It is no secret that the Supreme Court decision in Alice made securing a software-related patent difficult. Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 573 U.S. 208 (2014). Statutory subject matter eligibility criteria are murky and application of US Patent and Trademark Office (USPTO) rules is subjective. This has increased patent prosecution expense and frustration, and disincentivizes filing patent applications. In the more than 10 years since Alice was decided, the USPTO has issued multiple guidance updates aimed at curing problems created by this landmark decision. With each update, the pendulum swings back-and-forth from inventor-hostile to inventor-friendly postures regarding issuance of patents, particularly related to those that are AI-based.

For a number of years, the inclusion of artificial intelligence in the claims of a patent application appeared to help in overcoming Alice-based § 101 rejections. Under the Alice framework, if a claim recites an abstract idea, the next step is to determine whether that abstract idea is integrated into a practical application. The test for integration can be satisfied in a number of ways, including by “improvements to the functioning of a computer or to any other technology or technical field.” MPEP § 2106.05(a). The USPTO routinely accepted arguments that AI improved the functioning of a computer, and even when examiners resisted, success could often be achieved by including training of the AI model, even though training is required for AI to function.

At some point, the USPTO changed its position, and determined that this approach to AI was unsustainable. In July 2024, the USPTO issued a guidance update on patent subject matter eligibility focused specifically on AI, as well as a number of eligibility examples illustrating how AI-based applications should be evaluated. This updated guidance and the inclusion of new examples effectively ended token use of AI to overcome 101-based rejections (i.e., making the USPTO more inventor-hostile).

Recently, however, the pendulum once again has begun to swing back to inventor-friendly. First, in August 2025, the new Deputy Commissioner for Patents issued a memorandum with “[r]eminders on evaluating subject matter eligibility of claims under 35 U.S.C. 101.” This memorandum did not provide new guidelines, but strongly implied that the USPTO is, and has been, applying the existing framework too harshly against inventors.

A December 2025 memorandum followed, providing “notice of change to the MPEP in light of Ex Parte Desjardins,” a precedential Patent Trial and Appeal Board (PTAB) decision addressing patent eligibility under § 101 for a machine learning invention. The PTAB panel found that claims directed to a method of training a machine learning model were not merely an abstract idea, as they provided a specific technological improvement—namely, overcoming “catastrophic forgetting” and enhancing model performance—thus integrating the abstract idea into a practical application. The December memorandum provides notice that Desjardins serves to revise the latest version of the Manual of Patent Examining Procedure, but also states that “[t]hese updates are not intended to announce any new USPTO practice or procedure and are meant to be consistent with existing USPTO guidance.” 

Regardless, by recognizing the eligibility of AI-specific improvements, this shift again signals a return to a more inventor-friendly treatment of AI-related applications. The interpretation goes beyond prior USPTO examples that focused more narrowly on improvement to traditional hardware or operating systems. The latest memorandum also expands the scope of what qualifies as “improvements to the functioning of a computer,” informing examiners that improvements to the internal operation of AI systems, such as the way machine learning models are trained or retain information, should be viewed as practical applications, and thus as eligible subject matter.

While it is still too early to completely understand the full practical effect of the December 2025 memorandum, in combination with the August 2025 memorandum, it clearly signals a desired shift at the USPTO toward higher patent grant rates for AI-based patent applications.