In re Killian: Defining “Abstract Idea” and “Inventive Concept”

In re Killian is interesting because the Federal Circuit used that case to respond to the frustration many patent applicants feel regarding the amorphous and unpredictable nature of analysis under 35 U.S.C. § 101. Killian made that frustration quite plain, leading off his appeal with the argument that

the standard promulgated in Alice and Mayo is indefinite—so poorly defined that it renders all court and Board decisions finding a claim patent ineligible under the Alice/Mayo standard arbitrary and capricious under the Administrative Procedure Act (APA), 5 U.S.C. § 706(2)(A), and the Due Process Clause of the Fifth Amendment.

While there was basically no chance that the Federal Circuit would agree to an argument premised on the proposition that the standard laid out by its reviewing court (and applied by it) was incomprehensible, I think the Federal Circuit choosing to address that argument with a precedential opinion rather than affirming the board under rule 36 is interesting in and of itself. Even more is that the Federal Circuit acknowledged and responded to “Mr. Killian’s desire for ‘a single noncapricious definition or limiting principle’ for ‘abstract idea’ and ‘inventive concept.’” Thus, In re Killian not only shows that the Federal Circuit feels the need to respond to criticisms of the Alice/Mayo framework, it also shows what the Federal Circuit feels is the closest it can come to concrete definitions of the phrases “abstract idea” and “inventive concept” which are at the heart of the eligibility determination.


                Turning first to “abstract idea,” the Federal Circuit began by acknowledging that there is no single rule which could distinguish what was and was not an abstract idea.  It explained that such a rule did not exist because there were at least three different types of abstract ideas: (1) methods of organizing human activity, (2) claims to mental processes, and (3) claims to results rather than means of achieving them. It also said that it had provided guidance as to what constitutes an abstract idea in the form of cases which said that the abstract idea determination was made based on the claimed advance over the prior art, and cases which said that the intangibility of information meant that advance in the specified content of information were abstract. It finally turned specifically to software inventions, stating that, for that kind of invention eligibility “often” turns on whether the claims focus on “the specific asserted improvement in computer capabilities” or “on a process that qualifies as an ‘abstract idea’ for which computers are invoked merely as a tool.”


                The treatment of “inventive concept” was similar, with the Federal Circuit indicating that it had provided guidance, but stopping short of elaborating a clear definition. With respect to guidance, there was actually very little on offer, with the examples of guidance laid out by the Federal Circuit largely restating principles already laid out as “guidance” for what was an abstract idea. For instance, the first example of guidance in the discussion of “inventive concept” was that “methods that ‘improve[] an existing technological process’ include an inventive concept at step two.” However, the discussion of abstract ideas had already stated that the lack of “any particular assertedly inventive technology” would doom a claim to an invention where the improvement was information of specified content, and so it isn’t clear what would be added by treating an improvement to “an existing technological process” as an inventive concept.


                Ultimately, I don’t think the Federal Circuit’s explication of its “guidance” would be particularly satisfying to someone who, like the appellant in Killian, is frustrated by the current state of 101 caselaw. Indeed, the fact that the Federal Circuit could do no better than provide a non-exhaustive list of types of abstract ideas, then point to various past cases as “guidance” seems to confirm the ad-hoc and nature of contemporary subject matter eligibility analysis. I suspect that the Federal Circuit realized this, since it ended its discussion of “abstract idea” and “inventive concept” by explaining that, even if it was “persuaded by Mr. Killian’s argument that the Alice/Mayo framework is insolubly unclear, both this court and the Board would still be bound to follow the Supreme Court’s § 101 jurisprudence as best we can as we must follow the Supreme Court’s precedent unless and until it is overruled by the Supreme Court.” Hopefully, the Supreme Court will view this as an invitation to come up with a standard that a court couldn’t even hypothetically be persuaded is “insolubly unclear,” as otherwise I think the frustration felt by Mr. Killian (and many others) is likely only to get worse.


Disclaimer – “The statements and views expressed in this posting are my own and do not reflect those of my law firm, are intended for general informational purposes only, and do not constitute legal advice or a legal opinion.”

Cert denied in American Axle

Surprisingly, despite a solicitor recommendation, the Supreme Court denied cert in
American Axle. I can’t say I’m particularly disappointed. I recorded a
video last year (best viewed at 2x speed) where I talked about the potential outcomes of
the Supreme Court taking it up. I think my conclusion then – i.e., that
the Supreme Court was more likely to make things worse than better –
was basically right. Some clarity in the 101 analysis would have been
great, but I don’t think we would have gotten that, and the Supreme
Court passing up an opportunity to make things worse (again) is probably
all for the best.

Disclaimer – “The statements and views expressed in this posting are my own and do not reflect those of my law firm, are intended for general informational purposes only, and do not constitute legal advice or a legal opinion.”

American Axle v. Neapco: Some thoughts (2021)

I want to talk about American Axle v. Neapco.  This is a reasonably interesting case in its own right, since it applied section 101 to driveline propeller shaft manufacturing technology, which is obviously a bit different from the software and biotech inventions where section 101 is generally applied.  However, rather than talking about the case itself, I want to talk about what could happen if the Supreme Court grants cert, since they have asked for the solicitor’s views on the matter and that’s usually a sign that the case has a good chance of being heard.  I have seen articles out there saying that the Supreme Court should grant cert so they can fix the mess that is section 101.  What I want to do is explain why I think the Supreme Court granting cert would most likely make section 101 jurisprudence even more of a disaster than it already is.

Taking it doesn’t mean clarifying Alice/Mayo

I think part of the reasoning behind wanting the Supreme Court to take up American Axle is that the Supreme Court doesn’t take cases just to affirm, and so if they do take it they’re like to reverse the Federal Circuit which, by definition, would mean they would have to put some limits on the currently amorphous 101 analysis.  I think this is wrong.  Maybe the Supreme Court doesn’t take cases just to affirm them, but it did just that in Alice and Bilski.  You can say that those were special cases, since the PTO had basically teed up Bilski for Supreme Court review, and the Federal Circuit was so fractured on Alice that it basically amounted to a circuit split requiring review.  However, while the Federal Circuit might not be as fractured as it was for Alice, there are still significant differences between panels on how to apply section 101, and the fact that those differences persist even seven years after Alice may make the Court feel it needs to step in.  Additionally, while the patent bar, or at least the part of it that pays attention to these issues, is keenly aware of the drawbacks of section 101, the outside world is still focused on “patent quality” and the dangers of “Bad patents”.  Indeed, there was a hearing last month before the senate judiciary committee’s intellectual property subcommittee just last month that focused on exactly those issues, and I think the Supreme Court is much closer in outlook to the Senate’s intellectual property subcommittee than it is to the portion of the patent bar that is concerned about section 101.  Accordingly, my feeling is that if the Supreme Court does take up American Axle, it is likely to do so in order to emphasize that the Alice/Mayo analysis that has heretofore been mostly confined to software and biotech should be applied to other types of inventions as well, not to fundamentally clarify or make any changes to its application.

This isn’t a great case for clarifying anything

However, let’s say I’m wrong, and the Supreme Court does take up American Axle with the goal of significantly clarifying the analysis under section 101.  I think American Axle is a lousy case for the Supreme Court to take up if it wants to make things clearer.  First, the claims in American Axle were found ineligible because they were directed to a law of nature, so right off the bat this is a suboptimal vehicle for clarification because it doesn’t present the opportunity to define “abstract idea”, and I think the lack of a definition for an “abstract idea” is one of the biggest problems with 101 jurisprudence as it currently stands.  Also, because American Axle came through the courts rather than the PTO, it wasn’t analyzed using the patent office’s framework, which I think is the clearest framework anyone has come up with so far for applying section 101.  As a result, any significant change in the section 101 analysis is likely to be orthogonal to the patent office’s analysis, and so require major revisions to current PTO practice.  Does that necessarily mean it would be a step backward in terms of clarity?  Not necessarily, but after seeing what the Supreme Court has done in the past, I think the best we can reasonably hope for is that it leaves things largely intact so that we don’t all have to go back to the drawing board.

The only potential silver lining is that if the Supreme Court endorses broad application of section 101, we’ll finally see some meaningful limits 

So, is there any potential for a good outcome if the Supreme Court takes up American Axle?  There’s always the possibility that they’ll replace our current 101 mess with a meaningful, logical, predictable framework.  However, that’s so unlikely I think I can safely ignore it.  Are they likely to make things worse in the one context where section 101 is working at least relatively well?  Yes.  However, there is another possibility, which I alluded to earlier.  They could leave things essentially unchanged in terms of the substance of the 101 analysis, and instead take up the case to endorse applying 101 across the board, not just to software and biotech.  What happens then?  Well, obviously chaos for everyone, but maybe that chaos will lead to section 101 being applied a little more rationally in the long term.  In support of this optimistic outlook, I would like to draw an analogy to the broadest reasonable interpretation rule.  When it was only used during prosecution, the broadest reasonable interpretation rule could reasonably be criticized as essentially defaulting to the broadest possible interpretation, with reasonableness being an afterthought at best.  However, once it started being applied in inter partes review proceedings, the Federal Circuit started issuing opinions that found the broadest reasonable interpretation was a good deal narrower than I would have expected from the way it had traditionally been applied to patent prosecution.  That is, when an extremely broad broadest reasonable interpretation didn’t advance the policy goal of facilitating patent examination, it started narrowing in scope.  Similarly, if section 101 started being applied in areas where its aggressive use didn’t advance anti-software or anti-biotech policy preferences, it might happen that the Federal Circuit would find ways to make it easier to establish eligibility.  Just a thought, and something of a long shot.  However, given the Supreme Court’s track record I think it’s probably the best that could come from them hearing American Axle.

Anyway, I’d love to be proven wrong, and, right or wrong, I don’t think we’ll actually see any impact for at least another year.  In the meantime, I wish you good luck in dealing with section 101, and in your patent endeavors more generally.

Thanks a lot…bye bye.   

Disclaimer – “The statements and views expressed in this posting are my own and do not reflect those of my law firm, are intended for general informational purposes only, and do not constitute legal advice or a legal opinion.”

In re Stanford: Ruined by a Processor and a Memory

“The Federal Circuit’s focus on an aspect of the Stanford application which no one could (or did) argue was a protectable innovation not only damaged Stanford, it also squandered an opportunity to bring clarity to an area of law that desperately needs it.”

Computer boilerplate – such as including “a processor and a memory” in claims – is commonplace in patent applications. However, the recent case of In re Stanford shows that this can be a double-edged sword, having the potential to both undermine an application and to ruin an opinion that could otherwise have shed light on several of the thorniest open questions in patent eligibility jurisprudence. Skeptical that such a common practice could be so counterproductive? Read on.

Language with No Benefit

From the outset, the invention in Stanford is not one where computer components should have had any particular significance. The invention was not a new type of computer, or a new component, such as an innovative processor or memory. Instead, it was (at least according to Stanford) a significant advance in the field of genomics. Many types of analysis required for genome interpretation are dependent on haplotype phasing – i.e., determining whether a gene was inherited from a person’s father or mother. The Stanford invention used a particular type of model with particular states – a hidden Markov model with inheritance, compression fixed error, and MIE-rich fixed error states – to perform haplotype phasing. This significantly increased the resolution of haplotype phasing, moving from 80% in the prior art to 97.9% for phasing using the Stanford invention.

Despite the fact that the Stanford invention was not, and did not purport to be, an improved or a novel computer, both the claims and description of the Stanford application included computer components. The description explained that “the present invention relates to methods, techniques, and algorithms that are intended to be implemented in a digital computer system” (application 13/445,925, paragraph 56), before saying that “such a digital computer or embedded device is well-known in the art” (id.) and providing a description of a generic computer system that is repeated in a variety of other Stanford patents and applications (e.g., U.S. 8,966,413, U.S. 2013/0073329). The claims included similar generic computer language, with limitations requiring various steps of the innovative haplotype phasing method to be performed “on a computer system comprising a processor and a memory” or “using a computer system comprising a processor and a memory” being added to the claims in response to a rejection under 35 U.S.C. § 101.

From Ineffective to Detrimental

Stanford does not appear to have benefitted from the generic computer language in their application and, to the extent that language had any impact at all, the impact appears to have been negative. The addition of computer language to the claims did not convince the examiner to withdraw the 101 rejections. Instead, the examiner stated that “[t]here is no limitation in the claims that the claimed computer program or computer uses other than a generic computer,” maintained the rejection and made the successive office action final. (Office Action of 7-9-2015 on application 13/445,925). On appeal, the computer language appears to have gone from having no effect to being actively detrimental. The patent trial and appeal board explicitly referenced the application’s generic computer system description when explaining that “the claimed method does not teach a technical improvement in a computer processor, but rather use the computer as a tool to improve data analysis.” (Ex parte Dewey, Appeal No. 2018-005489, on application 13/445,925 (PTAB 2019) at 11). Similarly, the Federal Circuit specifically noted the claims’ recitation of a “processor” and a “memory” in support of its statement that “it is hard to imagine a patent than that recites hardware limitations in more generic terms than the terms employed in claim 1.” (In re Stanford, Case No. 20-1012 (Fed. Cir. 2021) at 11). Unsurprisingly, the attention paid to aspects of its application which were admittedly not innovative did not bode well for Stanford, and its appeals were unsuccessful at both the Patent Trial and Appeal Board (PTAB) and the Federal Circuit.

A Missed Opportunity for Clarity 

The attention paid to the non-innovative aspects of Stanford’s technology also made the Federal Circuit’s opinion much less useful in terms of clarifying the proper application of section 101. Looking purely at the law, it would have been reasonable to expect the Federal Circuit to have found the Stanford claims eligible. In McRO, Inc. v. Bandai Namco Games America, Inc., the Federal Circuit had already dealt with the eligibility of a method which used an assertedly innovative computational tool to improve an inherently computerized process, and found that claims which recited the tool with sufficient specificity were not directed to an abstract idea. Additionally, McRO clearly undercut the reasoning used by the PTAB to find the Stanford claims ineligible – i.e., that they recited steps that could be performed in the mind or recited mathematical concepts – since the McRO claims had also recited mathematical concepts (i.e., determining morph weight sets based on phoneme subsequence timing) and improved a process that not only could be, but actually had been performed in the human mind (i.e., synchronizing character and lip models in computer animation). If the Federal Circuit was going to find Stanford’s claims ineligible, it had the opportunity to significantly clarify the rules of subject matter eligibility by clearly distinguishing McRO. It did not. Instead, its only mention of McRO was to note that the PTAB had distinguished that case because the McRO claims “improve ‘the computer animation process itself.’”  In re Stanford, at 6. It then stated that “we are not persuaded that claim 1 is not directed to an abstract mathematical calculation” and “we are not persuaded that the process is an improved technological process.” (id. at 10). No explanation whatsoever was provided for why an improved haplotype phasing process was not “an improved technological process” while an improved computer animation process was. Similarly, no explanation was provided for why Stanford was doomed for reciting mathematical calculations while McRO was not. Indeed, it appears that the most important factor for the Federal Circuit was that the claims recited “a processor and a memory” without purporting to improve them, since it saw fit to explicitly state that:

Notably, claim 1 neither requires nor results in, a specialized computer or a computer with a specialized memory or processor. Indeed, it is hard to imagine a patent claim that recites hardware limitations in more generic terms than the terms employed by claim 1.

Considerations Going Forward

The Federal Circuit’s focus on an aspect of the Stanford application which no one could (or did) argue was a protectable innovation not only damaged Stanford, it also squandered an opportunity to bring clarity to an area of law that desperately needs it. More ominously, it also raised the question of whether the nigh-ubiquitous practice of including form computer language may need to be rethought, since if the Federal Circuit puts so much weight on “a processor and a memory” in claims which recite those terms, maybe it would be better to omit them all together to avoid falling into the trap of having to assert an improvement to the processor and the memory, rather than being able to focus on the actual invention an application is intended to protect.

Originally published on IPWatchdog.

Disclaimer – “The statements and views expressed in this posting are my own and do not reflect those of my law firm, are intended for general informational purposes only, and do not constitute legal advice or a legal opinion.”