Evolving Law for AI and Software Patents

At a Glance

• In Fintiv, Inc. v. PayPal Holdings, Inc., the Federal Circuit held that the term “payment-handler” invoked a 35 U.S.C. Section 112 paragraph 6 (pre-AIA) interpretation and is indefinite. As such, more detailed descriptions on various terms including algorithms are recommended.
• In Recentive Analytics, Inc. v. Fox Corp., the Federal Circuit held that “using machine learning to dynamically generate optimized maps and schedules based on real-time data . . . is no more than claiming the abstract idea itself.” Claims that include the use of machine learning could benefit from the description of implementation details in the use of machine learning models in the specification.

Patent law for artificial intelligence (AI) and software-related patents is evolving fast. For example, the U.S. Court of Appeals for the Federal Circuit recognized that “[m]achine learning is a burgeoning and increasingly important field and may lead to patent-eligible improvements in technology.” Issues such as the invocation of means-plus-function and the application of subject matter patentability to AI (e.g., machine learning, deep learning, generative AI) and software patents are among the recent Federal Circuit decisions. 

Fintiv, Inc. v. PayPal Holdings, Inc.

In Fintiv, Inc. v. PayPal Holdings, Inc. (Fed Cir. 2025), the appellate court held that the “payment-handler” invoked a 112 ¶ 6 (pre-AIA, corresponding to AIA 112(f)) interpretation and the claims are indefinite under 112(b) because the “specification fails to disclose adequate corresponding structure to perform claimed function.” Fintiv argued that the patents disclose a two-step algorithm: “(1) ‘wrap[s] APIs of different payment processors, such as, for example banks ...’ and (2) ‘exposes a common API to facilitate interactions with many different kinds of payment processors.’” 

The Federal Circuit rejected this argument, finding that “reciting the function of the payment-handler terms using generic terms without providing any details about an algorithm to carry out the functions of using APIs of different payment processors” does not constitute an algorithm. 

As such, more detailed descriptions on various terms including algorithms are recommended, including specific example implementation, to avoid such an outcome.

Recentive Analytics, Inc. v. Fox Corp.

In Recentive Analytics, Inc. v. Fox Corp. (Fed. Circ. 2025), the Federal Circuit held that claims towards training machine learning models to optimize event schedules, and claims towards the creation of network maps using training data and machine learning for broadcasters were ineligible under 35 U.S.C. Section 101. 

The court concluded that the disputed claims are directed to abstract ideas under the Alice test. As an initial matter, Recentive “asserted that its patents claim eligible subject matter because they involve ‘the unique application of machine learning to generate customized algorithms, based on training the machine learning model.” In the process of making their conclusion, the court concluded that “the requirements that the machine learning model be ‘iteratively trained’ or dynamically adjusted in the Machine Learning Training patents do not represent a technological improvement.” Iterative training and dynamic adjustment of machine learning models was held to be “incident to the very nature of machine learning.” 

It was held that even though the claims recite the use of machine learning, they simply “perform[ed] a task previously undertaken by humans.” Further, the “only thing the claims disclose about the use of machine learning is that [it] is used in a new environment” but they fail to “delineate steps through which the machine learning technology achieves an improvement.” Under step two of Alice, the Federal Circuit held that “using machine learning to dynamically generate optimized maps and schedules based on real-time data . . . is no more than claiming the abstract idea itself.”

Claims that include the use of machine learning could benefit from the description of implementation details in the use of machine learning models in the specification.

In Closing

The patent law for AI and software-related patents is still evolving. The Recentive decision is one of the first AI decisions of the Federal Circuit on 35 U.S.C. Section 101.