USPTO Releases 2024 Guidance on AI Patent Subject Matter Eligibility
The U.S. Patent and Trademark Office (USPTO) has issued its 2024 Guidance on Patent Subject Matter Eligibility, along with new examples for July 2024. Our DC IP lawyers see this update as it builds on the 2019 Guidance, offering enhanced clarity and consistency for evaluating the subject matter eligibility of patent claims related to artificial intelligence (AI) technologies.
Key Insights
The 2024 Guidance outlines the conditions under which AI and machine learning (ML) inventions may qualify for patent protection according to the Alice/Mayo test. It differentiates between claims that broadly describe a desired outcome—typically deemed ineligible—and those that specify a particular method for achieving that outcome, which may be eligible. The primary takeaway is that AI inventions are more likely to be recognized as subject matter eligible when claims demonstrate a technical improvement, focus on specific hardware implementations, detail the training or implementation of the AI model, or show how the AI model integrates into an improved technological process.
Practice Pointers
The 2024 Guidance offers useful tips for patent practitioners:
-
Highlight Practical Applications: Emphasize how the invention integrates a judicial exception into a practical application, such as improvements to computer functionality or a specific technological field.
-
Leverage New Examples: Use the examples provided in the 2024 Guidance to frame claims and arguments. Detail a specific application of AI to a technological field or the precise method for achieving a desired outcome.
-
Document Technical Improvements: Clearly outline how the invention enhances existing technological processes, demonstrating that the claims represent a practical application rather than just an abstract idea. Additionally, describe any specific technical challenges the invention overcomes.
PART 1: OVERVIEW OF THE 2024 GUIDANCE
This update aligns with the previous 2019 guidelines, maintaining a consistent analytical framework for assessing patent subject matter eligibility.
The USPTO evaluates AI-related concepts under frameworks such as “mental processes” or “mathematical calculations,” with potential implications for “organizing human activity” or “fundamental economic practices” in specific AI use cases. The guidance stresses that many AI invention claims may qualify as improvements to computer functionality or other technologies. The 2024 Guidance provides insights applicable to both AI/ML and other software inventions, offering hypothetical examples to illustrate these points.
Application of the Alice/Mayo Framework
The 2024 Guidance specifies how to assess AI inventions using the two-step Alice/Mayo framework:
- Step 2A, Prong One: Determine if a claim recites an abstract idea.
- Step 2A, Prong Two: Evaluate whether the claim integrates the judicial exception into a practical application.
Under Step 2A, Prong One, the guidance explains that AI inventions may risk reciting a judicial exception in several ways: by including “mathematical concepts,” by resembling a mental process, or by merely proposing a solution without detailing implementation.
In Step 2A, Prong Two, the guidance recognizes that many AI claims are eligible as enhancements to computer or technological functionality. Eligible claims must demonstrate specific improvements as outlined in the specification.
Additionally, the 2024 Guidance reaffirms the limited role of Step 2B in the eligibility analysis. No significant changes have been made in how the USPTO interprets Step 2B, and no examples are provided of claims that are ineligible under Step 2A but eligible under Step 2B.
PART 2: NEW AI-SPECIFIC EXAMPLES
The 2024 Guidance introduces three new examples (Examples 47-49) to underscore the necessity of detailing how AI provides a “technological solution to a technological problem.” Each example presents at least one ineligible claim and one eligible claim, with eligibility determined under Step 2A, Prong Two, where eligible claims incorporate sufficient steps following ML analysis to transform abstract ideas into practical applications.