The rapid development of artificial intelligence (AI) raises new questions about whether it could in the future be considered an inventor for the purposes of patenting.
The United States Patent and Trademark Office (USPTO) entered into the discussion on August 27, 2019 when it issued a request for comments on patenting AI inventions. In October 2020, it published the responses in "Public Views on Artificial Intelligence and Intellectual Property Policy" along with statistics on current AI component technologies in "Inventing AI: Tracing the Diffusion of Artificial Intelligence with U.S. Patents." Each lends important insight into the current and future role of AI in patents.
The Meaning of 'Inventor'
Laypersons often have a general, intuitive understanding of what it means to be an inventor. They may think of an inventor as someone who is responsible for the conception of the idea behind an invention. Often they understand an inventor to be someone who provides a disclosure, whether written or oral, describing a problem the invention solves and how to make and/or use the invention. But many important rights flow from inventorship, and a flawed identification of inventorship can invalidate a patent. Accordingly, from a patenting perspective, it is important to have a clear understanding of who is and is not an inventor.
In the United States, the legal definition of inventorship begins in Title 35 of the United States Code and Title 37 of the Code of Federal Regulations (Patents, Trademarks, and Copyrights). As a starting point, these sections establish that an inventor must be a "person." For example, 35 U.S.C. §100(f) defines an inventor as "the individual or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention."
Language under the headings of Pre-AIA (American Invents Act) and/or post-AIA 35 U.S.C. §115, on or after September 16, 2012, includes gendered personal pronouns such as "he," "himself," and/or "herself." Pre-AIA and post-AIA 35 U.S.C. §116 also define the inventor and use the genderless personal pronoun "they." Specifically, 35 U.S.C. §115(b) defines the required oath and declaration statements inventors must make when filing a patent application: "[A]n oath or declaration under subsection (a) shall contain statements that... such individual believes himself or herself to be the original inventor or an original joint inventor of a claimed invention in the application." (Emphasis added).
To obtain a patent, the claimed invention must fall within one of four statutory categories and improvements thereof that Congress regarded as appropriate pursuant to 35 U.S.C. §101. However, §101 also defines who is entitled to a patent and states "[W]hoever invents or discovers ... may obtain a patent therefor, subject to the conditions and requirements of this title." (Emphasis added.) Likewise, 35 U.S.C. §102 also defines who is entitled to a patent and states "[A] person shall be entitled to a patent unless..." (Emphasis added.)
The understanding that an inventor must be a natural person was recently highlighted in an issue that arose at the USPTO. Patent Application Serial No. 16/524,350, filed in July 2019, identified a single "inventor" named DABUS with a note stating that "the invention was autonomously generated by artificial intelligence." This caused the USPTO to issue a Notice of Missing Parts because each inventor was not identified by his or her legal name. The applicant petitioned the USPTO to reconsider the allegation in the Notice of Missing Parts. In a decision on petition, the USPTO denied the applicant's petition and opined that the word "Whoever" in 35 U.S.C. §101 suggests that an inventor must be a "natural person." The decision on the petition also referenced the language in 35 U.S.C. §§100, 115, and 116. Therefore, the USPTO has found that U.S. law does not recognize a machine as a "natural person" for the purpose of patenting an invention.
Type of AI Required to Facilitate Inventorship
According to the U.S. National Institute of Standards and Technology (NIST), AI systems "comprise 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."1 Patent practitioners are accustomed to current AI technology, which employs a machine to perform "individual tasks in well-defined domains"2 as a result of the execution of programmed instructions. An inventor (i.e., natural person) can use an AI system as a tool in the development of an invention. If a natural person contributed to the conception of the claimed invention, the fact that an AI system was used by the natural person to assist in the development process does not disqualify the person from being an inventor. Many take the position that current AI machines cannot "invent ... without human intervention."3 This paradigm is important when considering the question of inventorship as defined by Title 35 of the United States Code.
Current AI technology differs from that of artificial general intelligence (AGI). AGI will be achieved "when science agrees that machines can 'think' on their own."4 AGI has not yet been fully achieved, but an AGI machine will eventually be created. So the question arises, if an AGI machine conceives of an inventive idea, what entity would be properly identified as the inventor? On a more practical level, how can the AGI machine sign an oath and declaration with a legal name, as required by 35 U.S.C. §115? These are examples of the kinds of challenges that U.S. law will need to address when AGI arrives. U.S. laws have been amended in the past to keep up with advances in technology. For example, Title 35 of the United States Code was amended in 1990 with 35 U.S.C. §105, which is directed to "Inventions in Outer Space."
Patenting AI?
Inventions involving AI can be separated into three basic categories which include "a) inventions that embody an advance in the field of AI (e.g., a new neural network structure of an improved machine learning (ML) model or algorithm)"; "b) inventions that apply AI (to a field other than AI)"; and "c) inventions that may be produced by AI itself."5
The percentage of patent application filings at the USPTO that involved AI increased from 9% in 2002 to 16% 2018.6 The USPTO has identified distinct AI clusters, divided into various classes, such as: 1) knowledge processing and planning/control; 2) vision, machine learning and AI hardware; and 3) evolutionary computing, speech and natural language processing.7 The USPTO has also identified several component technologies where AI is active based on the clusters. The components include knowledge processing, speech, AI hardware, evolutionary computation, natural language processing, machine learning, vision and planning/control.8 There are already more than 100,000 published patent applications directed to the AI component technologies. The most exploited AI component technologies include knowledge processing and planning/control. Each of these component technologies includes around 40,000 published patent applications. The number of published patent applications attributed to these two AI component technologies is almost twice that of the next-ranked AI component technology, AI hardware, which has slightly fewer than 20,000 published patent applications.
Ownership of Inventions
There is also precedent for changing U.S. laws related to patent ownership. Prior to enactment of the American Invents Act (AIA) of 2011, 37 C.F.R. 3.73(a) stated that the owner of a patent application is the inventor (i.e., natural person) unless an assignment was in place. Post-AIA, 37 C.F.R. 3.73(a) now states that the "applicant" is the owner of the patent or patent application. Currently, an applicant may be a company (or another legally created organization) and/or natural persons. However, neither the pre-AIA nor the post-AIA regulations permit a machine to be an applicant. In order to keep up with advancement in AI technology, the U.S. patent ownership laws may need to change. Ownership of a patent or patent application may need to address "a natural person: (1) who trains an AI process, or (2) who owns/controls an AI system."9
One thing is certain, AI technology has and will continue to affect intellectual property far into the foreseeable future. More importantly, in addition to inventorship and ownership issues, AI technology may impact the "ordinary skill in the art" standard that we are accustomed to and is used by the USPTO and the courts in evaluating Intellectual Property.10
Footnotes
1 Inventing AI Tracing the Diffusion of Artificial Intelligence with U.S. Patents, pg. 3.
2 Public Views on Artificial Intelligence and Intellectual Property Policy, pg. ii.
3 Id.
4 Public Views on Artificial Intelligence and Intellectual Property Policy, pg. 6.
5 Public Views on Artificial Intelligence and Intellectual Property Policy, pgs. 1-2.
6 Inventing AI: Tracing the Diffusion of Artificial Intelligence with U.S. Patents, pg. 5.
7 Inventing AI: Tracing the Diffusion of Artificial Intelligence with U.S. Patents, pg. 8.
8 Inventing AI: Tracing the Diffusion of Artificial Intelligence with U.S. Patents, pg. 3.
9 Id.
10 Public Views on Artificial Intelligence and Intellectual Property Policy, pg. iii.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.