Partner
Here’s What Happened:
We know that courts are grappling with whether training of large language models (LLMs) results in copyright infringement. But what about patents?
Recentive Analytics Inc. obtained patents for two types of machine learning models in the entertainment industry. The patents involve a method of scheduling live events and refining “network maps” for broadcasting events. The patents employ machine learning techniques using generic computing machines.
Fox Corp. and others also used machine learning models to schedule live and broadcast events. Recentive sued them for patent infringement.
Fox moved to dismiss the complaint. Fox argued that the patents were not eligible for patent protection. Recentive admitted that the concept of networking maps has been around a long time. Recentive also admitted that its patents to not claim a machine learning technique. But, argued Recentive, their machine learning models generate custom algorithms which are new in the field of live broadcast events and event schedules.
The court disagreed with Recentive and granted Fox’s motion to dismiss. The court used the two step analysis established in Alice Corp. v. CLS Bank International Supreme Court decision is 573 U.S. 208 (2014).
- Step 1: The court determined that the claims in Recentive’s patents are “directed to the abstract ideas of producing network maps and event schedules, respectively, using known generic mathematical techniques.” Abstract ideas are not subject matters for patent protection.
- Step 2: The court determined that the machine learning models were broad and well known techniques that could be used on generic and conventional computing devices. So the models were not sufficiently “new” for patent protection.
In granting the motion to dismiss, the court also rejected Recentive’s request for leave to amend the complaint. The court did not believe Recentive could come up with a set of facts that would change the two factor analysis. The court was careful to limit the decision to machine learning models that use generic computing devices.
WHY YOU SHOULD KNOW THIS: Since the Alice decision, it is very hard to obtain a patent for an abstract idea that implemented without something new. Recentive overcame that hurdle in the patent prosecution phase and obtained the patents. But the patents did not hold up when they were challenged by an alleged infringer. Given the court’s two step analysis, as far as machine learning models go, patent protection will be hard to obtain without further innovation.
Cited Authority: Recentive Analytics, Inc. v. Fox Corp. et al., Case No. 23-2437 (Fed. Cir. Apr. 18, 2025) (Dyk, Prost, Goldberg, JJ.)