AI Meets Section 101: “Have AI Do It” is Not an Invention

 The headlines — and often our own experience — tell us AI is everywhere.  It promises to transform industries, automate decisions, and make predictions that seem almost magical.  It was inevitable that AI and Section 101 of the Patent Act would meet.  And, in the recent case of Recentive Analytics, Inc. v. Fox Corp., 2023-2437 (Fed. Cir. Apr. 18, 2025), they did.

Section 101

Section 35 U.S.C. 101 is the gatekeeper of the U.S. patent system. It says you can patent a process, machine, manufacture, or composition of matter, but not an “abstract idea,” a “law of nature,” or a “natural phenomenon.” Sounds simple, right?  After all, you don’t want someone patent an abstract idea of law of nature, so other inventors can’t use them.  That could bring innovation to a halt in some industries.

Section 101 has become a leading basis for finding patents invalid, following the Supreme Court’s decision in Alice Corp. v. CLS Bank International 573 U.S. 208 (2014). Alice says, to decide if an invention passes muster under Section 101, the court is to first examine whether a claimed invention is directed to a patent-ineligible concept, such as an abstract idea.  If not, it passes the section 101 hurdle.  If it fails the first part of the test, the court must next assess whether the claim includes an "inventive concept"—an element or combination of elements that transforms the claim into a patent-eligible application. Thus, a sufficiently inventive concept can render a claimed invention patent-eligible even if the first step suggests is not eligible.  Got all that?  Simple, right?

Not exactly.  The Alice test has been sharply criticized, even by members of the Federal Circuit.  For example, one judge described it in a dissent as “useless,” “an abstract,” “unworkable,” an “intellectual morass,” and “incoherent,” among other things. See Interval Licensing, LLC v. AOL, Inc., No. 16-2502 (Fed. Cir. 2018).

Not surprisingly, use of the Alice test often leads to differing result in the courts.  Still, the courts continue to apply Section 101 under the existing test. 

AI: The Double-Edged Sword of Innovation

AI promises to revolutionize everything from sports broadcasting to medicine. But there’s a catch: a lot of AI breakthroughs are about using smart algorithms to analyze data and make decisions.  These are things humans have always done.  AI just does it faster and at scale. So, when someone tries to patent an AI solution, the big question is this: is this a genuine technical leap, or just an abstract idea dressed up in code?

The Recentive Case: Machine Learning Meets Section 101 in the Courtroom

That brings us to Recentive Analytics v. Fox Corp.  Recentive had  patents for using machine learning to optimize TV broadcast schedules and network maps. TV broadcasters have always had to do this.  But the promise of using AI to complete this task faster, cheaper and more accurately sounds pretty cutting-edge, right?  Yet, when Recentive took Fox to court, claiming patent infringement, the judges weren’t so impressed.

The Federal Circuit looked at the patents and said, essentially, “These claims are just using generic machine learning techniques to solve a scheduling problem. That’s an abstract idea, not a patentable invention.” The court pointed out that the patents didn’t describe any new kind of machine learning model or technical improvement — just the idea of applying existing AI methods to a familiar problem.

How the Court Decided the Case

The court’s reasoning on the two part Alice test boiled down to this:

              1.          Is it an abstract idea? The judges said yes. It’s basically just decision-making —  something people have always done — but now with a computer.

              2.          Is there an inventive concept? The judges said no. The patents didn’t introduce any new technology or clever twist; they just used off-the-shelf AI tools in a specific context. Basically, it was a matter of having AI do something humans already did. That’s not enough for a patent, even if AI could do it better.

What Does It All Mean?

This decision isn’t all that surprising in light of Section 101’s prominent role in patent law these days.  But it is still a big deal in the AI world. The Federal Circuit has now held that courts must look for more than just smart uses of AI.  They must see real technical innovation. If your invention is just applying machine learning to a new kind of data, you will have a tough time convincing a judge it deserves a patent.

So, as AI keeps marching forward, the legal world continues to figure out how to keep up. Recentive was a critical meeting between AI and Section 101 at the appellate level. And it establishes that inventors are going to have to up their game to get patents on AI technology.  It’s not enough to claim an invention to say, in essence,  “Have AI do it.”

Next
Next

The College of Commercial Arbitrators Supports the Rule of Law