1. Home
  2. |Insights
  3. |The Conclusion of Am. Axle.: Considerations Relating to Section 101 Patent Eligibility in the Context of Artificial Intelligence & Machine Learning

The Conclusion of Am. Axle.: Considerations Relating to Section 101 Patent Eligibility in the Context of Artificial Intelligence & Machine Learning

What You Need to Know

  • Key takeaway #1

    Patent applicants and owners, especially those in the AI & ML spaces, should look to the guideposts established by American Axle to best position their application or patent to survive Section 101 challenges.
  • Key takeaway #2

    Patent Applicants and Owners should analyze claims and potential claims that may contain application of natural laws, complex logic, math, etc., and determine how they are focused on “tethered to tangible equipment,” and having “real-world impact.”
  • Key takeaway #3

    Patentees should assess the strengths and weaknesses of claim limitations and dependent claims where physical structure or steps may be lacking or missing, and how to blend more detail from the specification into the claims.

Client Alert | 8 min read | 08.21.23

Intro & Summary of MSJ Decision

In July 2023, the District of Delaware in Am. Axle & Mfg., Inc. v. Neapco Holdings LLC brought a long-running, closely-watched dispute to a close, leaving in its wake several guideposts for patent applicants and owners alike, including those pioneering technologies in the space of artificial intelligence (AI) and machine learning (ML).[1] 

The Delaware Court’s recent orders: 1) granted Plaintiff American Axle’s (“Axle”) Motion for Summary Judgment of No Invalidity Under 35 U.S.C. § 101; and, 2) denied Defendant Neapco’s (“Neapco”) Motion for Summary Judgment of Invalidity, both relating to claim 1 and its dependents of U.S. Patent No. 7,774,911 (“the ’911 patent”), as patent ineligible subject matter under the abstract idea exception to § 101.[2]Axle argued that the claims were “directed to a tangible, new and useful, multi-step industrial process manufacturing improved driveshaft—not an ‘abstract idea.’” As a representative claim, the Court found that claim 1 was directed to a tangible system with an observable real-world impact (including noise and vibration reduction), and recited technological improvements to tangible systems that performed concrete steps, including “tuning at least one liner” and “positioning the at least one liner within the shaft member,” as opposed to an “untethered” abstraction.

Procedural History

Am. Axle has had a torturous procedural history. Originally, on December 18, 2015, Axle sued Neapco for infringing three U.S. patents. Ultimately, however, the parties focused their dispute on the ’911 patent, and specifically, independent claims 1 and 22, and various dependent claims.

First District Court Summary Judgment Decision

In August 2017, the parties filed cross-motions for summary judgment as to the eligibility of the asserted claims of the ’911 patent under 35 U.S.C. § 101. The district court granted Neapco’s motion and held that the asserted claims are ineligible under § 101 because, “considered as a whole,” they were “directed to the mere application of Hooke’s law,” which deals with the frequency at which objects vibrate.[3] Axle appealed the district court’s decision to the Federal Circuit.

Federal Circuit Panel Decision and Modified Opinion

As an initial matter, the Federal Circuit panel affirmed the district court’s ruling in October 2019, but after a rehearing by Axle, the Federal Circuit modified its opinion in July 2020.[4]

The Federal Circuit concluded that independent claim 22 of the ’911 patent was patent ineligible under Section 101 because it: i) simply required the application of Hooke’s law to tune a propshaft liner to dampen certain vibrations; and, ii) on its face, does not identify the “particular [tuned] liners” or the “improved method” of tuning the liners to achieve the claimed result.[5] The Court reasoned that, while Axle may have discovered patentable refinements of the prior art process, such as particular uses of “sophisticated FEA [finite element analysis] models during its design process, . . . neither the specifics of any novel computer modelling, nor the specifics of any experimental modal analysis [were] included as limitations in claim 22.”[6] These unclaimed features, therefore, could not “function to remove claim 22 from the realm of ineligible subject matter.”[7]

Judge Kimberly Moore penned a particularly scathing dissent, focusing on three main points of contention.[8] First, the majority found the claims as being directed to natural laws, yet they clearly recite no such natural law.[9]Indeed, as she noted, “[u]nstated natural laws lurk in the operation of every claimed invention,” for example, “[e]very mechanical invention requires use and application of the laws of physics,”[10] such that “it cannot be that use of an unclaimed natural law in the performance of an industrial process is sufficient to hold the claims directed to that natural law.”[11] Second, the majority refused to consider certain unconventional claim elements. Third, the majority imbued § 101 with a new superpower—enablement on steroids.[12]Thus, the majority’s blended 101/112 analysis expands § 101, converted factual issues into legal ones, and is certain to cause confusion for future cases.[13]

In her view, Axle’s claims are patent-eligible because they “contain a specific, concrete solution (inserting a liner inside a propshaft) to a problem (vibrations in propshafts),” and a claim is not otherwise “directed to a natural law simply because compliance with a natural law is required.”[14]

In 2020, the panel issued a modified decision affirming the District Court’s findings on claim 22 and its dependent claims, but remanded for further proceedings on claim 1 and its dependents, saying they are “more general” than the ones the appeals court agreed were ineligible,[15]and explaining that claim 1 requires additional steps and has different claim constructions, where the specification may indicate or suggest that the “characteristics” that can be “tuned” include variables other than mass and stiffness.[16]Thus, because claim 1 also requires “positioning” in addition to tuning, and may reflect a broader definition of tuning, the Federal Circuit remanded the case to the district court to address the eligibility of claim 1 and its dependent claims in the first instance.[17]

Second District Court Summary Judgment Decision

On remand, the parties filed cross-summary judgment motions relating to patent eligibility of claim 1 and its dependents.[18] Finding claim 1 representative of the dependent claims, the Court found claim 1 was not directed to an abstract idea, because it recited a specific method for manufacturing a shaft assembly of a driveline system with specific steps.[19] The Court reasoned that: 1) the claims were directed to a tangible system, or a method of using such a system, with an observable real-world impact[20]; and, 2) the claims as a whole recite technological improvements to tangible systems or methods by performing concrete steps, including “tuning at least one liner” and “positioning the at least one liner within the shaft member.”[21]At bottom, the Court concluded that the asserted claims describe a real-world application of designing and manufacturing a liner for a propshaft system to achieve the desired damping of two modes of vibration, as opposed to an “untethered” abstraction.[22]

Lessons from Am. Axle

Am. Axle provides a framework for claims invoking computations, application of natural laws, physics, math, logic, and the like.  Ideally, such claims should be “directed to” or “tethered to tangible equipment” having “real-world impact.” To that end, Am. Axle calls for greater specificity in the claims. Particularly, in Am. Axle, the Federal Circuit explained that to avoid ineligibility, a claim must “ha[ve] the specificity required to transform [the] claim from one claiming only a result to one claiming a way of achieving it.”[23] At the other end of the spectrum, patentees have  “lost . . . claim[s] that encompassed all solutions for achieving a desired result” because the “claims failed to recite a practical way of applying an underlying idea . . . [and] instead were drafted in such a result-oriented way that they amounted to encompassing ‘the principle in the abstract’ no matter how implemented.”[24]

To be clear, the Federal Circuit does not suggest that specific novel computer or experimental processes cannot be claimed. Instead, the Am. Axle framework suggests that courts and patent applicants should be looking for more structure or steps that explain how principles can be implemented in an inventive way.

With that framework, the mechanical arts are likely safe, as Am. Axle seems to be an outlier in the mechanical arts.  Yet, there are considerations outlined in Am. Axle that are relevant to the software arts, particular to AI & ML, where software claims possibly contain complex logic or algorithms. With these types of complex claims, patent applicants can strengthen claim limitations by adding physical structures or steps for achieving the claimed results.[25] For example, claim limitations may specify “any novel computer modelling, [or] the specifics of any experimental modal analysis” or whatever structures or steps happen to work.[26]

As contemplated in Am. Axle, claim limitations should provide a step-by-step process on how computer modeling works to achieve the claimed results or the invention. For instance, how a computer takes advantage of any algorithms to manipulate large language models for training purposes, or to make accurate predictions. If there are any experimental modal analysis, what are relevant inputs or outputs? How is a computer changed by the inputs, outputs, the algorithm, etc. to achieve the claimed invention or results? Are there any data structures or hardware changes that happen with the algorithm or computer? These implemented steps can provide more specificity to strengthen the claims to survive a challenge under Section 101. Citing detail within the specification or figures, especially as related to examples of tangible equipment, and real-world impact, and blending those elements into limitations of the claims, can further fortify such claims against Section 101 eligibility challenges.

Takeaways

Patent applicants and owners, especially those in the AI & ML spaces, should look to the guideposts established by American Axle to best position their application or patent to survive Section 101 challenges, including by diminishing the application of natural laws, complex logic, math, etc., while focusing on “tethered to tangible equipment,” and having “real-world impact.” In addition, they should assess the strengths and weaknesses of claim limitations and dependent claims where physical structure or steps may be lacking or missing, and how to blend more detail from the specification into the claims. Nonetheless, the boundaries for patent eligibility are still in a state of flux, and despite following these guideposts, a court may still find patent claims ineligible under Section 101.

[1] No. CV 15-1168-GBW, (D. Del. Jul. 28, 2023).

[2] Id. at *6–18, *23–24.  

[3] Am. Axle & Mfg., Inc. v. Neapco Holdings LLC, 309 F. Supp. 3d 218, 225 (D. Del. 2018).

[4] Am. Axle & Mfg., Inc. v. Neapco Holdings LLC, 939 F.3d 1355, 1358, 1368 (Fed. Cir. 2019), reh’g granted, opinion withdrawn, 966 F.3d 1294 (Fed. Cir. 2020), and opinion modified and superseded on reh’g, 967 F.3d 1285 (Fed. Cir. 2020).

[5] Am. Axle & Mfg., Inc., 967 F.3d 1285, 1295 (Fed. Cir. 2020).

[6] Id. at 1295.

[7] Id. (See ChargePoint, Inc. v. SemaConnect, Inc., 920 F.3d 759, 766 (Fed. Cir. 2019), cert. denied, ––– U.S. ––––, 140 S. Ct. 983, 206 L.Ed.2d 135 (2020)).

[8] Id. at 1304–05.

[9] Id. at 1304.

[10] Id. at 1305, 1309.

[11] Id. at 1309.

[12] Id. at 1305.

[13] Id. at 1304–05.

[14] Id. at 1308.

[15] Id. at 1300.

[16] Id.

[17] Id.

[18] Am. Axle & Mfg., Inc., No. CV 15-1168-GBW, at *1 (D. Del. Jul. 28, 2023).

[19] Id. at *9–12.

[20] Id. at *16.

[21] Id. at *12.

[22] Id. at *16.

[23] Am. Axle & Mfg., Inc., 967 F.3d 1285, 1296 (Fed. Cir. 2020).

[24] Id.

[25] Id. at 1308.

[26] Id. at 1295.

Insights

Client Alert | 8 min read | 11.21.24

New Legislation Introduced in Congress Proposes Ending Normal Trade Relations with China and More

On November 14, 2024, Rep. John Moolenaar (R-Mich.), chair of the House Select Committee on the Chinese Communist Party, introduced the Restoring Trade Fairness Act, seeking to suspend China’s Permanent Normal Trade Relations (“PNTR”) status....