I think it is fair to say that the theme that comes through the loudest in the IBM brief is this: The abstract idea doctrine is unworkable. To that I say a resounding AMEN! If the Supreme Court cannot or will not tell us what an abstract idea is how can we any longer pretend that the jurisprudential path the Court has taken will lead to predictability? At least insofar as software is concerned there is a complete and total lack of predictability. There is also no uniform application of the law, which at least conceptually should raise concerns of disparate treatment of those similarly situated.
Below I provide additional thoughts on the IBM Summary of the Argument in the format of comments from the peanut gallery, or perhaps as a patent law equivalent to Mystery Science Theater 3000. In order to differentiate my thoughts/comments from IBM amicus brief, my comments are italicized, colored, indented and tagged with the IPWatchdog logo.
Before proceeding allow me to make one general observation. I wish the IBM brief had made a bigger deal about Watson, which as far as I can tell is the first generation version of the Star Trek omnipotent computer. A review of Watson claims rather clearly suggests to me that if the claims at issue in Alice v. CLS are not patent eligible than Watson is not patent eligible. Have we really reached the point where artificial intelligence is not patent eligible? I think IBM could have made a compelling argument on this point, but their brief is still very good and quite persuasive. Despite my wish that Watson had played a more central exemplary role I fully appreciate that space is very limited. Nevertheless, if my thoughts on the patentability of Watson and similar innovations is of interest please see:
- Is IBM’s Watson Still Patent Eligible?
- Are Robots Patent Eligible?
- Did the Federal Circuit Ignore the Supreme Court in CLS Bank?
Without further ado, the Summary of the Argument from IBM’s amicus brief:
Software is not a new technology. It has been around in various forms for well over half a century. During that time, it has become one of the fundamental building blocks of innovation and technological advancement, and a critical part of our Nation’s economy. Software is the medium for innovation in every field, from automobiles to manufacturing to medicine. The fact that the Court is now—in 2014—actively considering such a basic question as whether computer-implemented inventions such as software are even eligible for patent protection is deeply troubling.
This Court’s felt need to review whether computer-implemented inventions such as software are patent-eligible is an unmistakable indicator that § 101 jurisprudence in this area has proven unworkable. This Court has repeatedly grappled with § 101 and the judicially-created exceptions to that provision: “laws of nature, natural phenomena, and abstract ideas.” Diamond v. Diehr, 450 U.S. 175, 185 (1981). With respect to this last exception—abstract ideas—despite the Court’s best efforts, not to mention those of IBM and others that saw promise in § 101’s ability to weed out problematic patents, a workable test for when a computer-implemented invention is a patent-ineligible abstract idea has not emerged.
MY TAKE: I agree 100%, as I do with virtually everything written in the IBM brief. The law of patent eligibility is a mess because Supreme Court decisions over the past several years have lacked judicial discipline. Previous Courts have cautiously counseled to use 101 as a coarse filter, not a fine filter, because only that which is unequivocally unworthy of a patent should be weeded out prior to inquiring whether the innovation is new, nonobvious and adequately described. Still, in the Mayo case the Supreme Court conflated novelty and nonobviousness with patent eligibility.
It is mind-boggling that the Supreme Court needs to even address the patent eligibility of software. Software patents have been granted by the U.S. Patent Office since 1968, the Patent Act explicitly mentions software and computer implemented methods as patent eligible and according to the GAO somewhere between 50% and 60% of all innovations described in patent applications filed at the USPTO relate in one way or another to software.
There should be no serious question that computer-implemented inventions such as software constitute patent-eligible subject matter under § 101. Courts and amici agree that the technological functions carried out via software could also be implemented through computer hardware in the form of integrated circuits (chips) or other discrete electronic components, and the patent eligibility of those tangible devices is beyond question. Accordingly, there is no principled reason to draw distinctions between hardware and software when it comes to § 101 patent eligibility. Still, arguments that computer-implemented inventions such as software are not patent eligible persist and are proffered with increasing frequency as computer-implemented inventions become more pervasive and as innovative software brings computer implementation within the ken of many ordinary computer users.
MY TAKE: Amen! If you look at a well written patent application that appropriately describes the technology the powers a computer implemented process there will be a lot of tangible components (i.e. “hard”—ware) discussed and required in order for the computer implemented process (“soft”—ware) to actually work. Those familiar with the technical reality of software also know that any computer implemented process that can be accomplished using software can also be accomplished using logic gates. Indeed, logic gates are foundation for all digital electronic circuits and microprocessor based systems. Thus, software can be explained on its core level as a process for manipulating logic gates.
Questions concerning the patent eligibility of even the most technologically innovative software are being raised as a consequence of the unworkability of the abstract idea doctrine as applied to computer- implemented inventions. Drawing on this Court’s case law, courts and commentators have repeatedly expressed concern that a patentee can “preempt” an entire field or fundamental idea by claiming the implementation of a well-worn concept on a computer. See Diehr, 450 U.S. at 187 (“Their process admittedly employs a well-known mathematical equation, but they do not seek to pre-empt the use of that equation.”); Parker v. Flook, 437 U.S. 584, 589 (1978) (expressing concerns with claims that would “‘wholly preempt’” a field); Gottschalk v. Benson, 409 U.S. 63, 72 (1972) (“[I]f the judgment below is affirmed, the patent would wholly pre-empt the mathematical formula and in practical effect would be a patent on the algorithm itself.”); see also Bilski v. Kappos, 130 S. Ct. 3218, 3231 (2010) (concluding that the disputed claims “would pre-empt [risk hedging] in all fields, and would effectively grant a monopoly over an abstract idea”). In many cases, courts appear to have adopted the view that implementation on a computer has become so commonplace as to be technologically meaningless and that, as a result, a patentee claiming the implementation of an abstract idea on a computer is claiming nothing more than the abstract idea itself.
MY TAKE: The reason the abstract idea doctrine is unworkable is because the Supreme Court has never defined what is an abstract idea. The Supreme Court has treated the term “abstract idea” much as they have the term “obscenity”; they know it when they see it. Such a level of subjectivity leads to chaos, which is exactly how the Judges on the Federal Circuit can manage to find themselves evenly split on the issue of whether software is patent eligible. The Supreme Court abhors bright line rules unless they are the ones who announce them. Such an irrational fear of certainty and predictability is curious given how those concepts are so fundamentally important to a functioning judicial system. Still, if they don’t like bright line rules that everyone can follow as announced by the Federal Circuit they at least owe us a workable test that they are willing to endorse.
While it may be true that the steps required to implement an abstract idea on a computer are well established and thus likely within the capability of one skilled in the art, the implementation of that idea on a computer, however minimal the effort, definitively separates that specific technological implementation from the idea itself. Technology has enabled an astounding diminishment of the man-machine interface, but it has not erased the need for that interface and, as a result, even the most problematic computer-implemented invention will be meaningfully different from and narrower than the associated abstract idea. Thus, computer- implemented inventions will always fall outside the abstract idea exception and are patent eligible under § 101.
To be sure, patents that do no more than implement an otherwise abstract idea on a computer through conventional programming raise concerns reminiscent of those identified in this Court’s § 101 jurisprudence, and such patents are part and parcel of a larger challenge facing our Nation’s patent system. Patentees have all too often obtained patent protection for very straightforward ideas by simply stating in their patent application that the idea will be executed in a specific technological medium. See, e.g., Bilski, 130 S. Ct. at 3230 (the “prohibition against patenting abstract ideas ‘cannot be circumvented by attempting to limit the use of the formula to a particular technological environment’” (quoting Diehr, 450 U.S. at 191-92)). And patentees have repeatedly been able to secure overbroad patents that threaten to undermine innovation. See Mayo Collaborative Servs. v. Prometheus Labs., Inc., 132 S. Ct. 1289, 1303 (2012) (discussing concerns with foreclosing future innovation).
MY TAKE: On page 22 of the IBM brief Mayo v. Prometheus is taken on more directly. It is true that virtually everyone knowledgeable about the claimed “innovation” in that case would have found the claims invalid, but invalid either under 102 for lack of novelty or under 103 because the claimed invention was obvious. So yes, there is a justifiable concern about foreclosing future innovation, but that is NOT a topic appropriately addressed under 101. The question under 101 should be whether there IS an invention. If there is an invention with the requisite tangible description then the law should force the analysis to push forward to 102, 103 and 112. A tangible requirement to the claimed invention works to protect against the patenting of ideas. The fact that an invention is a pioneering invention, fundamentally important or widely used is of no importance under 101. I would think the Supreme Court would understand this after their decision in KSR, which presumably wanted to discourage trivial contributions from being patented. One would have thought by discouraging the trivial the Supreme Court was exalting the foundational, pioneering invention.
But as IBM’s day-to-day real-world experience has repeatedly confirmed, a muscular application of the abstract idea doctrine is not a useful tool for addressing these problems in the computer-implemented invention context. As this Court has recognized, “all inventions at some level embody” an abstract idea. Mayo, 132 S. Ct. at 1293. And unlike laws of nature and natural phenomena, abstract ideas defy the demarcation of meaningful boundaries that make the development of clear and administrable rules possible. As a result, efforts to apply the abstract idea doctrine to computer-implemented inventions quickly devolve into an undisciplined parsing and rewriting of the relevant claims such that courts end up evaluating a claim of their own making—not what the inventor actually claimed. See 35 U.S.C. § 112(b) (the “inventor” must “particularly point out and distinctly claim the subject matter which” he “regards as the invention” (emphasis added)). Moreover, the unwieldy nature of the abstract idea doctrine makes it impossible to apply that doctrine to computer-implemented inventions as a “threshold test.” Bilski, 130 S. Ct. at 3225. Every patent implements some abstract idea and attempting to isolate and assess the extent to which a computer-implemented invention “preempts” or “monopolizes” an abstract idea will always require more than the quick look that § 101 contemplates.
While the prospect of preempting abstract concepts through their implementation on a computer understandably leads courts to ask § 101-type questions, the answer to this § 101-type concern actually lies in § 103’s bar on obvious patents. The very same innovations that have eroded the man-machine interface and made programming a much more accessible feat than it once was have also made it far more likely that the computer implementation of an idea would be readily within the grasp of skilled programmers and thus obvious.
MY TAKE: I believe 103 is a far better focus than 101. There is no intellectually honest way to distinguish between inventions that are capable of being patented and inventions that are not capable of being patented. Prior to the confusion ushered in by ill-considered Supreme Court decisions relative to 101, the law could be summarized as this: If you have an invention it is patent eligible.
Ideas, laws of nature and natural phenomena are not inventions. If you have an invention, such as a computer implemented process it is and properly should be patent eligible. Analysis should push forward primarily to 112 to first see if the invention has been adequately described. That is why I have said that Judge Rich was largely correct in announcing the rule in State Street, which focused on whether the method was concrete and tangible. If there is a concrete and tangible description that fully describes the underlying technology then 101 and 112 should be satisfied. Then attention should move to 102, which likely won’t be an issue in most instances, and on to 103 where the trivial will be weeded out.
In addition to avoiding the pitfalls of applying the abstract idea doctrine in an area where it is has proved unworkable, focusing on obviousness when evaluating computer-implemented inventions has much to recommend it. The test for whether a patent is invalid for obviousness under § 103 is clear and easy to apply. See KSR Int’l Co. v. Teleflex Inc., 550 U.S. 398, 402, 406 (2007). First, courts and patent examiners must determine the scope and content of the relevant prior art. Second, the level of skill of the ordinary artisan must be identified. Third, the scope and content of the claimed invention must be defined through the process of claim construction. See KSR, 550 U.S. at 406. And finally, with that information in hand, courts must determine whether § 103’s requirements have been met.
MY TAKE: When I first read this paragraph I cringed. The test for obviousness is clear and easy to apply? In what universe?
What IBM describes as clear and easy to apply is really the test announced in Graham v. John Deere, which is of course true. The obviousness framework is clear and easy to understand, but KSR has allowed some patent examiners and some Judges to practically do whatever they want because the KSR test, if you can call it that, is subjective and can be easily manipulated to achieve a predetermined result. When KSR is coupled with the Federal Circuit’s love of de novo review there is no real certainty with respect to obviousness. I do agree, however, that 103 is the far better place to determine patentability. 101 is not supposed to weed out inventions.
The prior art analysis required under § 103 entails construing and evaluating the actual claim and each claim element, thus ensuring a focused inquiry into the actual invention at issue (as opposed to an unjustified parsing and effective rewriting of the claims). Relatedly, while continually evaluating patents under § 101 does nothing to combat problems with broadly drafted patents, further developing § 103 doctrine can play a critical role in increasing patent quality and can also assist in keeping any inquiry into “inventiveness” where it belongs. And perhaps most importantly, the more fine-grained analysis that § 103 mandates ensures that the patent-eligibility of broad swaths of technology will not be called into question, fostering the certainty that investment in innovation demands.
MY TAKE: There is little doubt that obviousness is the far more appropriate way to handle computer-implemented inventions. The true test for patentability should not rest on 101. The Patent Statute has different sections that should each be allowed to function. Short-circuiting the work of the analysis and ramrodding it into all into 101 ignores the law even if that is what the Supreme Court did in Mayo.
If computer implemented processes are not patentable then 50% to 60% of all disclosed innovations will wind up being patent ineligible in some respect. That will drive innovation under ground, which defeats the Constitutional purpose of the patent system. With innovation driven underground those who come later will not able to stand on the shoulders of those who came earlier. Therefore, the Supreme Court runs the real risk of making a decision that plainly and clearly frustrates the Constitutional purpose of a patent system.