The Application of Generic Machine Learning to New Data Environments Requires “Something More” to be Patent Eligible

On April 18, 2025, the Federal Circuit issued an opinion in Recentive Analytics, Inc. v. Fox Corp. addressing for the first time whether patents that claim no more than the application of generic machine learning to a new data environment are patent eligible under 35 U.S.C. § 101.¹ The Federal Circuit held that they are not, absent an inventive concept that transforms the claims into something more than an abstract idea.²

Recentive Analytics, Inc. (“Recentive”) is the owner of the four disputed patents, which seek to optimize the scheduling of live events and to generate network maps for broadcasters.³ The patents fall into two groups: the “Machine Learning Training” patents⁴ and the “Network Map” patents.⁵ The two Machine Learning Training patents share a specification that teaches that the machine learning model can be trained using a set of data, including historical data from prior events, venue locations, and ticket sales, and may be trained to optimize schedules based on target features that the user inputs.⁶ The specification also states that the patented method can use any suitable machine learning technique.⁷ The two Network Patents also share a specification and use training data, such as weather data, news data, or gambling data, in conjunction with a machine learning model to generate optimized network maps.⁸ Like the Machine Learning Training patents, the specification for the Network Map patents clarifies that the method uses any suitable machine learning technique, and that users can input target features to achieve a desired result.⁹

Recentive filed a patent infringement suit against Fox Corp., Fox Broadcasting Company, LLC, and Fox Sports Productions, LLC (collectively, “Fox”) in the District Court for the District of Delaware on November 29, 2022.¹⁰ Fox moved to dismiss the complaint for failure to state a claim and the district court granted Fox’s motion, finding that the patents were ineligible under the Supreme Court’s two-step analysis set forth in Alice Corp. v. CLS Bank Int’l.¹¹ Under Alice, courts determine patent eligibility under § 101 by first determining whether the claims are directed to a patent-ineligible abstract idea.¹² If the claims are found to be directed to ineligible subject matter, the court then determines whether the claims possess an inventive concept that is sufficient to ensure that the patent amounts to significantly more than a patent upon the ineligible concept itself.¹³ The district court found that the asserted claims of the Machine Learning Training and Network Map patents were directed to abstract ideas “using known generic mathematical techniques.”¹⁴ The district court also found that because the machine learning limitations were broad, well-known techniques and claimed generic computing devices, the claims were not directed to an inventive concept and therefore failed Alice step two.¹⁵ The district court denied Recentive’s request for leave to amend, finding that any amendment would have been futile, and Recentive appealed.¹⁶

On appeal, a panel of three judges reviewed Recentive’s challenges to the district court’s dismissal of the complaint and the determination of patent eligibility de novo.¹⁷ Under the first step of Alice, the Federal Circuit found that it was clear that the disputed claims are directed to ineligible subject matter because the claims do not represent a technological improvement, holding that the district court correctly concluded that the Machine Learning Training and Network Map patents were directed to abstract ideas.¹⁸ Recentive argued that the application of machine learning is not generic because the algorithms are functionally dynamic and can automatically customize the maps and schedules.¹⁹ The court rejected this argument and instead emphasized that the claims and specifications do not describe how the improvement was accomplished, stating that “the claims do not delineate steps through which the machine learning technology achieves an improvement.”²⁰ The court further rejected Recentive’s argument that the patents are patent eligible “simply because they introduce machine learning techniques to the fields of event planning and creating network maps” because this rationale “directly conflicts with our § 101 jurisprudence.”²¹ The court cited SAP Am., Inc. v. InvestPic, LLC, which held that patents that disclose the application of existing technology to a novel context may be directed to abstract ideas, and concluded that the same reasoning applies in the machine learning context.²² Lastly, the court found that the fact that the claimed methods perform a task with greater speed and efficiency than humans does not render them patent eligible.²³ The court noted that it has consistently held that “the increased speed and efficiency resulting from use of computers (with no improved computer techniques) do not themselves create eligibility,” regardless of whether the issue is raised at Alice step one or step two. ²⁴

The Federal Circuit additionally affirmed the district court’s conclusion that Recentive’s claims do not include an inventive concept, and therefore they fail to satisfy step two of Alice.²⁵ Recentive argued that the inventive concept is using machine learning to optimize the maps and schedules and update them based on real-time information, but the court found that Recentive failed “to identify anything in the claims that would ‘transform’ the claimed abstract idea into a patent-eligible application.”²⁶ The court concluded that since the claims do not “transform the Machine Learning Training and Network Map patents into something ‘significantly more’” than an abstract idea, Fox’s motion to dismiss was properly granted.²⁷ The Federal Circuit also affirmed the district court’s decision to deny Recentive’s request for leave to amend in light of its holding with respect to the ineligibility of Recentive’s patents.²⁸

The Federal Circuit’s holding emphasizes the importance of including an inventive concept in future applications of machine learning if an inventor wishes to obtain or enforce a patent. As the Federal Circuit noted, “[m]achine learning is a burgeoning and increasingly important field” that is likely to lead to improvements in technology.²⁹ However, the Recentive v. Fox Corp. decision reiterates that there must be an improvement to the machine learning model, otherwise the claimed application may be a patent ineligible abstract idea. As machine learning technology advances, clients, industry professionals, and practitioners are encouraged to follow the latest developments surrounding the intersection of artificial intelligence and intellectual property law.