In a precedential decision implicating patent filings related to machine learning and artificial intelligence processes, the Federal Circuit in Recentive Analytics, Inc. v. Fox Corp., No. 2023-2437 (Fed. Cir. Apr. 18, 2025), held that an application of a generic machine learning process to a new data environment is insufficient to confer patent eligibility absent a technical innovation in the underlying machine learning process. While the Recentive decision represents the Federal Circuit’s first foray into the subject-matter eligibility of machine learning innovations, the Recentive decision is consistent with the Federal Circuit’s existing § 101 jurisprudence for software and computer-implemented inventions.
The Recentive decision
Recentive Analytics, Inc. filed a suit against Fox Corp. in the District of Delaware, asserting patent infringement of four patents related to the use of machine learning to generate network maps and optimize scheduling of TV broadcasts and live events. Fox moved to dismiss for failure to state a claim on the grounds that the patents were ineligible under § 101. Recentive argued that the claims of the patents were indeed patent eligible, as they recited a unique application of a machine learning process to the generation of network maps and the scheduling of live events and broadcast programming. The District Court granted Fox’s motion, and on appeal, the Federal Circuit affirmed.
In the Recentive decision, the Federal Circuit held that “claims that do no more than apply established methods of machine learning to a new data environment” are not patent eligible. Recentive, slip op. at 10. In reaching its decision, the Federal Circuit emphasized that:
- The claimed invention relied on generic and conventional machine learning technology and generic computers and processors.
- Iterative training or dynamic adjustments are not a technological improvement because they are within the nature of machine learning.
- The claims fail to disclose how the machine learning technology achieves an improvement.
- Limiting machine learning to a particular field of use or technological environment, or applying existing technology to a new database, does not make it patent eligible.
- Increasing the speed and efficiency of a process through the use of machine learning technology without improving the underlying machine learning process does not confer patent eligibility.
Takeaways from the Recentive decision
The Recentive decision emphasizes that the mere addition of a known or generic machine learning or artificial intelligence process to even a new technical environment is alone insufficient to confer eligibility under 35 U.S.C. § 101 absent an improvement to the underlying machine learning technology. To survive § 101 challenges after Recentive, applicants should consider disclosing and actively claiming technical improvements to the claimed machine learning processes, not simply increases in computational speed or efficiency directed from the machine learning process or new use cases for known machine learning processes.
Further, while the Federal Circuit indicated that “improvements to machine learning models” could confer patent eligibility under § 101, questions remain as to what qualifies as an improvement in the context of machine learning processes. The metes and bounds of these improvements will be defined as the US Patent and Trademark Office, the district courts, and the Federal Circuit digest and interpret the Recentive decision.