Inventorship, Patenting and AI: The Public Comments on Patenting Artificial Intelligence Inventions

By Christopher P. King

Use of artificial intelligence has grown enormously in recent years. A decade ago, machine learning was a new and exotic technology—at least, for mainstream commercial applications—with few companies patenting ML-based solutions. Since then, use of AI has gradually become mainstream, with more and more companies employing ML solutions in their inventions by the mid-to-late 2010s, and concepts like neural networks becoming household names. Indeed, interest in AI has become so keen that questions previously found only in works of science fiction have begun to receive more serious consideration. One such question is the possibility of according rights in intellectual property—formerly accorded only to natural persons—to AI software, if the AI software indeed generated the IP in question. Recently, for example, a patent application was filed listing AI software “DABUS” as the sole inventor. The patent examiner rejected the application, and in a decision on April 22, 2020, the U.S. Patent and Trademark Office upheld the rejection, arguing that under the current law only natural persons may be considered inventors.

This decision based on current law, however, leads to more theoretical questions regarding how AI should be treated under the IP laws, such as the theoretical plausibility, and practical value, of granting inventorship and/or ownership in IP to AI software. Indeed, the USPTO had previously solicited input on these very questions. In August 2019, the USPTO issued a request for comment on the patenting of AI inventions. The 12 questions posed by the USPTO were wide-ranging and comprehensive, addressing not only hot-button issues such as patent eligibility under § 101 of the Patent Act (question 5), but also questions of inventorship and ownership (questions 3 and 4), whether AI impacts the level of a person of ordinary skill in the art (question 8), and prior art (question 9). They also addressed more foundational questions, such as what constitutes an “AI invention” (questions 1 and 2) and whether the patent laws should be revised to better address AI issues, such as inventorship by AI (question 3). Other questions addressed the required degree of written description and enablement (questions 6 and 7), and whether non-patent forms of protection of AI-related data would be beneficial (question 10).

Several months later, in October 2019, the USPTO additionally issued a request for comment on non-patent IP issues for AI, such as copyright, trade secret and trademark. This article will focus solely on the patent issues, however.

Areas of Agreement

The responses received from the public included 56 from individuals and 43 from organizations. Perhaps surprisingly, there was near-unanimity on a number of issues: AI is merely a tool, like any other tool. The popular imagination is captivated by the vision of an artificial general intelligence, a strong AI with intellectual capability equivalent to that of a human being, and the USPTO’s questions appear to some degree to be influenced by that vision. In contrast, the public responses by both corporations and individuals unequivocally stated that artificial general intelligence is currently of merely philosophical interest, and that in practice AI is simply a tool used by humans.

Hostility to the concept of bestowing legal personhood upon AI. Although many areas of the law have granted quasi-personhood to entities such as corporations, commentators were opposed to legislation under which an AI program could invent or own intellectual property. Many commentators argued that awarding inventorship to an AI program would not comport with the constitutional requirement to “promote the progress of science and useful arts,” since machines lack the volition that would allow them to act based on incentives to create. It was additionally noted that of the IP5—the five leading IP jurisdictions—only Europe fails expressly to define inventorship as being limited to human persons.

Need for analytical distinctions between AI categories. Commentators noted that it is important to distinguish between the degrees to which AI is involved in an invention when analyzing the relevant legal issues. For example, is the invention directed to an improvement in fundamental AI techniques themselves, or are existing AI techniques merely being used to solve a particular problem in a particular domain? Such distinctions may lead to sharply different analyses of the issues posed by the USPTO.

Eligibility concerns. AI is almost invariably implemented in software, and the law of eligibility post-Alice can be problematic for software, particularly for AI software that can be considered to represent mathematical processes, one of the categories considered by USPTO guidance to constitute an ineligible abstract idea. Some responses argued that Federal Circuit eligibility law should be revised, e.g., to focus primarily on preemption, as in the Federal Circuit’s McRO case.

Areas of Disagreement

Commentators were more divided on other issues. For example, they split over the issue of whether existing forms of IP protection are sufficient for protecting AI inventions. Several responses opined about possible need for sui generis protection of certain data used in AI inventions, such as training databases or outputs. Others, however, suggested that AI can be adequately protected by existing forms of IP or technology, such as training databases being protected by trade secret law (or perhaps even copyright law if the requisite creativity is present), machine learning models being protected by technological protections such as DRM and machine learning model output being protected by trade secret or licensing law.

Relatedly, although commentators generally agreed that existing IP standards can be properly applied to AI in the same way that they applied to other technologies, there was some disagreement about how patent disclosure requirements should be implemented. Specifically, commentators differed on the degree of detail needed to establish written description and enablement under § 112 of the Patent Act, given the black-box nature of trained models. Some (including the American Bar Association’s Intellectual Property Law section and an individual commentator) opined that vigorous enforcement, including requiring disclosure of low-level details such as model weights, may be required; others, in contrast, were more concerned about the difficulties that such disclosure would pose in the case of technologies such as deep-learning systems.

Several commentators cited a need for weakening of IP protections in certain cases to foster competition. For example, the “large data collectors” that have access to huge training datasets due to their ability to read network data or other user data were cited as guarding access to those datasets for anticompetitive reasons, rather than reasons of user privacy, leading to a recommendation that there be “controlled sharing” of the datasets.

Considerations for Those Applying for AI-Related Patents

If the public comments are representative of the future state of the law, there will be little, if any, move toward the establishment of AI personhood for purposes of inventorship or ownership, nor will there likely be significant changes to the IP laws specifically to address AI issues. And as noted on Fenwick's Bilski Blog, it seems unlikely that patent eligibility law will be altered any time soon, either judicially or legislatively. Regarding disclosure questions, applicants for AI-related patents will continue to need to weigh the risk of disclosing aspects such as parameter values that may ultimately be considered required for enablement and written description under § 112 against the option to keep such details as trade secrets.

Obviousness, Common Sense and Sensibility: Federal Circuit Ruling Offers Cautionary Tale for Patent Applicants

By Emily Bullis and John Kind

The U.S. Supreme Court’s 2007 decision in KSR International v. Teleflex altered the obviousness inquiry under 35 U.S.C. § 103 in determining whether a claimed invention passes muster under the Patent Act. The KSR Court admonished the U.S. Court of Appeals for the Federal Circuit for its rigid application of the “teaching, suggestion, or motivation test” and instead introduced flexibility into the inquiry by allowing factfinders to exercise common sense. The Federal Circuit recently revisited this topic in B/E Aerospace v. C&D Zodiac, holding that the U.S. Patent Trial and Appeal Board did not err when it used common sense to supply a missing claim limitation and invalidate two B/E Aerospace patents directed to space-saving technologies for aircraft enclosures, such as lavatories, closets and galleys.

Case Background

Zodiac challenged two B/E Aerospace patents, U.S. Patent No. 9,073,641 and U.S. Patent No. 9,440,742, in an inter partes review proceeding before the PTAB after being sued by its rival in the Eastern District of Texas for patent infringement. Each of the patents-in-suit contained a two-page written description teaching “an enclosure with contoured walls designed to ‘reduce or eliminate the gaps and volumes of space required between lavatory enclosures and adjacent structures.’” Claim 1 of the ’641 patent, which the parties agreed was representative of the challenged claims, requires that an exterior wall of the lavatory includes a first recess for part of an inclined seat back and a second recess for a portion of a seat support. The pair of recesses allow the seat to be positioned closer to the lavatory, making more efficient use of the space available.

Zodiac alleged that the challenged claims were obvious over the “Admitted Prior Art,” which Zodiac defined as certain portions of the challenged patents, and U.S. Patent No. 3,738,497 to Betts. The Board agreed, concluding that Betts’ contoured wall design met the “first recess” claim limitation and that one of skill in the art (in this case, an airplane interior designer) would have been motivated to modify the Admitted Prior Art’s flat wall with Betts’ contoured wall because such a modification would add space to the airplane cabin by allowing the seat to be moved further aft. The Board determined that adding the second recess was obvious both because it was merely “application of a known technology (i.e., Betts) for its intended purpose with a predictable result” and “‘it would have been a matter of common sense’ to incorporate a second recess.”

Federal Circuit Appeal

B/E appealed the Board’s decision to the Federal Circuit. While B/E did not challenge the finding that Betts taught the claimed “first recess” or dispute that a skilled artisan, interested in maximizing cabin space, would have been motivated to combine Betts with the Admitted Prior Art, it argued that the Board improperly incorporated a second recess not disclosed in the prior art.

The Federal Circuit panel disagreed, affirming the Board’s conclusion that a skilled artisan would have used common sense to incorporate the second recess into the combination of references. In doing so, the panel focused on the Board’s conclusion that the claims did not pass muster either under the traditional obviousness analysis or the common-sense approach and noted that the Board’s invocation of common sense was supported by more than eight pages of “reasoned analysis and evidentiary support.”

The court cited to KSR and its progeny, reiterating that while “common sense cannot be used as a ‘wholesale substitute for reasoned analysis and evidentiary support,’” “courts must ‘consider common sense, common wisdom, and common knowledge’ in analyzing obviousness.” Drawing parallels to Perfect Web Technologies v. InfoUSA, a 2009 Federal Circuit decision affirming a district court’s use of common sense to supply a missing claim limitation, the court noted that the technology at issue was similarly simple and merely involved repetition of an existing element (the “first recess”) to achieve the missing claim limitation (the “second recess”).

On appeal, B/E also argued that the Board improperly relied on three “design drawings” that Zodiac had submitted as part of its IPR petition as evidence that lower recesses to receive a seat support were known in the art. B/E asserted that the drawings were not “patents” or “printed publications” under 35 U.S.C. § 311(b) and should therefore be excluded from the proceeding. The Board denied B/E’s motion to exclude, saying that the drawings were considered only for the purpose of identifying the relevant skill in the art and thus need not qualify as “printed publication prior art.” The Federal Circuit did not reach the issue of whether the Board “ran afoul of § 311(b)” by considering the drawings, however, noting that “[t]he Board fully articulated its conclusion of obviousness” and that its finding of obviousness independent of the design drawings was supported by “substantial evidence.”

The precise contours of when common sense can be applied to provide claim elements that are missing from the prior art remain unclear. However, B/E Aerospace v. C&D Zodiac provides several pointers. The court’s analysis frequently referred to the simplicity of the technology and the fact that the value of saving space in aircraft cabins was well known at the time of invention. In other words, the panel analyzed obviousness from the perspective of an airplane interior designer—who already had the goal of moving a seat as far back as possible—faced with the problem of the seat support butting up against the lavatory wall. Furthermore, the designer already had a recess in the lavatory wall for the seat back. In that context, it is easy to see how the panel concluded adding another recess to the wall is merely common sense.

Implications for Patent Applicants

In evaluating inventions and preparing patent applications, applicants should consider the complexity of the technology and what they admit as prior art. In particular, with simple technologies, applicants should avoid admitting the problem addressed by the invention is known or discuss prior solutions in their applications unless doing so supports an argument for non-obviousness (e.g., if the known solutions are deficient and the invention is a radically different approach). Applicants should also consider the range of possible solutions and how similar their inventions are to existing solutions for related problems. In the case of B/E’s patent, to move the seat further back, there were only two options for addressing the seat support: make it smaller or add a recess for it. Furthermore, there was no meaningful distinction between adding a recess for the seat back and adding a recess for the seat support. Both problems involved a physical component of the seat colliding with the lavatory wall, and both solutions were adding a recess. Therefore, to the extent possible, applicants should identify differences between the novel features of the invention included to solve a problem and prior art solutions to similar problems.

Quick Updates

Trademark Modernization Act: Congress to Combat Fraudulent Trademark Applications

By Garner Kropp

A bipartisan, bicameral group of legislators in the U.S. Congress recently published a draft Trademark Modernization Act to reform the trademark application process and the remedies for trademark infringement. The draft TM Act—which is significant for both brand owners and consumers—has three parts. First, it allows third parties to submit evidence during a mark’s examination. Second, it provides new ex parte procedures to cancel registration of marks. Third, it restores the rebuttable presumption of irreparable harm as a remedy in a successful infringement lawsuit.

The TM Act is motivated by a dramatic increase in trademark applications from countries outside the U.S. fraudulently claiming use in domestic commerce. In Congress’ hearings last year, the U.S. Patent and Trademark Office testified that between 2013 and 2017, trademark filings from China increased 1,200%, and a large number of these filings were paid for by only three credit cards. Furthermore, two experts testified that, in a random sample of apparel trademark applications from China, two-thirds contained fraudulent specimens. Some examples included doctored photos, marks affixed to another company’s product or claiming foreign transactions as domestic ones.

To combat these fraudulent applications, the TM Act allows third parties to participate in a mark’s registration. Any third party can submit evidence “relevant to a ground for refusal of registration.” If a mark has already been registered, the TM Act creates two new procedures for a third party to petition for cancellation of a mark: expungement and reexamination. In both a petition for expungement and a petition for reexamination, a third party can submit findings from a “reasonable investigation” that the mark has never been used in commerce or in connection with the goods in petition. After review of this evidence, an examiner may cancel a mark’s registration.

These procedures have limits, however. A petition for expungement must be filed within three years of registration, and a petition for reexamination must be filed within five years. Also, upholding the registration would estop subsequent petitions against the mark.

The third part to the TM Act restores the rebuttable presumption favoring an injunction following a successful infringement lawsuit. Following eBay v. MercExchange, U.S. Circuit Courts of Appeal have been split in applying a four-factor test to determine irreparable harm and award an injunction in trademark cases. The Third, Ninth and Eleventh Circuits apply the four-factor test when evaluating remedies in trademark cases. The TM Act restores the pre-eBay rebuttable presumption for plaintiffs seeking a preliminary injunction, temporary restraining order or permanent injunction.

Congress has yet to pass the TM Act, but the strong coalition behind the bill suggests the act is likely to become law. Once it does, trademark holders should consider increasing defense of their marks before the USPTO to reduce the need for litigation.

U.S. Copyright Office: DMCA Is “Tilted Askew,” Recommends Remedies for Rightsholders

By Aylin Kuzucan

On May 21, 2020, the U.S. Copyright Office released its first full report—based on 92,000 written comments, five roundtables and decades of case law—on the Digital Millennium Copyright Act (17 U.S.C. § 512). The analysis was intended to determine whether the DMCA’s safe harbor provisions effectively balanced the needs of online service providers and rightsholders. The Copyright Office concluded that the balance is “tilted askew,” with largely ineffective copyright infringement protections for rightsholders.

When Congress enacted the DMCA in 1998, the safe harbors were designed to balance two goals: (1) preventing the potentially large economic impact of copyright infringement liability for OSPs, and (2) protecting rightsholders against “the threat of rampant, low-barrier online infringement.” In its recent report, however, the Copyright Office found that a balance has not yet been achieved.

To remedy the imbalance of the current implementation of the DMCA and better reflect the technological advances of the 21st century, the Copyright Office offered several recommendations. Notable suggestions include: (1) narrowing the eligibility of OSPs, (2) strengthening the deterrent factors of the DMCA, (3) clarifying the knowledge requirement and (4) making procedural modifications.

In narrowing the eligibility of OSPs, the Copyright Office reported that activities shielded from liability for being “related to” storage under § 512(c) is interpreted too broadly and should be limited to acts of storage or providing access to content. The Copyright Office also questioned whether technology services beyond providing internet, such as peer-to-peer systems and payment processors, should be included under § 512(a).

To strengthen the deterrent capabilities of the DMCA, the Copyright Office suggested requiring a clear, documented and publicly available repeat infringer policy and for Congress to clarify the “appropriate circumstances” for termination of a user’s account based upon repeated acts of infringement. It also suggested increased penalties for knowingly making material misrepresentation in takedown notices and counter-notices. However, the Copyright Office noted that in Lenz v. Universal Music Corp., the Ninth Circuit may have been too heavy handed in imputing a fair use in good faith requirement when issuing a takedown notice and that Congress should monitor the impact and potential liability imposed on rightsholders who submit takedown requests on infringing works without a fair use inquiry.

To better understand the knowledge requirement, the Copyright Office suggested drawing a more definitive distinction between “actual” and “red flag knowledge, under section 512(c) or (d) safe harbors” and adding a reasonableness standard to the knowledge requirement that would account for the differences among OSPs.

Procedurally, the Copyright Office suggested shifting the submission process of takedown notices to a regulatory process, which would enable the Copyright Office to set more flexible rules and “future-proof” the statute against changing communications methods. It also suggested creating an alternative statutory method for the adjudication of online infringement claims. The Copyright Office noted that the current 10 – 14 days is both too long for legitimate speech to be blocked, and too short for a rightsholder to prepare and file a federal lawsuit.

Going forward, the Copyright Office plans to post a new website—copyright.gov/DMCA—with several educational and practical elements, including model takedown notices and counter-notices. In addition, the U.S. Senate Judiciary Committee’s Subcommittee on Intellectual Property plans to draft changes to the DMCA by the end of 2020. Any changes made will be critical for the copyright community to monitor closely.

Can a Non-Owner of a Trade Secret Sue for Misappropriation? Courts Say Yes.

By Sapna S. Mehta

In Advanced Fluid Systems v. Huber, the U.S. Court of Appeals for the Third Circuit held that a company that did not own, but lawfully possessed, a trade secret could bring a misappropriation claim. Other courts, including the Fourth and Tenth Circuits, have ruled similarly, signaling a growing consensus among courts to focus on possession rather than ownership when evaluating standing in trade secret cases.

Huber involved competing hydraulics companies vying for contracts with a NASA rocket launch facility in Virginia. The trade secrets at issue were Advanced Fluid Systems' engineering drawings from its design and installation of a hydraulic system for the facility. AFS’s contract with the Virginia Commonwealth Space Flight Authority deemed these materials “work for hire” and the “exclusive property” of the space authority. Still, AFS retained them and marked them as AFS’s confidential information. The company’s retention of these materials became key in the tale of “disloyalty and deception” that followed. When a company using the facility for its rocket launch took over from the space authority, it solicited new vendors. An AFS sales engineer who was “intimately involved” in AFS’s project gave the drawings to an AFS competitor. He helped that competitor obtain the contract, while simultaneously working on (and undermining) AFS’s bid. The engineer resigned from AFS, formed his own business, and successfully outbid both companies for another project contract.

AFS, having lost out on the contracts, sued its ex-employee, his company, the other competing company, and the company that took over the facility alleging trade secret misappropriation under Pennsylvania’s Trade Secret Act. AFS won on summary judgment when the district court found the ex-employee and his company liable for misappropriation. After a subsequent bench trial, the court held the competitor to which the employee had been feeding information jointly and severally liable.

On appeal, the Third Circuit affirmed the district court’s ruling that AFS had standing to assert trade secret misappropriation even though AFS’s contract with the space authority designated the information as the space authority’s “exclusive property.” The district court had rejected the defendants’ reliance on common law references to ownership based on more recent understanding of the “knowledge-driven value” of trade secrets. The Third Circuit agreed, reasoning that neither the statutory text nor the legislative history of either the uniform law or the Pennsylvania statute discuss legal ownership as a prerequisite. Adopting the Fourth Circuit’s reasoning in an earlier case, it explained that the value of a trade secret comes from its secrecy: an ownership requirement does not account for “the substantial interest that lawful possessors of the secrets have in the value of that secrecy.” Even without an “explicit license” to AFS in its contract, the Third Circuit found sufficient facts establishing AFS’s “permission to hold and use the secrets,” thus entitling it to sue.

The Huber Court qualified its holding to “circumstances like this.” Still, the Third Circuit’s ruling provides a path for those who possess, but do not own, trade secrets to sue for misappropriation. On the other hand, AFS’s naming of the company that later operated the facility (which claimed to be the legal owner of the trade secrets following its acquisition of the space authority) as a defendant provides a caution to trade secret owners. A trade secret owner deciding to license or provide its trade secret to third-parties should be mindful that doing so could later involve litigation over its trade secrets that it does not initiate.

Federal Circuit’s Cardionet Reversal Sheds Light on Patent Eligibility in Medical Diagnosis Inventions

By Fredrick Tsang

The COVID-19 pandemic has caused many pharmaceutical companies to focus R&D on both the diagnosis and treatment of infectious diseases. When it comes to patenting potentially blockbuster products, however, diagnosis and treatment inventions can have night and day results. The U.S. Patent and Trademark Office, after the case Vanda Pharmaceuticals v. West-Ward Pharmaceuticals International, issued a memo stating that “‘method of treatment’ claims that practically apply natural relationships should be considered patent eligible.” In contrast, patent applicants seeking broad patent protection in some medical diagnosis inventions that may involve natural biological relationships may face uphill battles after the denial of rehearing en banc Athena Diagnostics v. Mayo Collaborative Services (Fed. Cir. 2019) and the Supreme Court’s denial of cert. (The Supreme Court’s refusal to take the case was surprising given that almost every judge on the Federal Circuit (including those who had voted against rehearing en banc), the United States Solicitor General, and many significant amici urged the Supreme Court to take case as a vehicle for clarifying its decision in Mayo.) Now, in a recent ruling in Cardionet v. Infobionic, the Federal Circuit has provided insight on potential avenues to patent eligibility for diagnostic inventions. Specifically, diagnosis techniques that involve improved devices or laboratory techniques may still be eligible for patent protection under 35 U.S.C. § 101.

Eligibility for life science patents has been in flux since the Supreme Court’s Mayo and Alice line of cases, which set forth a now widely used two-step framework to determine whether a claim is directed to an ineligible concept such as an abstract idea or a law of nature. After Mayo/Alice, on the diagnosis front, in Athena, the Federal Circuit held that a diagnostic method that involved the use of a man-made molecule “may still leave the claim directed to a natural law.” It is also noteworthy that diagnostic methods, other than in vivo studies, are usually patent-eligible in Europe if properly claimed.

In Cardionet, the Federal Circuit examined the patent eligibility of a diagnostic device for heart rhythm disorders claimed in U.S. Patent No. 7,941,207. The claimed device includes a logical unit to “identify a relevance of the variability in the beat-to-beat timing to at least one of atrial fibrillation and atrial flutter.” A district court, in an order granting a Rule 12(b)(6) motion to dismiss, found the claims ineligible as being drawn to automating basic diagnostic processes of collecting and analyzing a patient’s heartbeat, which doctors have long performed. The Federal Circuit reversed, holding that the claims are directed to an improved cardiac monitoring device. It noted that the district court erred in disregarding a statement in the specification touting the claimed invention as achieving “more accurate and clinically significant” detection of heart rhythm disorders.

The Cardionet case appears to indicate that courts continue to favor claims that involve a physical device. The independent claim in Cardionet broadly covers a device that determines variability of heartbeat for diagnosing heart rhythm disorders. Although the parties focused on the judicial exception of abstract ideas and did not appear to raise the issue of law of nature, it would have been interesting had the court also addressed the law of nature exception. (Is it not a natural law that a patient having a heart rhythm disorder will naturally exhibit variability in the beat-to-beat timing?) The difference in the outcomes in Cardionet and cases like Mayo shows that phrasing a diagnostic invention as an improvement to a machine (or a laboratory technique) remains a good avenue to eligibility.

For patent applicants, the Cardionet case demonstrates the importance of an early engagement with good patent counsel with expertise in subject matter eligibility of diagnosis inventions, particularly given the divergence of eligibility standards in the United States and Europe. To build a successful diagnosis patent portfolio under the challenging eligibility legal framework requires advanced planning and deliberate effort in phrasing the specification and claims in a favorable way. For example, a well-crafted statement in the specification could very well become critical evidence in defeating a 12(b)(6) motion to dismiss, as a district court is required to accept the statement as true and draw all reasonable inferences in favor of the patentee—as noted by the Federal Circuit in Cardionet.