Not A Categorical Ban: Federal Circuit Narrowed Spectrum of Patent Eligible Machine Learning Claims

Recentive Analytics, Inc. v. Fox Corp., No. 23-2437 (Fed. Cir. 2025) – On April 18, 2025, the Federal Circuit upheld the district court’s dismissal of the case on the ground that the patents were ineligible under § 101. 

Background

Recentive owns four patents related to television broadcast program scheduling optimization (US Patent Nos. 11,386,367; 11,537,960; 10,911,811; and 10,958,957). Belonging to two different families, all four patents nonetheless rely on machine learning techniques to carry out the claimed methods. As described by the specification of one of the patents, the claimed inventions could utilize “any suitable machine learning technology” including “a gradient boosted random forest, a regression, a neural network, a decision tree, a support vector machine, a Bayesian network, [or] other type of technique.” 

Recentive asserted all four patents against Fox in the U.S. District Court for the District of Delaware. Fox moved to dismiss the case for failure to state a claim. Fox argued that all asserted claims were invalid because they did not recite patent eligible subject matter under 35 U.S.C. § 101. Applying the two-step inquiry of Alice, the district court found the asserted claims unpatentable. Accordingly, the district court dismissed the case. Recentive appealed.

Issue

Whether claims that do no more than apply established methods of machine learning to a new data environment are patent eligible.

Holdings and Reasoning

The Federal Circuit answered no. Reviewing the district court’s dismissal de novo, the Federal Circuit conducted its own Alice analysis. 

At Alice step one, the Federal Circuit held the claims were directed towards the abstract ideas of applying generic machine learning techniques. The Federal Circuit first observed from the record that Recentive had conceded the claims recited broadcast scheduling concepts that were performed by human beings and existed even prior to computers. Recentive argued it was the application of machine learning to carry out these concepts that made the claims patent eligible. However, the Federal Circuit found that the claims merely applied conventional machine learning techniques, with no improvement to the training models. Recentive then argued that the claims were directed to the application of machine learning techniques to a new field of use and that the application speeds up human activity itself and represents a technical improvement. Relying upon several precedents, the Federal Circuit rejected this argument. 

At Alice step two, the Federal Circuit highlighted that Recentive’s argument overlapped with its argument for step one and “plainly fail[ed] to identify anything in the claims that would ‘transform’ the claimed abstract idea into a patent-eligible application.”

In the end, recognizing machine learning as an important field, the Federal Circuit caveated that its decision should not be read as a categorical ban on machine learning claims. Rather—“[t]oday, we hold 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.”