August USPTO Memo Signals Nuanced Approach to AI and Software Patent Eligibility
"Examiners should..." for each prong of the Step 2A analysis
On August 4, 2025, the United States Patent and Trademark Office (USPTO) issued a memorandum to its examiners in key technology centers, offering "important reminders" on evaluating subject matter eligibility for claims under 35 U.S.C. § 101. The document, authored by Deputy Commissioner for Patents Charles Kim, specifically targets the challenges examiners face with software, Artificial Intelligence (AI), and Machine Learning inventions. The result may be a more nuanced approach.
While the memo asserts it "is not intended to announce any new USPTO practice or procedure," its clarifications—particularly around abstract ideas, practical applications, and technical improvements—may provide a clearer path for patenting innovations in these critical fields (Memo, p. 1).
The guidance arrives as patent professionals continue to work through the framework established by court decisions, which can be difficult to apply to rapidly advancing technologies like AI. This memo appears to be an effort by the USPTO to ensure consistent and careful application of its existing examination procedures, which are detailed in the Manual of Patent Examining Procedure (MPEP).
A Limit on the "Mental Process" Exception
A significant portion of the memo addresses the "mental process" grouping of abstract ideas, a category that frequently ensnares software-related claims. The USPTO’s guidance distills case law into three groupings of abstract ideas: "(1) mathematical concepts, (2) certain methods of organizing human activity, and (3) mental processes" (p. 2).
The memo reminds examiners that a mental process is one that "can be performed in the human mind, or by a human using a pen and paper" (p. 2). However, it firmly cautions against an overreach of this exception. It states, "Examiners are reminded not to expand this grouping in a manner that encompasses claim limitations that cannot practically be performed in the human mind" (p. 2).
This is particularly relevant for AI and machine learning inventions. The guidance clarifies that if a claim involves processes that a human mind is simply not equipped to perform, it does not fall into the mental process category.
The memo points out that, "Claim limitations that encompass AI in a way that cannot be practically performed in the human mind do not fall within this grouping" (p. 2). This clarification is a welcome signal for inventors whose creations, by their very nature, operate at a scale and speed far beyond human cognitive ability.
Reciting vs. Involving an Exception
The memo also draws a critical distinction for practitioners to note in their claim drafting: the difference between a claim that recites a judicial exception and one that merely involves one (p. 1).
A claim that recites an exception is subject to the full eligibility analysis, whereas a claim that only involves one is eligible without further scrutiny (p. 3).
To illustrate, the memo contrasts two examples from their subject matter eligibility guidance:
The claim limitation "training the neural network in a first stage using the first training set" of example 39 does not recite a judicial exception. Even though "training the neural network" involves a broad array of techniques and/or activities that may involve or rely upon mathematical concepts, the limitation does not set forth or describe any mathematical relationships, calculations, formulas, or equations using words or mathematical symbols (p. 3).
Contrast this with the limitation "training, by the computer, the ANN based on the input data and a selected training algorithm to generate a trained ANN, wherein the selected training algorithm includes a backpropagation algorithm and a gradient descent algorithm" of claim 2 of example 47. This limitation requires specific mathematical calculations by referring to the mathematical calculations by name, i.e., a backpropagation algorithm and a gradient descent algorithm, and therefore recites a judicial exception, namely an abstract idea (p. 3).
This distinction highlights the importance of how an invention is described in the claims. Focusing on the application and function of a process, rather than naming the specific underlying mathematical formulas, could be a key strategy in drafting claims that avoid an initial finding of ineligibility.
The Importance of the Claim "As a Whole"
For claims that do recite a judicial exception, the analysis proceeds to Step 2A Prong Two, which asks if the claim as a whole integrates the exception into a "practical application" (p. 3).
The memo stresses that examiners must not evaluate claim limitations in isolation.
The analysis in Step 2A Prong Two considers the claim as a whole. The way in which the additional elements use or interact with the exception may integrate the judicial exception into a practical application. Accordingly, the additional limitations should not be evaluated in a vacuum, completely separate from the recited judicial exception. Instead, the analysis should take into consideration all the claim limitations and how these limitations interact and impact each other when evaluating whether the exception is integrated into a practical application (pp. 3-4, citing MPEP 2106.04(d)(III)).
This holistic approach is critical for complex software and AI inventions, where the inventiveness often lies in the specific combination and interaction of elements, rather than in a single component.
The Search for a "Technological Solution"
A key pathway, under Step 2A Prong Two, to establishing a practical application is showing that the invention provides an "improvement to the functioning of a computer or to another technology or technical field" (p. 4). The memo also refers to this as a "search for a technological solution to a technological problem" (p. 4).
The memo offers some clarity as to what is required in the application to support such an improvement. The specification still is the key, but it does not have to use magic words. The memo states, "The specification does not need to explicitly set forth the improvement, but it must describe the invention such that the improvement would be apparent to one of ordinary skill in the art" (p. 4).
Furthermore, the "claim itself does not need to explicitly recite the improvement described in the specification" (p. 4). This provides at least a modicum of flexibility, allowing the specification to tell the full story of the invention's technical contribution, which can then be used to argue for the eligibility of the claims.
The memo also cautions examiners against misusing the "'apply it' consideration," where claims are dismissed as merely instructing the implementation of an abstract idea on a generic computer (p. 4). It notes the overlap between the "apply it" and "improvements" considerations and directs examiners to consider (pp. 4-5):
“Whether the claim recites only the idea of a solution or outcome” with insufficient details of “how” or no particular solution/way;
“Whether the claim invokes computers or other machinery merely as a tool to perform an existing process, or whether the claim purports to improve computer capabilities or to improve an existing technology”; and
“The particularity or generality of the application of the judicial exception”
A More Favorable Path for AI Patents?
This memo, while not new law, provides a consolidated and focused set of instructions that appear favorable for AI and software patent applicants. The explicit limitation on the "mental process" doctrine is a significant benefit. By stating that processes which "cannot practically be performed in the human mind" are not abstract mental steps, the USPTO is acknowledging the unique nature of AI (p. 2).
This guidance could prevent examiners from dismissing complex AI models as mere automation of human thought. The emphasis on finding a "technological solution to a technological problem" and allowing the specification to provide the evidence of that solution, even if not recited in the claims, gives patent practitioners a clear target (p. 4).
Despite the optimistic signals, challenges remain. The determination of what "cannot practically be performed in the human mind" is still subjective and will likely be a point of debate between examiners and applicants (p. 2).
Likewise, successfully arguing that a claim offers a "technological improvement" rather than just "applying" an abstract idea on a computer requires a well-drafted application with a strong technical disclosure (p. 4).
Applicants who fail to clearly articulate the technical problem and their specific solution in the specification will struggle to take advantage of this guidance. There is also the risk that some examiners may continue to "oversimplify claim limitations," a practice the memo specifically cautions against (p. 4).
A Higher Bar for Rejection
Perhaps one of the most impactful reminders for day-to-day prosecution is the memo’s instruction on the standard for making a rejection.
Examiners are told that if eligibility is a "close call," a rejection should only be made if it is "more likely than not (i.e., more than 50%) that the claim is ineligible" (p. 5).
The memo plainly states, "A rejection of a claim should not be made simply because an examiner is uncertain as to the claim's eligibility" (p. 5).
This is a powerful statement that should encourage examiners to resolve doubt in favor of the applicant, potentially reducing the number of marginal § 101 rejections.
In practice, arguing about the rejection threshold on the record may undermine a long-term strategy, discussing fact-driven factors of a “close call” in an examiner interview should be fair game.
Nice Quotes to Keep Around
2A Prong 1 – Alice Step 1
"Examiners are reminded not to expand this grouping in a manner that encompasses claim limitations that cannot practically be performed in the human mind." (p. 2).
“Examiners should be careful to distinguish claims that recite an exception (which require further eligibility analysis) from claims that merely involve an exception (which are eligible and do not require further eligibility analysis)” (p. 3, citing MPEP 2106.04, subsection II(A)(1)).
2A Prong 2 – Alice Step 2
“…the analysis should take into consideration all the claim limitations and how these limitations interact and impact each other...” (pp. 3-4).
“The examiner is reminded to consult the specification to determine whether the disclosed invention improves technology or a technical field, and evaluate the claim to ensure it reflects the disclosed improvement. The specification does not need to explicitly set forth the improvement, but it must describe the invention such that the improvement would be apparent to one of ordinary skill in the art. The claim itself does not need to explicitly recite the improvement described in the specification." (p. 4).
"Examiners are cautioned not to oversimplify claim limitations and expand the application of the 'apply it' consideration." (p. 4).
“Claims that are determined to improve computer capabilities or improve technology or a technical field support a finding that the claim integrates the judicial exception into a practical application or amounts to significantly more than the judicial exception itself.” (p. 5).
Last Thoughts
Sneaking one past the goaltenders at the USPTO should never be the primary goal—patents should be drafted with enough support to survive a district court or Federal Circuit challenge. Following CAFC case law is still the best bet.
This USPTO memorandum provides a valuable window into the agency's thinking on patent eligibility for software and AI. It signals a move towards a more nuanced analysis that recognizes the unique characteristics of these technologies.
Hopefully this leads to more consistent § 101 examinations where claims are viewed as a whole and technical elements are not ignored
For patent attorneys, in-house counsel, and inventors, the memo serves as a guide for drafting stronger applications. The key is to focus on the technical substance: clearly describe how the invention improves computer functionality or solves a technical problem in a (particular) way that a human mind practically could not.
While the path to patenting AI is not without its hurdles, this guidance provides helpful signposts that, if followed, could lead to more successful outcomes at the prosecution level.
Citation: “Reminders on evaluating subject matter eligibility of claims under 35 U.S.C. 101,” August 4, 2025 (https://www.uspto.gov/sites/default/files/documents/memo-101-20250804.pdf) (“2025 Reminders Memo”).
Disclaimer: This is provided for informational purposes only and does not constitute legal or financial advice. To the extent there are any opinions in this article, they are the author’s alone and do not represent the beliefs of his firm or clients. The strategies expressed are purely speculation based on publicly available information. The information expressed is subject to change at any time and should be checked for completeness, accuracy and current applicability. For advice, consult a suitably licensed attorney and/or patent professional.