2025 Federal Circuit IP Appeals: Recentive Analytics, Inc. v. Fox Corp., 134 F.4th 1205 (Fed. Cir. 2025) (Dyk, Prost, Goldberg)

Recentive held that the application of machine learning to new tasks and contexts cannot render an invention patent-eligible under 35 U.S.C. § 101.

Recentive’s claims recited iteratively training a machine-learning model to identify relationships between different event parameters and desired outcomes for the events, generating a schedule for future events based on user-defined event parameters and desired outcomes, and updating the generated schedule based on real-time changes to event parameters. In other words, the patents claimed various machine-learning-generated network maps and schedules for TV broadcasts and live events.

The district court dismissed Recentive’s infringement case against Fox, holding that the patents at issue were directed to ineligible subject matter under the two-step Alice inquiry. Under step one, the court found that the claims were “directed to the abstract ideas of producing network maps and event schedules, respectively, using known generic mathematical techniques.” Under step two, the court found no “inventive concept” because the machine learning techniques described were “broad, functionally-described, well-known techniques” that did not amount to significantly more than a patent on the abstract concept itself.

The Federal Circuit affirmed. Recentive had argued that its claims were patent-eligible because they involved a “unique application of machine learning to generate customized algorithms, based on training the machine learning model, that can then be used to automatically create … event schedules that are updated in real-time.” The Federal Circuit rejected this argument, concluding that iterative training using updated data is “incident to the very nature of machine learning” and “do[es] not … mak[e] machine learning better.” The court explained that the claimed systems and methods simply applied conventional machine-learning techniques to a new data environment.

Under Recentive, the process of using artificial intelligence (AI) to perform tasks previously done manually or using more basic computer processing (like sorting through large amounts of sequencing data to find patterns or using a generic large language model to generate a compound that binds to a target of interest) is unpatentable. But the court stated that patent protection should remain available for inventions that improve AI itself, such as machine learning tools that provide a new technological advantage. Recentive held “only that patents that do no more than claim the application of generic machine learning to new data environments, without disclosing improvements to the machine learning models to be applied, are patent ineligible under § 101.”

Recentive carries forward several § 101 jurisprudential concepts into the AI space. In particular, field‑of‑use limitations (for example, using AI in a new data environment) do not transform an abstract idea into a patentable improvement. And increased speed and efficiency relative to human performance of a task are also not sufficient for eligibility. The decision thus shows that the court views AI and machine learning similarly to general-purpose computing for purposes of patentability.

Going forward, patent applications for AI innovations would be well served to include robust disclosure of how a claimed invention provides a technical solution to a technical problem (for example, a novel training process or novel data preprocessing technique). Surface-level claiming of “iteratively training” models to improve accuracy will likely not be sufficient.

This article appeared in the 2025 Federal Circuit IP Appeals: Summaries of Key 2025 Decisions report.