Executive Summary
Patent practitioners have seen a shifting landscape for patent eligibility under 35 U.S.C. § 101 since the Supreme Court’s 2012 and 2014 seminal decisions in Mayo and Alice. Now, the United States Patent and Trademark Office issued guidance on November 28, 2025, rescinding large portions of artificial intelligence (AI) guidance that was issued under President Biden. The new framework treats AI systems as tools only and decrees that no separate eligibility standard applies when examiners consider applications for AI-assisted inventions.
Litigators and in-house counsel who are frequently targeted by lawsuits related to software and business logic patents should reassess how and when to assert § 101 eligibility challenges. The Biden-era guidance, Inventorship Guidance for AI-Assisted Inventions, 89 Fed. Reg. 10043 (Feb. 13, 2024), allowed broader acceptance of such patent claims by requiring application of the Pannu factors (addressing joint inventorship among multiple natural persons) to AI-assisted inventions; in contrast, the new guidance pulls examiners back to traditional Alice/Mayo principles and human conception, narrowing the penumbra afforded to AI-assisted inventions.
In addition, the Federal Circuit has approved of § 101 challenges presented as early motions to dismiss in U.S. district courts, even on machine learning claims, when the claims deviate from the new requirements. As a result, recently issued AI/software patents may be newly vulnerable to eligibility challenges and should be scrutinized under the new framework.
USPTO’s November 28, 2025 Guidance
The USPTO formally rescinded the February 2024 Biden-era guidance that allowed AI systems to perform acts that, if performed by a human, could constitute inventorship, consistent with decisions such as Thaler v. Vidal, 43 F.4th 1207, 1213 (Fed. Cir. 2022), cert. denied, 143 S. Ct. 1783 (2023) (holding that an inventor must be a natural person, and that an AI system cannot be named as an inventor on a patent application). The February 2024 guidance provided that “a natural person must have significantly contributed to each claim in a patent application or patent,” requiring careful scrutiny of prompts presented to the AI system to determine whether the natural person made significant contributions. Inventorship Guidance for AI-Assisted Inventions, 89 Fed. Reg. 10048 (Feb. 13, 2024), available here.
Under the November 2025 policy, the USPTO states that inventorship factors under Pannu v. Iolab Corp., 155 F.3d 1344, 1351 (Fed. Cir. 1998), apply only when determining whether multiple natural persons potentially qualify as joint inventors and cannot be applied to hold AI systems as “joint inventors.” Revised Inventorship Guidance for AI-Assisted Inventions, 90 Fed. Reg. 54636 (Nov. 28, 2025), available here.
Additionally, the USPTO clarified that there are no special pathways for prosecution on AI-assisted inventions and that examiners should review inventorship under the same legal standards for all applications. In practical terms, this means that an AI system listed as an inventor on an application, regardless of the system’s sophistication, cannot be named as an inventor or a joint inventor and is subject to rejection under 35 U.S.C. § 101 or §115 because generative AI and other computational models are merely tools for use by inventors who are natural persons. This also has implications for claiming the benefit of priority for earlier applications or international applications where a listed inventor was an AI tool, as there must be either the same natural person as the inventor or at least one joint inventor who is a natural person in common between the applications. Requiring applicants to anchor inventive concepts in human effort creates fertile ground for litigation challenges.
Intersection with § 101 Litigation Trends
This reset for the USPTO is in accord with the latest Federal Circuit guidance as well. In the Recentive Analytics decision from April 2025, the Federal Circuit affirmed a district court’s dismissal of the complaint under § 101 for patents that merely applied known machine learning methods within a new data environment. Recentive Analytics, Inc. v. Fox Corp., 134 F.4th 1205 (Fed. Cir. 2025). The court held that merely implementing an abstract idea of data analysis and scheduling by using generic machine-learning models on conventional computing infrastructure does not represent technological improvement sufficient to move the claims out of the realm of abstract ideas. Id. at 1212-14.
Nor is the use of an algorithm a get-out-of-jail-free card for inventive concept. The Federal Circuit also stressed that, even if inventors point to improved accuracy or efficiency achieved by using known models for machine learning, this does not transform an abstract idea into patent-eligible subject matter under § 101 (id. at 1214-15), which is consistent with Alice/Mayo and their progeny. Thus, under this precedential decision, practitioners can anticipate new challenges in the USPTO and district courts to AI or software patents that rely on generic computing functions or high-level references to algorithms without concrete details on how to implement the claims.
In litigation in particular, patents related to AI or software with claims that merely automate existing methods using generic computing are prime targets for early motions to dismiss addressing the subject-matter eligibility. Such motions, evaluated and closely tracked with other early defenses, such as venue or sufficiency of pleadings, empower defendants to identify the most vulnerable weakness and present aggressive early motions where appropriate. Other times, where eligibility might not be so clear, parallel prior art invalidity challenges by way of inter partes review or ex parte reexamination may be the most cost-effective route.
Venable will continue to monitor the shifting landscape of 35 U.S.C, § 101 and stands ready to assist companies facing software or AI-related claims of patent infringement.