Recentive Analytics v. Fox: The Federal Circuit Provides Analysis on the Patent Eligibility of Machine Learning Claims
On April 18, 2025, the Federal Circuit remained consistent with previous Alice decisions by holding that four machine learning patents involved in a dispute between Recentive Analytics, Inc. and Fox Corp. were ineligible under 35 U.S.C. § 101 because they “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.” With this ruling, the Federal Circuit continues to warn patent practitioners and Patent Owners that the mere application of generic machine learning models to a new field or environment will be insufficient to clear the subject matter eligibility hurdle.
Recentive’s patents fell broadly into two families:
- Family #1: Machine Learning Training Patents (U.S. Patent Nos. 11,386,367 and 11,537,960) were directed towards systems and methods for determining event schedules by dynamically generating optimized event schedules for live events using machine learning models trained on historical data.
- Family #2: Network Map Patents (U.S. Patent Nos. 10,911,811, and 10,958,957) were directed towards automatically and dynamically generating “network maps” that determine which programs are displayed on which channels according to geographic markets and timings.
The Recentive patents were asserted in an infringement suit against Fox Corp., which moved to dismiss Recentive’s infringement suit on the grounds that the patents claimed ineligible subject matter under 35 U.S.C. § 101. The district court concluded that the patents were directed to patent-ineligible subject matter.
On appeal, while the Federal Circuit acknowledged the importance of machine learning and the fact that machine learning “may lead to patent-eligible improvements in technology,” it nevertheless held that Recentive’s claims did not meet that standard. Specifically, Recentive’s claims were directed to no more than applying established machine learning techniques to a new data environment, without reciting improvements to machine learning technology.
Under Alice step one, the Federal Circuit found that Recentive’s claims were directed towards an abstract idea that relied on the use of generic machine learning technology, for at least the following reasons:
- The Federal Circuit held that Recentive’s claims, at their core, were directed to activities that humans have long performed “before the introduction of machine learning”; for example, event planners had already looked at information “such as prior ticket sales, weather forecasts, and other data to determine when and where to schedule a particular event”, and had manually created network maps to determine what content would be displayed on which channel at a certain time. Further, the Federal Circuit reiterated that merely performing a task previously undertaken by humans with greater speed and efficiency by using existing machine learning technology, will not render the claims patent-eligible. As Mintz has noted previously, courts consistently hold that claims reciting a process that could theoretically be performed by a human are patent-ineligible, even if the human-performed process would be inefficient and even when the claimed invention seeks to avoid that inefficiency.
- The Federal Circuit further criticized the claims for “not delineat[ing] steps through which the machine learning technology achieves an improvement” over existing technology; for example, while Recentive argued that its claims were directed to improving algorithms for generating event schedules and network maps, or for making machine learning better, the Federal Circuit noted that the claims “do not claim a specific method for ‘improving the mathematical algorithm or making machine learning better.’” Instead, Recentive’s claims merely “functionally” describe “a mere concept without disclosing how to implement that concept.”
- Recentive’s arguments that the claims were directed to patent-eligible subject matter because they applied machine learning technique to a new field of use (i.e., event scheduling, creation of network maps) were rejected. The Federal Circuit reiterated that “an abstract idea does not become nonabstract by limiting the environment” and held that “Recentive’s argument that its patents are eligible simply because they introduce machine learning techniques to the fields of event planning and creating network maps directly conflicts with our § 101 jurisprudence.”
- The Federal Circuit was also unconvinced that training a machine learning model made the model specific or tailored, holding (as other courts have previously held) that the “requirements that the machine learning model be ‘iteratively trained’ or dynamically adjusted in the … patents do not represent a technological improvement… [and] are incident to the very nature of machine learning.”
Under Alice step two, the Federal Circuit found that Recentive’s claims merely recited the abstract idea itself, i.e., of using a machine learning model to generate optimized maps and schedules. “[N]othing in the claims, whether considered individually or in their ordered combination, that would transform the Machine Learning Training and Network map patents into something ‘significantly more’ than the abstract idea of generating event schedules and network maps through the application of machine learning.” Accordingly, the claims were directed to patent-ineligible subject matter
At Mintz, we continue to follow the Federal Circuit and the Patent Office’s developing body of law surrounding subject matter eligibility of patents that focus on machine-learning-based techniques, as we develop robust patent prosecution and litigation strategies for machine learning technologies.
