Subject matter eligibility
The basic principles of subject matter eligibility doctrine are easy to recite. Patents can protect inventions that can be classified as a process, machine, manufacture, or composition of matter. Judicial exceptions to patentable subject matter include abstract ideas, laws of nature, and natural phenomenon. Abstract ideas include mathematical concepts, certain methods of organizing human activity, and mental processes. Inventions that merely involve a judicial exception are ineligible for patent protection, while inventions that apply the judicial exception to improve technology are protectable by patent.
Yet applying these basic principles to specific inventions is surprisingly challenging. Indeed, even though 10 years has passed since the Supreme Court’s Alice opinion - Alice Corp. v. CLS Bank Int’l, 573 U.S. 208 (2014) - which set forth the legal framework for determining whether an invention is abstract, the lower courts and patent bar continue to struggle with applying that framework to the specific facts of each case. As a result, an answer to the foundational question – is an invention patent eligible? – is often difficult to predict and riddled with uncertainty. And despite the USPTO’s recent guidance, this unpredictability will continue to plague AI-related inventions.
USPTO guidance
The USPTO’s AI guidance is largely a summary of Supreme Court and Federal Circuit patent eligibility decisions, none of which involved AI. This is likely because the USPTO does not have the authority to set the law, and the guidance itself recognizes that it “does not have the force and effect of law.” Although courts will eventually shape the law on eligibility for AI-related inventions, the USPTO’s guidance is nonetheless valuable because it reflects the agency’s understanding of Section 101 precedent, sets out agency’s policy in this area of law, and instructs patent examiners on how to apply the law.
3 AI-related patent examples
The most informative part of the USPTO’s AI guidance actually lies in a separate publication on the agency’s website, in which the agency analyzes three example AI-related patents.
-
a neural network to detect anomalies in network traffic (example 47)
-
technology for distinguishing speech from background noise (example 48)
-
an AI model designed to personalize medical treatment for a patient (example 49)
In these examples, the USPTO provides detailed analysis regarding the application of subject matter eligibility law to hypothetical claims, both on a limitation-by-limitation basis and for the claim as a whole. Yet even such detailed analysis and explanation (spanning 35 pages for 3 example patents and 8 claims), the USPTO’s guidance actually underscores the unpredictable and sometimes unintuitive outcomes that result from the agency’s analytical framework.
Example 49 – counter-intuitive outcomes
A great illustration of counter-intuitive outcomes is present in the USPTO’s third hypothetical AI patent (Example 49), pertaining to personalizing medical treatment using an AI model and presenting two claims:
-
[Claim 1] A post-surgical fibrosis treatment method comprising:
-
(a) collecting and genotyping a sample from a glaucoma patient to a provide a genotype dataset;
-
(b) identifying the glaucoma patient as at high risk of post-implantation inflammation (PI) based on a weighted polygenic risk score that is generated from informative single-nucleotide polymorphisms (SNPs) in the genotype dataset by an ezAI model that uses multiplication to weight corresponding alleles in the dataset by their effect sizes and addition to sum the weighted values to provide the score; and
-
(c) administering an appropriate treatment to the glaucoma patient at high risk of PI after microstent implant surgery.
-
[Claim 2] The method of claim 1, wherein the appropriate treatment is Compound X eye drops.
The USPTO guidance concluded that claim 1 was patent ineligible as an abstract idea, while claim 2 – which merely adds that the “treatment is Compound X eye drops” – is patent eligible. What follows is a high-level summary of the agency’s reasoning, without delving into the particular steps and sub-prongs used in its analysis.
The USPTO concluded that claim 1 was patent ineligible by first identifying judicial exceptions that appear in the claim. Specifically, the guidance determined that the “identifying” limitation (b) recites: (1) a “mental process” because comparing a patient’s risk score to quartile scores can be performed in the human mind; (2) a “law of nature” because it describes the naturally occurring relationship between a patient’s genes and their risk for post-implantation inflammation; and (3) a “mathematical concept” because calculating a weighted risk score using multiplication and addition is arithmetic.
Next, the USPTO looked at the limitations that don’t recite judicial exceptions – i.e., limitations (a) and (c) – to determine whether the claim “as a whole integrates the recited judicial exception into a practical application of the exception.” The USPTO concluded it did not. The “collecting” step (a), it reasoned, was mere data gathering and thus insignificant extra-solution activity. The “administering” step (c) was no more than a generic instruction to “apply” the judicial exception to a field of use because it was not constrained to a particular manner or type of treatment.
In contrast, the USPTO concluded that claim 2 was patent eligible because requiring the administration of a particular treatment placed meaningful limits to the judicial exceptions in claim 1. Accordingly, the guidance concluded that claim 2 “integrates the judicial exception into a practical application” by administering a particular treatment (Compound X eye drops) to a specific patient population (glaucoma patients at high risk of post-implantation inflammation).
Under the USPTO’s logic, a claim which simply stated “A treatment method comprising administering Compound X eye drops to glaucoma patient at high risk of post-implantation inflammation” should similarly be patent eligible. Which begs the next question, would the courts agree? Perhaps only time can tell.
Next steps
As patent applications claiming AI-related inventions make their way through the Patent Office, many applicants will face significant challenges overcoming rejections based on subject matter eligibility. If you or your company are seeking patents on AI-related inventions, our team of patent attorneys are ready to help you navigate these murky waters. In addition to helping our clients obtain strong eligible patents, our familiarity with Patent Office policy also enables us to provide wholistic advice to clients facing patent disputes, including defending against AI-related patents in litigation or IPRs.
