post-Alice. Brian Kearns & Matteo Sabattini, Request for Information: Patent Eligibility Jurisprudence Study, Comments submitted by Ericsson, USPTO, Oct. 15, 2021, at 4, 8.

Google, on the other hand, which has obtained “nearly 11,000 U.S. patents related to AI through 2018,” claims that under the current framework, it “generates higher-quality patent applications that meet with more success both in the U.S. and in foreign patent offices” and “has not experienced issues obtaining patent protection in the U.S. for inventions in AI and QC technology, as compared to foreign jurisdictions.” Laura A. Sheridan, Request for Information: Patent Eligibility Jurisprudence Study, Comments submitted by Google, LLC, USPTO, Oct. 15, 2021, at 4, 8.

In a similar manner as Google, Dell expressed appreciation for Alice and other patent eligibility decisions, claiming “[i]n no part of this ecosystem has patent-eligibility case law factored in as a significant impediment to Dell’s business strategies” nor has patent prosecution for artificial intelligence been deterred. Krisha Gupta, Request for Information: Patent Eligibility Jurisprudence Study, Comments submitted by Dell Technologies, Inc., USPTO, (Oct. 18, 2021), at 3.

On the related issue of if and how the current patent eligibility doctrine affects global research and development and intellectual property strategies, some commented that the current doctrine incentivizes investment in other countries with more predictable eligibility frameworks, as current U.S. law is favorable to foreign companies who wish to enter the domestic market and copy inventions of U.S. companies. Others expressed a concern that under the current doctrine, patent holders may be less likely to enforce U.S. patents, instead opting to litigate disputes in other countries and taking away opportunities for the U.S. to influence global patent policy.

Some who stood against patent eligibility reform noted that applicants can obtain patents in one or many countries no matter where research and development occurs, and as such, a shift in patent eligibility would bear little on where such research and development activities are carried out.

To be clear, other factors besides modification to the existing patent eligibility framework will likely have a greater influence on whether AI is developed, deployed, or patented domestically. Data privacy is one countervailing policy consideration that may degrade the attractiveness of the U.S. market to AI developers. In some cases, AI may require large amounts of input data for “training” purposes—much as the human brain learns from experience. Depending on the regulatory regimes governing a region and a sector (e.g., healthcare, finance, or social media), such large amounts of data may be expensive or impossible to obtain in the U.S. or Europe when compared to data in other countries such as in Asia. For example, China generally imposes more relaxed data privacy and security restrictions on companies who collect information from consumers, and as of recent has seen a far higher number of patent filings for AI than the US. Patrick Thomas & Dewey Murdick, Patents and Artificial Intelligence: A PRIMER, CTR. FOR SECURITY & EMERGING TECH., 11-12 (2020).

In any case, the recent bipartisan gathering of information by congressional members regarding patent eligibility and AI may be an encouraging sign to those in favor of reform that legislative action is forthcoming. The growing ubiquity of AI will likely be a point of discussion if and when legislative hearings take place. It remains to be seen, though, what level of importance policy makers will attach to AI in the greater debate, especially considering the diverging viewpoints of prolific AI developers.

Reprinted with permission from the October 28, 2021 issue of The Legal Intelligencer ©2021 ALM Media Properties, LLC. Further duplication without permission is prohibited. All rights reserved.