In recent years, the landscape of software patent eligibility in the United States has undergone significant transformations, particularly in the realm of artificial intelligence (AI) and machine learning (ML). These changes are largely influenced by evolving guidelines from the U.S. Patent and Trademark Office (USPTO) and pivotal court decisions.
Historically, the Supreme Court’s decisions in Alice Corp. v. CLS Bank International and Mayo Collaborative Services v. Prometheus Laboratories, Inc. established a framework that rendered many software-related inventions ineligible for patents under 35 U.S.C. § 101. These rulings deemed that merely implementing abstract ideas on a computer does not qualify as patentable subject matter.
However, the USPTO has recently signaled a shift towards a more inclusive approach to software and AI-related patents. In November 2025, the USPTO issued new guidelines emphasizing that AI systems are to be considered tools that assist human inventors in the creative process. The guidance reaffirms that under U.S. law, AI cannot be designated as an inventor. However, the USPTO has moved away from a prior approach used during the Biden administration that applied joint-inventorship criteria to AI-related inventions. Instead, it asserts that the standard for determining inventorship remains the same regardless of AI involvement, meaning there is no special standard for AI-assisted innovations. US Patent Office issues new guidelines for AI-assisted inventions
Further reinforcing this perspective, the USPTO’s Appeals Review Panel issued a precedential decision in Ex parte Desjardins in September 2025. This decision vacated a previous rejection under § 101, highlighting that improvements to the functioning of machine learning models can constitute practical applications under the governing subject matter eligibility framework. The decision emphasized that eligibility must be assessed carefully and consistently, with proper consideration of technological improvements reflected in the claims and specification.
Despite these developments, challenges persist. The Electronic Frontier Foundation (EFF) has raised concerns about proposed USPTO rules that could limit the public’s ability to challenge improperly granted patents. The EFF argues that these rules would make inter partes review (IPR)—a key tool for correcting patent office errors—unavailable in many critical situations, potentially allowing questionable patents to remain unchallenged. EFF Tells Patent Office: Don’t Cut the Public Out of Patent Review
In the judicial arena, the Federal Circuit’s decision in Recentive Analytics v. Fox Corp. clarified that merely applying machine learning to a specific field does not render an invention patentable unless it improves the ML model itself or enhances a computer’s functionality. This ruling underscores the necessity for patent applications to demonstrate specific technological improvements rather than abstract applications of AI.
As the USPTO continues to refine its guidelines and courts provide further clarity, legal professionals must stay vigilant. The evolving standards necessitate a nuanced understanding of what constitutes patent-eligible subject matter in the software and AI domains. Ensuring that patent applications clearly articulate technological advancements and practical applications will be crucial in navigating this shifting terrain.