AI Bulletin – December 2021
Want the AI Bulletin mailed to you?
- Thaler v. Hirshfeld: AI cannot be an “inventor” under the Patent Act
- Artificial intelligence technology increasingly deployed during COVID-19 pandemic
- Facebook ends facial recognition
- Proposed Restoring America’s Leadership in Innovation act would expand the scope of patentable subject matter, particularly for artificial intelligence technology
- Numerous state bills about artificial intelligence proposed and enacted in 2021
- National Artificial Intelligence Advisory Committee (NAIAC) call for nominations
- QE Victory: $14 million verdict for Proofpoint in trade secret and copyright trial
- QE’s Kathleen Sullivan selected for lifetime achievement award
SPOTLIGHT – Thaler v. Hirshfeld
District Judge holds that an artificial intelligence cannot be an “inventor” of a patent.
In the closely-watched Thaler v. Hirshfeld in the Eastern District of Virginia, District Judge Leonie Brinkema granted summary judgement on September 2, 2021 to Andrew Hirshfeld, acting director of the USPTO, holding that an artificial intelligence machine cannot be named as an “inventor” on a patent application.
Background. Plaintiff Stephen Thaler had filed two patent applications with the USPTO which listed Thaler’s artificial intelligence system “DABUS” (Device for the Autonomous Bootstrapping of Unified Sentience) as the inventor. One application was addressed to a “Neural Flame” light beacon “that flashes in a new and inventive manner to attract attention.” Another claimed a “beverage container based on fractal geometry.” The USPTO refused to process the applications. Thaler sought judicial review under the Administrative Procedure Act and requested that the applications be reinstated, as well as a judicial declaration “that a patent application for an AI-generated invention should not be rejected on the basis that no natural person is identified as an inventor.”
Decision. Judge Brinkema found that the USPTO’s interpretation of the statutory language was entitled to deference. Additionally, Judge Brinkema independently examined the statute and found that the USPTO’s interpretation was in accord with the plain language. For example, 35 U.S.C. § 100 defines an “inventor” as “the individual, or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention.” Judge Brinkema found the Supreme Court’s analysis of the term “individual” in the context of another statute, the Torture Victim Protection Act, in Mohamad v. Palestinian Authority, 566 U.S. 449 (2012) to apply equally to the Patent Act. In Mohamad, the Supreme Court ruled that “‘individual’ ordinarily means ‘a human being, a person.’” Judge Brinkema found that the same reasoning should apply in the context of patents because “artificial intelligence machines or systems are not normally referred to as ‘individuals’ in ordinary parlance.” There was no evidence that Congress had intended anyone other than human beings to be “inventors.”
That interpretation was buttressed by several other statutory provisions about inventorship, including the requirement in 35 U.S.C. § 115(b)(2) that in executing the oath or declaration accompanying a patent application, the inventor must include a statement that “such individual believes himself or herself to be the original inventor or an original joint inventor of a claimed invention in the application.”
The interpretation was also consistent with two prior Federal Circuit cases. In University of Utah v. Max-Planck-Gesselschaft, 734 F.3d 1315 (Fed. Cir. 2013), the Federal Circuit held that a university could not be a named inventor and that “inventors must be natural persons and cannot be corporations or sovereigns.” Similarly, in Beech Aircraft Corp. v. Edo Corp., 990 F.2d 1237 (Fed. Cir. 1993), the Federal Circuit held that a corporation could not be a named inventor and that “only natural persons can be ‘inventors.’”
Thaler relied heavily on policy arguments rather than canons of statutory interpretation, contending that “AI-Generated Inventions will result in more innovation” and that “allowing people to take credit for work they have not done would devalue human inventorship.” Judge Brinkema found that policy concerns could not “override the plain meaning of a statutory term.” Judge Brinkema also noted that the USPTO had solicited comments on invention by artificial intelligence entities and had issued a report in October 2020. The USPTO report summarized the views of numerous commentators that allowing artificial intelligences to be named inventors would be bad policy chiefly because it would be premature: no true artificial general intelligence exists, so it doesn’t make sense to ascribe concepts of “inventorship” or “conception” to the limited forms of artificial intelligence that do exist.
Treatment in Other Jurisdictions. Thaler has sought similar declarations that an artificial intelligence machine may be a named inventor in the UK, EU, Australia, and South Africa. To date, the UK and EU (EPO) have concurred with Judge Brinkema and found that under their relevant statutes, an inventor must be a human being. However, an Australian Federal Court issued a decision that none of the provisions of the Australian Patent Act “exclude an inventor from being a non-human artificial intelligence device or system.” Thaler’s patent applications were also allowed in South Africa.
Further Developments. Thaler’s appeal to the Federal Circuit is currently pending. Given the clear language of the statute and lack of any appetite in Congress to allow artificial intelligences to be named inventors, the status quo seems likely to persist in the United States for the near future.
OTHER RECENT DEVELOPMENTS
- On November 2, 2021, Facebook ended use of facial recognition technology for automated photo tagging in the wake of a previous $650 million settlement with users who alleged that Facebook tagged them in photos without their permission. In additional to not automatically tagging photos in the future, Facebook will also delete previous automatic tags and “more than a billion people’s individual facial recognition templates.” Facebook noted that although facial recognition helped generate image descriptions for blind users, there were “many concerns about the place of facial recognition technology in society.”
ARTIFICIAL INTELLIGENCE TECHNOLOGY
- A Harvard Business Review article examined the trends in adoption of artificial intelligence technology over the COVID-19 pandemic and found that businesses accelerated AI adoption plans. 86% of survey respondents said that AI is becoming a mainstream technology at their company in 2021. The labor shortages caused by the pandemic increased the deployment of AI. AI is becoming particularly useful in navigating supply chain issues and dealing with large sets of data.
GOVERNMENT & REGULATION
- On November 4, 2021, several Republican Representatives introduced H.R.5874, Restoring America’s Leadership in Innovation Act of 2021. Among other things, the bill would abrogate the Supreme Court’s recent jurisprudence on subject matter eligibility to broaden the scope of patentable subject matter. The bill would allow patents “if the claimed invention as a whole, as understood by a person having ordinary skill in the art, exists in nature independently of and prior to any human activity, or exists solely in the human mind.” If enacted, the bill would eliminate several of the patentable subject matter doctrines that have been used in recent years to limit the patenting of artificial intelligence technology.
- Artificial intelligence-related bills were introduced in at least 17 states in 2021. Many of the bills related to the risks posed by biased or discriminatory AI systems. Colorado’s enacted S.B. 169 prohibits insurers from using algorithms or predictive models that unfairly discriminate based on protected characteristics. Illinois’ enacted IL H.B. 53 establishes a data gathering system for employers that use artificial intelligence systems in hiring, in order to determine whether the systems have a racial bias. Other proposed bills related to the auditability of artificial intelligence systems and potential discrimination in creditworthiness decisions, housing, unemployment, and other areas.
- On September 7, 2021 the Department of Commerce issued a call for nominations to serve on the National Artificial Intelligence Advisory Committee (NAIAC). NAIAC was created by the National Artificial Intelligence Initiative Act of 2020 and has a broad mandate to provide advice to the President and others regarding the development and deployment of artificial intelligence systems. The USPTO encouraged intellectual property professionals to apply.
QE IN THE NEWS
- After a four-week jury trial before Judge Chesney in the Northern District of California, Quinn Emanuel won a $14 million verdict for client Proofpoint in a trade secret misappropriation and copyright case involving AI-assisted cybersecurity technology. Proofpoint brought the action against Vade Secure (a French cybersecurity company) and its CTO Olivier Lemarie (a former Proofpoint employee) in 2019 based on evidence that Lemarie and Vade Secure had misappropriated highly confidential source code for Proofpoint’s anti-spearfishing software. In addition to finding that the defendants had misappropriated fifteen separate Proofpoint trade secrets and infringed copyrights in its proprietary source code, the jury found that defendant Vade Secure’s misappropriation was willful and malicious. The QE team was led by Sean Pak (Co-Chair of the firm’s National Intellectual Property Litigation Practice) and Iman Lordgooei. Proofpoint’s motion for enhanced damages, based on the jury’s finding of willfulness, is currently pending before the district court.
- Quinn Emanuel partner Kathleen Sullivan was selected by the American Lawyer for a Lifetime Achievement Award. https://www.law.com/americanlawyer/2021/10/01/the-american-lawyer-announces-2021-attorney-of-the-year-finalists-and-lifetime-achievement-awards/. Kathleen is the former Dean of Stanford Law School. Kathleen is widely recognized as one of the nation’s top appellate advocates. She has argued eleven cases in the United States Supreme Court, and is the first female name partner at an Am Law 100 firm. Kathleen was also profiled in the December issue of the American Lawyer, which described her path from academia to leading Quinn Emanuel’s appellate practice. https://www.law.com/therecorder/2021/11/23/from-law-school-to-the-courtroom-kathleen-sullivan-has-been-an-educator-at-every-turn/.
Quinn Emanuel Urquhart & Sullivan, LLP is a 875+ lawyer business litigation firm – the largest in the world devoted solely tobusiness litigation and arbitration. Our lawyers have tried over 2,500 cases and won 86% of them. When we representdefendants, our trial experience gets us better settlements or defense verdicts. When representing plaintiffs, our lawyers havewon over $70 billion in judgments and settlements. We have also obtained seven 9-figure jury verdicts, four 10-figure juryverdicts, fifty-three 9-figure settlements, and nineteen 10-figure settlements.
No longer want to receive these emails? Just reply and ask to unsubscribe