Tag Archives: Abstract Ideas

abstract-idea exception, software and business-method eligibility

Hurdling Eligibility with a “Separable Interface”

by Dennis Crouch

Amazon has asked the Federal Circuit to rehear Ceiva Opco, LLC v. Amazon.com, Inc., No. 2024-1721 (Fed. Cir. July 23, 2026) (nonprecedential), a Section 101 case about internet-connected digital picture frames.

The panel affirmed summary judgment of ineligibility as to representative claims of three patents (US9203930, US9654562, and US9124656), but reversed as to claim 19 of US6442573, finding that particular claim was not directed to an abstract idea.

The feature separating claim 19 from the rest was a single limitation: a user interface "physically separable" from the digital picture frame. The panel read that limitation as capturing a technological advance, remote user access to a picture frame, rather than a bare desired result.

Ceiva conceded below that it did not invent the interface technology. But, of course, Alice step one does not ask whether a claim is novel, but rather whether it is directed to an abstract idea. Conventionality (novelty) is a question for step two.

 

Amazon's rehearing petition, filed August 24 by folks at Fenwick & West, argues that the limitation cannot carry the weight assigned by the panel.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Refilling the Pipeline: Two New Section 101 Petitions Heading to the Supreme Court

by Dennis Crouch

Two more Section 101 petitions are taking shape at the Supreme Court. Over the past two weeks, two patentees that lost on summary judgment of ineligibility, then lost again in nonprecedential Federal Circuit affirmances, filed applications to extend their certiorari deadlines: Innovaport LLC v. Target Corp., No. 26A8, and Oasis Tooling, Inc. v. Siemens Industry Software Inc., No. 26A34. The Chief Justice has already granted Innovaport's application, setting a September 4, 2026 petition deadline; Oasis Tooling's request would move its deadline to October 12. Both filings arrive on the heels of the Court's May 18 denial in United Services Automobile Ass'n v. PNC Bank N.A., No. 25-853.

  • Innovaport is likely to target the disconnect between USPTO examination and Article III adjudication: some of its patents overcame Section 101 rejections during prosecution, only to be invalidated by courts that owe the agency no deference.
  • Oasis Tooling is likely to focus on the mental process doctrine, asking how a computerized process can be dismissed as one that "can be performed in the human mind" when the court concedes it could not actually be performed in the human mind.

A third application involving 101 issues is also still in the queue, US Patent No. 7,679,637 LLC v. Google LLC, No. 25A1262. All other eligibility petitions filed this past year have been denied certiorari.

By my count, 89 petitions for writ of certiorari raising patent eligibility issues have been filed with the Supreme Court since the fall of 2017, where my tracking data begins. None of those petitions have been granted.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Pairing Down: Federal Circuit Affirms § 101 Dismissal of Vehicle-Mode Notification Patent

The Federal Circuit's nonprecedential decision in TJTM Technologies, LLC v. Google LLC, No. 25-1218 (Fed. Cir. May 5, 2026), affirms a N.D.Cal. R.12(b)(6) dismissal of TJTM's infringement suit on § 101 grounds. The patent at issue, U.S. Patent No. 8,958,853, claims a mobile device that automatically enters an "inactive mode" upon pairing with a vehicle, suppressing call and text notifications and transmitting an automated away message to senders. TJTM accused certain Google Android distracted-driving features of infringement. Judge Trina Thompson granted Google's motion to dismiss after applying the two-step framework of Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014), and a Federal Circuit panel of Judges Dyk, Chen, and Stark (with Judge Chen writing) affirmed in a brief opinion.

Representative claim 1 recites a generic mobile device with three structural components (a wireless communication module, a processor, and a memory) followed by a sequence of programmatic steps: providing a graphical user interface through which a user customizes inactive-mode functions, receiving a user selection to auto-initiate inactive mode upon vehicle pairing, receiving the away-message text, automatically initiating inactive mode in response to pairing, and upon an incoming communication transmitting the away message while suppressing the audible, visual, or vibration cues. The claim says nothing about how the Bluetooth pairing handshake works, how the OS-level interrupt suppression is accomplished, or how the away message is generated and routed. The Federal Circuit found this combination of generic components and conventional steps insufficient to clear either Alice prong.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Guest Post: Improving PERA’s Eligibility Exclusions

Guest Post by Professors Jeffrey A. Lefstin, UC College of the Law, San Francisco & Peter S. Menell, UC Berkeley Law

As Dennis reported last year, Senators Thom Tillis and Chris Coons re-introduced the Patent Eligibility Reform Act (PERA) in the Senate in May 2025, accompanied by a parallel bill in the House. Despite significant pushback on earlier versions of the bill, PERA 2025 changes only one aspect of the proposed new text for § 101: human genes that are merely “isolated,” but neither purified, enriched, altered by human activity, nor employed in a useful invention or discovery, will remain ineligible. That revision is essentially illusory, since vanishingly few useful claims will fail to meet the latter criteria.

The criticisms of our current patent eligibility regime are familiar. Jurisprudentially, the Mayo/Alice regime was founded on serious misunderstandings of both the historical record and of the intent behind the 1952 Patent Act. The Supreme Court’s reshaping of patent eligibility was based largely on its economic hunches, rather than the careful study such a foundational doctrine deserves. The Federal Circuit’s ‘abstract idea’ jurisprudence denies patentability to specific and inventive advances in data processing without ever justifying why it should do so, and often depends on distinctions so fine they are unintelligible to most observers. In biomedicine, between the decisions of the Supreme Court and the Federal Circuit, the United States has almost unintentionally backed into a regime where nearly all methods of diagnosis are ineligible, but nearly all methods of treatment are eligible — again with no economic justification for why that should be the case.

We agree with PERA’s ambition to write on a clean slate, and to separate patent eligibility under § 101 from questions such as claim scope and inventiveness that the drafters of the 1952 Act intended to be resolved under §§ 102, 103, and 112. But PERA’s approach to subject matter exclusions has serious flaws, as we and many others have pointed out. PERA excludes from patentability “[a] process that is substantially economic, financial, business, social, cultural, or artistic,” without suggesting how to determine whether a process is within those categories. More seriously, PERA then neuters that exclusion by exempting any process that “cannot practically be performed without the use of a machine or manufacture.” Under this language, any invention that requires the use of any human artifact — such as pencil or paper — is patent eligible. PERA would meaningfully exclude from the patent system only mental processes, and pure motions of the human body such as dance or sport moves. When it comes to the biomedical space, PERA’s exclusion of human genes and natural materials is redundant — but the real issue is that PERA makes no attempt to address concerns that patents on genes and other natural materials could encumber research.

(more…)

The Razor Returns: AIPLA Tells the Supreme Court That Alice Step Two Has Revived the Pre-1952 ‘Invention’ Requirement

by Dennis Crouch

The American Intellectual Property Law Association has filed an amicus brief urging the Supreme Court to grant certiorari in United Services Automobile Association v. PNC Bank N.A., No. 25-853 (U.S. filed Mar. 2, 2026). The AIPLA brief does not focus on the particular merits of USAA's mobile check deposit patents. Instead, it frames the case as an opportunity for the Court to correct what the brief calls a historical regression: the Federal Circuit's application of the Alice/Mayo eligibility framework has revived the same subjective "inventive concept" inquiry that the Patent Act of 1952 was specifically designed to eliminate. The brief, filed by Barbara Fiacco and Valerie Orellana of Foley Hoag on behalf of AIPLA, traces this argument through more than a century of patent law, from the commingled eligibility/obviousness/definiteness standards of the 19th century through the 1952 reforms and back to the present "confusion."

USAA's petition presents two questions:

  • Whether the Federal Circuit has wrongly extended the "abstract idea" prohibition to cover concrete technological processes; and
  • Whether it has improperly held that computer-implemented inventions are patent-eligible only if they improve the computer's own functionality.

In 2025, the Federal Circuit reversed a $218 million jury verdict for USAA, holding that its claims directed to mobile check deposit using a handheld device were ineligible. The Supreme Court has signaled some interest by issuing a "Response Requested" and PNC's response is now due April 8, 2026, following an extension. See Dennis Crouch, USAA Petitions SCOTUS: Is Mobile Check Deposit Just an "Abstract Idea"?, Patently-O (Jan. 19, 2026).


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Tie Goes to the Runner? Three Months of SMED Practice at the USPTO

by Dennis Crouch

In baseball, there is a folk rule on force-outs that "tie goes to the runner." The idea is straightforward: if the ball and the runner arrive at the base simultaneously, the runner is safe. In fact, the Official Rules of Baseball do not admit that a tie is even possible, but rather the question is simply whether the fielder tagged the base before the runner arrived. But the folk rule persists because it reflects an instinct about how close calls should break: and in baseball it is the fielder's duty to force the out.

Director John Squires has brought a version of this thinking to patent examination. Since taking office in September 2025, Squires has repeatedly signaled that close calls on patent eligibility should favor the applicant. The December 4, 2025 memoranda on Subject Matter Eligibility Declarations (SMEDs) formalized the invitation: applicants facing Section 101 rejections should submit evidentiary declarations under 37 C.F.R. § 1.132, and examiners should treat that evidence seriously when evaluating eligibility under the preponderance-of-the-evidence standard. Dennis Crouch, Subject Matter Eligibility Declarations (SMEDs) to Overcome Eligibility Rejections, Patently-O (Dec. 5, 2025). The theory is that a well-drafted SMED, supported by concrete technical evidence, should create enough disputed facts to tip the balance in the applicant's favor. But the question remains whether examiners on the ground will agree.

After looking through several hundred R. 132 declarations, I eventually found seven SMEDs that specifically target Section 101 eligibility. The sample is small but instructive. These seven declarations span a range of technologies: data center infrastructure automation, financial analytics, robotic process automation, cybersecurity risk modeling, plant genomics, and blockchain systems. The declarants  range from solo inventors with decades of software experience to PhD scientists at venture-backed startups. And the early results are mixed. So far, none have received a notice of allowance and others have been rejected (with the bulk still awaiting response from the examiner).


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

The Recentive Ratchet: RPI’s NLP Patent Falls to the New-Environment Rule

by Dennis Crouch

In Rensselaer Polytechnic Institute v. Amazon.com, Inc., No. 2024-1725 (Fed. Cir. Feb. 24, 2026), a panel led by Judge Dyk affirmed summary judgment invalidating U.S. Patent No. 7,177,798 as ineligible 35 U.S.C. § 101. The '798 patent claims a method for processing natural language inputs using case-based reasoning applied to a metadata database. Rensselaer and its exclusive licensee CF Dynamic Advances had asserted the patent against Amazon's Alexa virtual assistant technology. The district court (N.D.N.Y., Judge Sannes) granted Amazon's motion for summary judgment, finding the claims directed to patent-ineligible subject matter under Alice Corp. v. CLS Bank Int'l, 573 U.S. 208 (2014). The nonprecedential opinion applies the rule from Recentive Analytics, Inc. v. Fox Corp., 134 F.4th 1205 (Fed. Cir. 2025) (cert. denied) that applying a well-established AI technique to a new field does not overcome patent ineligibility. Judge Dyk authored both opinions.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Pre-Alice Patents Keep Falling: Three Section 101 Decisions from the Federal Circuit

by Dennis Crouch

When I see a computer-related patent with a pre-2010 filing date, my first instinct is to check for eligibility problems. Patents drafted before the Supreme Court's 2014 decision in Alice Corp. v. CLS Bank International, 573 U.S. 208 (2014), were written in a different era of patent law where Section 101 was often an absent watchman even for software and business method claims. Specifications from that period routinely described the problem being solved in business terms and the solution in generic technological ones. This made sense under the legal regime that existed at the time. But times have changed and the same specifications now must support claims that need to demonstrate a concrete technical improvement - and that can be a very hard task when the underlying disclosure treats computer components as interchangeable commodities.

Two decisions from the Federal Circuit this past week illustrate the problem. In the precedential GoTV Streaming, LLC v. Netflix, Inc., No. 2024-1669 (Fed. Cir. Feb. 9, 2026), Judge Taranto reversed an jury infringement verdict and held three related patents invalid under Section 101. The patents, with a 2007 priority date, claim methods for tailoring content display specifications to the capabilities of a wireless device. In the nonprecedential Innovaport LLC v. Target Corporation, No. 2024-1545 (Fed. Cir. Feb. 6, 2026), the court affirmed summary judgment invalidating all 55 asserted claims across six related patents claiming priority back to a 1999 application. Those patents cover systems for providing in-store product location information.

Some of the asserted patents were prosecuted after Alice and overcame eligibility rejections during examination. The Federal Circuit found that fact irrelevant, citing its prior holding that courts "are not required to defer to Patent Office determinations as to eligibility." But, in my review, neither prosecution history developed a factual record supporting inventive concept at Alice step two. A more robust prosecution history, one that built up specific factual showings rather than simply amending around the examiner's rejection, might have given the courts something to work with.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Tu Quoque Denied: Google’s Own Patents Irrelevant to §101 Defense

by Dennis Crouch

Google won today's eligibility case with an affirmance that  US7679637 is ineligible.  For me, the following are the two most interesting aspects of the decision:

  • Tu Quoque Rejected: The patent owner argued that  Google's own video conferencing patents use similar "functional claiming" techniques, and therefore Google must believe such claims are patent-eligible. The Federal Circuit rejected this hypocrisy-style argument, holding that the eligibility of Google's patents "is not before us and has no bearing on our analysis."
  • Where Does Functional Claiming Fit? At oral argument, Chief Judge Moore candidly admitted doctrinal uncertainty: "Is this result-oriented functional claiming problem . . . a Step 1 issue or a Step 2 issue? I'll be honest with you, I don't always know where that line is." The opinion resolves this by treating functional claiming as part of the Step 1 "directed to" inquiry.

US Patent No. 7,679,637 LLC v. Google LLC, No. 2024-1520 (Fed. Cir. Jan. 22, 2026).


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

USAA Petitions SCOTUS: Is Mobile Check Deposit Just an “Abstract Idea”?

by Dennis Crouch

I only recently started using mobile check deposits.  I don't get very many checks these days, but when I do it is a real pain to drag myself to the the bank to deposit -- particularly because I don't carry my ATM card with me.  Great convenience -- and also a great patent story.

USAA is a financial services association cooperatively owned by U.S. military members, veterans, and their families. Because USAA serves military families whether stationed at home or deployed overseas to places like Iraq and Afghanistan, it does not operate a traditional branch network. This created a persistent problem: how could USAA members deposit checks when stationed far from any bank? Starting around 2005, a team at USAA began researching technology that would allow members to deposit checks from anywhere in the world using devices they already owned. The challenge was formidable: as of the 2006 priority date, the banking industry conventional wisdom was that specialized check-scanning equipment was the only way to reliably capture check images meeting the technical requirements for remote deposit under the Check 21 Act. USAA's engineers developed a system using a downloaded mobile app to assist customers in capturing compliant check images without specialized equipment, launching Deposit@Mobile in 2009 as the first mobile check deposit service offered by any U.S. bank. Within a month, it became the number one financial app in the Apple app store, with members depositing over 1.5 million checks in the first year.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Ariscale’s Failed Ordered Combination 101 Argument

by Dennis Crouch

In Technology in Ariscale, LLC v. Razer USA Ltd., No. 2024-1657 (Fed. Cir. Jan. 6, 2026), the Federal Circuit affirmed the district court's judgment that claims 1 and 14 of U.S. Patent No. 8,139,652 are invalid under 35 U.S.C. § 101 for claiming patent-ineligible subject matter. The nonprecedential decision, authored by Judge Cunningham, illustrates continuing difficulty patentees face when attempting to frame mathematical signal processing operations as concrete technological improvements. The case also demonstrates how a patent's own specification can undermine eligibility arguments in situations where the disclosed benefits do not map directly to the actual claim language.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Section 101 and Functional Claiming: The Unexplored § 112(f) Path in 10Tales v. TikTok

by Dennis Crouch

I'm watching the pending eligibility appeal in 10Tales, Inc. v. TikTok Inc., No. 2024-1792 (Fed. Cir.).  The appellant is challenging a Northern District of California's grant of judgment on the pleadings finding claim 1 of U.S. Patent No. 8,856,030 directed to the abstract idea of "presenting personalized content to a user based on information about the user." In some ways, the case presents a straightforward application of the Court's targeted-content precedents, but the patent's timing raises interesting questions about how we should evaluate innovation with the benefit of hindsight.

The '030 patent, filed in April 2003, is directed to a system that retrieves information about users from their interactions in online communities and uses that information to select and substitute digital media assets, creating personalized content displays. The district court held that claim 1 fails both steps of the Alice framework: at step one, the claim is directed to the abstract idea of personalizing content; at step two, the claim elements are conventional computer functions that do not supply an inventive concept. On appeal, 10Tales argues that the district court over-generalized the claim at step one and failed to recognize that retrieving "user social network information" from external sources and performing "rule based substitution" of digital assets constituted inventive features at step two.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Empty Formalism: How Patent Eligibility Lost Touch with Preemption

by Dennis Crouch

Earlier this year, the Federal Circuit decided its first "do it on AI" case - holding that Recentive Analytics' inventions were ineligible. The key patents at issue focused on training machine learning models for specific applications (event scheduling and network maps).

Recentive has now petitioned the Supreme Court for writ of certiorari.  Conventional wisdom is that a good certiorari petition needs to show the legal error below and also show why the case is important and timely.  The petition follows this dual framing strategy: (1) a doctrinal claim that the Federal Circuit has abandoned preemption as the touchstone of patent eligibility; and (2) a policy argument tied to what I think of as the "new great game" and what the Trump Administration calls "Winning the AI Race."  The case also arrives at the Supreme Court as the USPTO has begun to move aggressively toward limiting its use of eligibility in patent prosecution.

The petition particularly asks whether eligibility doctrine requires model-architecture improvements, whether it is enough to create a practical tool by applying machine learning to new data environments.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

PERA 2025: Eligibility Reform Returns to Capitol Hill

by Dennis Crouch

On October 8, 2025, the Senate Judiciary Subcommittee on Intellectual Property held a hearing on the Patent Eligibility Restoration Act (PERA) of 2025 (S. 1546), legislation that would fundamentally reshape 35 U.S.C. § 101 and effectively overturn the Supreme Court's Mayo, Myriad, and Alice decisions.  The hearing featured testimony from former USPTO Directors Andrei Iancu and David Kappos, China IP expert Mark Cohen, biotech executives, patient advocates, and retail industry representatives, all addressing whether Congress should replace the judge-made "abstract idea" and "law of nature" exceptions with a set of narrow statutory exclusions. Chairman Thom Tillis (R-NC) and ranking member Adam Schiff (D-CA) both acknowledged the widespread confusion in current eligibility jurisprudence, with Tillis noting that the status quo "continues to be hopelessly confused" and expressing determination to move the bill forward to committee markup. The Bill has bipartisan support in the Senate, but it is unclear  how hard those in opposition would push if the bill started to move forward.

PERA was originally introduced back in 2019 as a draft proposal. It was later reintroduced as S.4734 in 2022 and has been refined through successive versions, with the current 2025 bill (S.1546) representing a more measured approach than earlier drafts while still fundamentally overhauling § 101. The 2019 proposal included explicit language stating that all judicially created exceptions were "hereby abrogated." Later versions achieve the same result more subtly by specifying that eligibility is determined "subject only to" the enumerated statutory exclusions. The 2022 version would have made isolated human genes patent-eligible, directly contradicting Myriad's core holding, but the 2025 bill narrows this by removing "isolated" from the human gene provision—meaning mere isolation of natural DNA wouldn't suffice for eligibility, though purified, modified, or engineered genes used in inventions would still qualify. Similarly, where the 2022 draft used the ambiguous term "non-technological" to define excluded business methods, the 2025 version asks whether the machine is "necessary to practically perform" the invention. Probably most importantly, the 2025 bill adds new guardrails against patenting superficial computerization of otherwise abstract ideas.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

Guest Post by Ryan Abbott: Reflections on Recentive v. Fox: To Do or not To Do it with AI?

By Ryan Abbott. Professor Abbott, MD, JD, MTOM, PhD, is Professor of Law and Health Sciences at the University of Surrey School of Law, Adjunct Associate Professor of Medicine at the David Geffen School of Medicine at UCLA, partner at Brown, Neri, Smith & Khan, LLP, and a mediator and arbitrator with JAMS, Inc.

Recently, the Federal Circuit handed down its decision in Recentive v. Fox which the Court stated presented a question of first impression: “whether claims that do no more than apply existing methods of machine learning to a new data environment are patent eligible.” Recentive Analytics, Inc. v. Fox Corp., 134 F.4th 1205 (Fed. Cir. 2025). As the framing of the question suggests, the Court answered in the negative, confirming that patents that simply take an abstract idea and “do it with AI” are not statutory subject matter.

Since this decision, questions have come up about what this might mean for the patenting of inventions relating to artificial intelligence (AI) in the United States, and by extension, America’s leadership in AI development. Even before this case, the argument has been made that getting patents on AI innovations is difficult or impossible under the current law, and therefore patent eligibility reform is needed. These are important questions. As I have argued for years in my work on the Artificial Inventor Project, and as I testified on in Congress, AI technology is playing a rapidly growing role in the innovation process. We want to make sure the patent system is promoting the development and use of AI so that it can be used to generate socially valuable innovation (and that there are no legal concerns when that happens, including related to inventorship).

My view is that current patent eligibility law as it pertains to abstract ideas best supports the AI development and innovation that we need in the United States. The current approach permits the patenting of all aspects of AI technology – from the development and training of machine learning models to the applications of those models. Recentive gave us a glimpse into how patent eligibility law applies at one end of the AI patenting spectrum, and it reached the right conclusion.

Before delving into Recentive, it is helpful to understand the application of patent eligibility law in a traditional software patenting context. Eligibility analysis relies on the two-step Alice/Mayo test, which first determines whether claims are directed to judicial exceptions to eligible matter, namely abstract ideas, laws of nature, or natural phenomena. Alice Corporation v. CLS Bank International, 573 U.S. 208 (2014); Mayo Collab. Servs. v. Prometheus Lab’ys, Inc., 566 U.S. 66 (2012). If claims are directed to a judicial exception then to be patent eligible they must recite additional elements that amount to “significantly more” than the judicial exception. Mayo, 566 U.S. at 72-73. The Supreme Court has described the second part of the test as the search for an “inventive concept”. See MPEP § 2106. This is the Supreme Court’s way of ensuring that a true, technological invention lies at the heart of the patent and that it isn’t just an abstract idea dressed up to look like a real invention.

For more than a decade, the Alice/Mayo two-part test has been the exclusive test for software patent eligibility analysis. It balances patent incentives with the need to allow the public to utilize “the basic tools of scientific and technical work.” Alice Corp., 573 U.S. at 216. Patents on such tools could do more harm than good and “might tend to impede innovation more than it would tend to promote it.” Id. Claims are not ineligible simply because they involve judicial exceptions, the applications of such concepts remain protectable. Indeed, the Federal Circuit has subsequently clarified that claims directed to software and improvements in computer-related technologies are not inherently abstract. See, e.g., Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1335 (Fed. Cir. 2016); see also Visual Memory LLC v. NVIDIA Corp., 867 F.3d 1253, 1258 (Fed. Cir. 2017). Judicial exceptions have existed for more than 150 years, and a significant amount of case law has now applied Alice/Mayo in the context of software patents.

The Federal Circuit has tended to require patent applications to describe specific improvements over the prior art that improve a process, whether software- or hardware-based, rather than merely performing an existing process on a computer. This has allowed a wide range of computer-implemented claims to be found patent-eligible. See, e.g., McRO Inc. v. Bandai Namco Games America Inc., 837 F. 3d 1299 (Fed. Cir. 2016); Amdocs (Israel) Ltd. v. Openet Telecom Inc., 841 F. 3d 1288 (Fed. Cir. 2016); Ancora Technologies Inc. v. HTC America Inc., 908 F. 3d 1343 (Fed. Cir. 2018). This standard thus encourages research and development into improvements in computer processes, while preventing applicants from broadly tying up the use of computers to perform existing processes and relying on functional claiming (e.g., filtering content on a computer, unless an improvement over the prior art for filtering content; parsing, editing, sorting, and comparing data, absent a specific inventive method for carrying out these functions). See BASCOM Glob. Internet Servs., Inc. v. AT&T Mobility LLC, 827 F.3d 1341, 1348 (Fed. Cir. 2016); see Berkheimer v. HP Inc., 881 F.3d 1360, 1366 (Fed. Cir. 2018).

With the foundation of applying patent eligibility law to traditional software patents, we can more fully understand the Federal Circuit’s assessment of patent eligibility in an AI context in Recentive. This case concerned the eligibility of using machine learning, a subset of AI, to generate network maps and schedules for television broadcasts and live events. Recentive was appealing the district court’s ruling that its patents were directed to ineligible subject matter under 35 U.S.C. §101. The Federal Circuit affirmed the lower court on the basis that the patents lacked an inventive concept and were directed to the abstract idea of using a generic machine learning technique in a particular environment.

Preparing network maps is a long-standing practice, if a sophisticated and niche one that has traditionally been performed by human beings. Recentive characterized its patents as applying existing, generic AI approaches to this specific context. In other words, “do it with AI.”

This type of “do it with AI” claim is ineligible under the two-step Alice inquiry.  On the first step, the Court found that the claims were directed to the abstract idea of producing network maps and event schedules using known standard mathematical techniques. Recentive’s patent did not claim or describe how the machine learning technology resulted in a technological improvement. Instead, the patent was just about using AI to perform a task previously done by humans more quickly or efficiently without actually explaining how to accomplish that, which does not create eligibility. This is the same application of the Alice/Mayo test to traditional software patents.

On the second step, which looks at whether additional elements add an “inventive concept” that transforms an otherwise ineligible claim, the Court also found the patents failed. The patents only provided broad, functionally described common computing techniques and devices—again merely claiming an abstract idea.

While the Court framed Recentive as a question of first impression, the Federal Circuit has long held that software claims do not become eligible by limiting their application to a particular field or environment – eligibility does not result from taking an abstract idea and throwing in “do it on a computer” or “do it over the internet.” See, e.g., Intell. Ventures I LLC v. Capital One Bank (USA), 792 F.3d 1363, 1366 (Fed. Cir. 2015). Similarly, that applying an existing technology to a new database does not create eligibility.  See, e.g., SAP Am., Inc. v. InvestPic, LLC, 898 F.3d 1161, 1168 (Fed. Cir. 2018). Like the software patent cases that came before it, Recentive just applied AI to an existing framework.

While simply describing an abstract idea and adding “do it with AI” is not patentable (nor should it be), patent protection is available for all aspects of AI innovation. Practitioners have outlined best practices for drafting these patent applications, emphasizing the importance of providing sufficient detail and explaining the technical advantage of the claimed solution. https://www.fr.com/insights/thought-leadership/articles/ai-patent-eligibility-observations-and-lessons-for-the-us-and-china/ (“The key for applicants is to clearly show that their claim improves a technological process rather than merely uses technology to perform an existing process.”) https://www.troutman.com/insights/locke-lord-quickstudy-patenting-ai-inventions.html. (“Draft a strong patent application… Focus on technical aspects and algorithms… Provide experimental evidence and results…”). Of course, applicants must also ensure what is being claimed is commensurate with what was actually invented and not improperly tying up the basic tools of further scientific development. The USPTO’s most recent guidance on patent subject matter eligibility in 2024 provided some examples as well.  See, e.g., https://www.uspto.gov/sites/default/files/documents/2024-AI-SMEUpdateExamples47-49.pdf.

Recent district court decisions applying Recentive have also upheld claim eligibility for AI-related claims. Aon Re, Inc. v. Zesty.AI, Inc., No. CV 25-201, 2025 WL 1938214 (D. Del. July 15, 2025) (“The claim thus implements machine-learning technology in a specific way to address a practical technical problem: how to accurately and quickly assess the physical characteristics and condition of a property from an aerial image using machine learning models.”); Nielsen Co. (US), LLC v. Hyphametrics, Inc., No. CV 23-136-GBW, 2025 WL 1672002 (D. Del. June 13, 2025) (Claimed method provides “an improvement to the technical field of image processing, and, in particular, to improving speed and reliability of detecting overlays in images, with reduced or eliminated human oversight… These claims contain structure because they, inter alia, recite the specific [neural] networks used to perform the detection”).

Though all aspects of AI can be patented under Alice – and our patent eligibility law aims to ensure that AI patents do not over-claim in a way that would harm further AI innovation and usage – AI has nonetheless become yet another argument in support of patent eligibility reform. Various changes have been proposed over the past decade to expand the scope of section 101. Most recently, Senator Thom Tillis re-introduced the Patent Eligibility Restoration Act (PERA) in the Senate. https://www.congress.gov/bill/119th-congress/senate-bill/1546/text. This Act would eliminate all judicial exceptions to patent eligibility, replacing them with explicit and narrow statutory exclusions: 1) a mathematical formula not integrated into a claimed invention; 2) a mental process performed solely in the human mind; 3) an unmodified human gene as it exists in the human body; 4) an unmodified natural material as it exists in nature; and 5) a process that is primarily economic, financial, business, social, cultural, or artistic in nature. Most relevant to the issue of AI patenting, and what would be patent-eligible if PERA were to become law, the bill includes an eligibility escape hatch. It states that “the claimed invention shall not be excluded from eligibility for a patent if the invention cannot practically be performed without the use of a machine or manufacture.” This would mean that any abstract ideas could be patented if implemented on a computer leveraging computational capabilities. Effectively all AI inventions “cannot practically be performed without the use of a [computer].” This means that under PERA, “do it with AI” patents like the one found to be ineligible in Recentive would be eligible.

As a result, PERA would be a significant shift for patent eligibility of computer-related patents, and one that would harm rather than help the development and use of AI. Problematically, by untethering patent eligibility for abstract ideas from the requirement that patents be about technological improvements, PERA would allow patents that harm the development and widespread usage of actual technology. This in turn is likely to result in more patent thickets which will make AI development increasingly difficult and expensive, it will give incumbents more leverage over startups, and it risks holdups from foreign adversaries abusing the patent system. This is precisely the type of approach that allowed a flood of “do it on a computer” and “do it on the internet” patents to harm software development and usage for years. It is not a standard that is likely to help the United States remain the world’s leader in AI development.      Nor is it a standard that will support innovators using AI as part of their development process.

And that is where our focus should lie: making sure our patent system is helping to incentivize AI technology development and usage in order to move the field forward. Patent eligibility has a pivotal gatekeeping role in the system, as the Federal Circuit correctly showed in Recentive. The line that has been drawn and applied in a software context for years is also the right one for AI.

The author represents that he is not being paid to take a position in this post. In addition to being an academic, the author is a practicing attorney and represents clients with interests on both sides of the post’s subject matter.

Director Squires Takes Aim § 101 Rejections

by Dennis Crouch

In a striking early display of directorial authority, newly sworn-in Under Secretary of Commerce and USPTO Director John Squires has vacated a Patent Trial and Appeal Board (PTAB) decision that had entered a new ground of rejection under 35 U.S.C. § 101 against DeepMind Technologies' machine learning patent application. The September 26, 2025 decision in Ex parte Desjardins, Appeal 2024-000567, Application 16/319,040, represents one of Director Squires' first substantive actions since taking office one week ago and offers yet another clear signal that Section 101 reform is building steam. SquiresPTABDecision.

Squires writes:

Categorically excluding AI innovations from patent protection in the United States jeopardizes America's leadership in this critical emerging technology. . . the panel essentially equated any machine learning with an unpatentable 'algorithm' and the remaining additional elements as 'generic computer components,' without adequate explanation. . . . most troubling [the Panel] eschewed the clear teachings of Enfish, and instead substituted only a cursory analysis that ignored this well-settled precedent. Panels should treat such precedent with more care, especially when acting sua sponte.

Finally, and most clearly a signal to all USPTO employees to back-off with regard to patent eligibility, he writes:

This case demonstrates that §§ 102, 103 and 112 are the traditional and appropriate tools to limit patent protection to its proper scope. These statutory provisions should be the focus of examination.

The application, filed by DeepMind (owned by Google) and prosecuted by Fish & Richardson, claims methods for training machine learning models on multiple sequential tasks while avoiding "catastrophic forgetting," which is defined as the phenomenon where neural networks lose knowledge of previous tasks when trained on new ones.  As someone with some ADHD tendencies, I recognizing this issue.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

The PowerBlock Problem: A Roadmap for Dodging Patent Eligibility Hurdles

by Dennis Crouch

Accused infringer iFit Inc. has filed a combined petition for panel rehearing and rehearing en banc in PowerBlock Holdings, Inc. v. iFit, Inc., No. 24-1177 (Fed. Cir. filed Sept. 10, 2025), challenging the Federal Circuit's recent decision finding the  automated exercise equipment claims eligible. The petition argues that the panel's holding directly conflicts with the court's precedent in Yu v. Apple Inc., 1 F.4th 1040 (Fed. Cir. 2021) as well as Supreme Court precedent by allowing systems to avoid ineligibility simply by reciting conventional components that perform basic functions at a high degree of generality.  The patent in this case seems to me similar to so many others -- it claims a known physical system at a fairly high level of generality and couples that with previously unknown information processing.  This combination was enough to overcome an abstract idea challenge - and suggests a roadmap for artful dodging of the eligibility hurdles.  In my view, the PowerBlock panel also seems to have have applied a mechanical versus software distinction that is not supported by Supreme Court precedent.

I hope that the original panel will take a hard look at their reasoning. As noted below, I believe there are a couple of specific errors that could be fixed while still reaching the same outcome. That result would help insure that this decision avoids working potential mischief.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

The Federal Circuit’s Heavy Lifting on Automation Patents

by Dennis Crouch

In a decision that should provide welcome clarity for mechanical device patents, the Federal Circuit reversed a district court's dismissal of patent infringement claims involving automated exercise equipment, holding that claims incorporating an electric motor to automate previously manual dumbbell weight selection are not directed to abstract ideas under Alice step one. PowerBlock Holdings, Inc. v. iFit, Inc., No. 24-1177 (Fed. Cir. Aug. 11, 2025).  While the decision provides a victory for mechanical automation patents, the court's reasoning appears to turn primarily on the physical nature of the claimed system - using an electric motor to move mechanical components.

A few issues buried inside:

  • Mechanical vs. software distinction - Court emphasized the physical nature of the invention (electric motor moving mechanical components) as distinguishing it from purely software-based automation patents. In other words physical automation is much more likely eligible.
  • CAFC Erred in decided on narrow claim - Court focused analysis on detailed claim 1 then summarily declared broader claim 20 also eligible, despite parties arguing the claims separately and meaningful distinctions between them.  I see this as an error that should get a rehearing.
  • Section 101 boundary warning against ignoring old elements - The decision included an explicit caution to "parties and courts" against conflating patent eligibility with novelty/obviousness inquiries by ignoring conventional elements in the eligibility analysis.  The interesting bit is that the court inverted quotes from Mayo and Flook to justify its statements.

To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

En Banc Ineligibility Petition in Longitude Licensing v. Google

by Dennis Crouch

Before digging into this en banc petition, I just wanted to note that patent eligibility issues have been almost shockingly quiet over the past six months.  But, to me it feels something like a powder keg waiting to explode. While Acting Director Stewart has been incredibly active on other fronts since taking the helm in January 2025, she has remained notably silent on Section 101 reform initiatives. I believe there is some likelihood that this will change once John Squires is confirmed by the U.S. Senate -- given that he has already signaled support for patent eligibility reform measures like the Patent Eligibility Restoration Act (PERA).

Back to the case: In Longitude Licensing Ltd. v. Google LLC, No. 2024-1202 (Fed. Cir. Apr. 30, 2025), the Federal Circuit affirmed a N.D. Cal. dismissal of infringement claims under 35 U.S.C. § 101, holding Longitude’s asserted image processing patents were directed toward digital image processing.  Longitude’s patents (U.S. Patents Nos. 7,668,365, 8,355,574, 7,454,056, and 7,945,109) cover methods of identifying a primary object within an image and adjusting image characteristics based on the object’s properties. The panel, consisting of Judges Lourie, Dyk, and Chen, determined that representative Claim 32 of the ’365 patent recited only generic, functional limitations without adequately describing how these limitations achieved the purported technological improvements.

32. An image processing method comprising:

determining the main object image data corresponding to the main object characterizing the image;

acquiring the properties of the determined main object image data;

acquiring correction conditions corresponding to the properties that have been acquired; and

adjusting the picture quality of the main object image data using the acquired correction conditions;

wherein each of the operations of the image processing method is executed by an integrated circuit.

Longitude has now petitioned the Federal Circuit for rehearing en banc, asserting that the panel improperly isolated the claim language from the specification and ignored evidence that the claimed methods improved computer functionality itself. The en banc petition frames the question presented as whether courts may determine patent eligibility under § 101 by analyzing claim language alone, divorced from supporting specification evidence, particularly where the specification describes specific technological improvements. The Federal Circuit has shown some interest in the case - requesting responsive briefing from Google, which is due in Mid-July 2025. [Longitude Licensing En Banc Petition].


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.

“Do it on AI” claims are Abstract Ideas

by Dennis Crouch

The Federal Circuit's AI-patent decision this week affirmed a district court's dismissal of the claims as ineligible, holding that "generic" machine learning technology is itself an abstract idea. Recentive Analytics, Inc. v. Fox Corp., No. 2023-2437 (Fed. Cir. Apr. 18, 2025).

In its decision, the court acknowledged the growing importance of machine learning, and exerted some care in limiting its holding:

Machine learning is a burgeoning and increasingly important field and may lead to patent-eligible improvements in technology. Today, 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.

At oral arguments, the patenee's attorney Robert Fredrickson (Goodwin) explained that one reason his client did not claim a new algorithm is that it would have raised 101 issues:

This wasn't an invention of a new machine learning technique, because that would fall into another one of this court's section 101 traps, is if the claim is improving the mathematical algorithm or making machine learning better, then we're claiming the natural law, the mathematical algorithm itself.

That question --  scope of protection for new machine learning algorithms -- will be left for another day.


To continue reading, become a Patently-O member. Already a member? Simply log in to access the full post.