Friday, September 11, 2020

CIPO's Approach to Patentability of Computer-Implemented Inventions

 Choueifaty v Canada (Attorney General) 2020 FC 837 Zinn J rev’g and remanding CD 1478

            Application 2,635,393

The vexed issue of patentable subject matter has reared its head once again. Choueifaty applied for a patent for a computer-implemented method for selecting an investment portfolio with the lowest level of risk for a given return [CD 25]; in brief, a computer-implemented business method. CIPO, applying its problem-solution approach to claim construction set out in MOPOP 12.02.02e [13], determined that the “essential elements” of the invention “are directed to a scheme or rules involving mere calculations used to construct the anti-benchmark portfolio and thus not directed to patentable subject matter” [CD 52], [16]. In CIPO’s view, the computer itself was not an essential element; had it been, the claim would have been allowed [17]: PN 2013-03.

In a brief decision, Zinn J reversed on the basis that the problem-solution approach is not the correct way to determine the essential elements of the claim; rather, the approach set out by the SCC in Whirlpool 2000 SCC 67 and Free World 2000 SCC 66, must be used [40]. CIPO relied on Genencor 2008 FC 608 for the proposition that the Whirlpool test is not applicable to patent examiners [34]. In light of the subsequent FCA decision in Amazon 2011 FCA 328, [43] which expressly held that CIPO must use the Whirlpool approach, Zinn J held that Genencor “is no longer good law” [35]. Zinn J therefore remitted the application to the Commissioner for reassessment in accordance with his reasons.

In my view, Zinn J’s decision is entirely correct so far as it goes, but in this post I want to step back and take a brief look at the bigger picture.

Ultimately, the problem is that CIPO is trying to create a prohibition on patenting of business methods (and certain diagnostic methods). There is no clear basis in the case law for such a prohibition, and certainly there is none in the Act, so it appears that CIPO is trying to work some sleight of hand with s 27(8), which provides that “No patent shall be granted for any mere scientific principle or abstract theorem.” But on its face, this applies broadly, to any kind of invention, not just business methods. So how to create a specific connection, to justify striking down only business method patents?

The Practice Notice Respecting Computer-Implemented Inventions PN 2013-03 states that “disembodied inventions” are unpatentable, including “inventions where the claimed subject-matter is a mere idea, scheme, plan or set of rules.” This is correct, to the extent it reflects s 27(8) (the Practice Notice does not actually cite s 27(8) for this proposition). The Practice Notice goes on to say that “Where. . . it is determined that the essential elements of a construed claim are limited to matter excluded from the definition of invention (as noted above), the claim is not compliant with section 2 of the Patent Act, and consequently, not patentable.” This sounds plausible, but note the shift from “the claimed subject matter” to “the essential elements.” If the statement were that “Where the claimed subject matter is excluded from s 2, the claim is not patentable,” as with, eg a claim to “A transgenic mouse,” this statement would be clear law. The focus on the essential elements of the invention, rather than the claimed invention, is a key move that sets up the final element, determination of what constitutes the “essential elements.” In CIPO’s approach, identifying the essential elements boils down to identifying the inventive concept: see [37], and see MOPOP 12.02.02d says “The common general knowledge in the art provides the baseline of information to which the description is expected to add.”

Stepping back from the details, the logic of this approach is that abstract ideas are unpatentable, and if the essence of the invention is an abstract idea, then it is unpatentable. That conclusion, however, is wrong, wrong, wrong (and also wrong). An abstract idea is not patentable as such, but an abstract idea absolutely can provide the inventive concept for a valid patent. That is the core holding of Shell Oil [1982] 2 SCR 536, 554:

 A disembodied idea is not per se patentable. But it will be patentable if it has a method of practical application.

Thus it is perfectly clear law that an abstract idea can supply the inventive concept that makes the invention new and non-obvious: Tye-Sil (1991) 35 CPR(3d) 350 (FCA) provides another useful discussion. For a recent example, see Georgetown Rail 2018 FC 70 (here) where the patented method and system for inspecting railroad track were inventive “only in respect of their algorithms” [129]. In s 27(8), the words “mere” and “abstract” are crucial. That provision does not say that there can be no patent for a theorem, on the view that theorems are abstract; the prohibition is directed to “abstract theorems”; and a theorem or scientific principle that is not abstract, but is claimed as a method of practical application, is patentable.

CIPO’s analysis gains superficial plausibility by playing on different meanings of the term “invention.” Sometimes “invention” means a particular embodiment of the claimed subject matter. More to the point for our present purposes, it is common to speak of the inventive concept as being “the invention”; after all, it is the inventive concept that is the patentee’s contribution to knowledge that justifies the reward of a patent. And it is also common to speak of the invention as being the subject matter defined by the claim. So, while it is fair enough to say that the inventive concept is the invention and it is also true that the invention (as claimed) cannot be an abstract idea, these statements implicate different senses of “invention.” Consequently, it does not follow that the patent is directed to unpatentable subject matter if the inventive concept is an abstract idea. The patent is directed to unpatentable subject matter pursuant to s 27(8) only if the invention as claimed is an abstract idea. A claim to “F=ma” would be unpatentable, but a claim to a method of aiming a cannon would be patentable, even if the only inventive step involved in the method was the discovery that F=ma.

Indeed, the inventive concept is always an abstract idea. As Lord Hoffmann pointed out in Merrell Dow [1995] UKHL 14, [28], “An invention is a piece of information,” and in Shell Oil at 549, the SCC noted that an invention is “the practical embodiment of the new knowledge.” A computer related invention that, for example, makes the computer run faster – which CIPO considers patentable – is no different. The invention in that case is not the computer itself, which is at best an embodiment, it is the information as to how to arrange the architecture for greater speed, and that is abstract information.

So, yes, CIPO’s approach is wrong because it fails to use the legally required Whirlpool / Free World method of determining the essential elements. But, less obviously but perhaps more importantly, the approach is wrong because it is inconsistent with Shell Oil; by equating the invention with the essential elements which are in turn equated with the inventive concept, CIPO has avoided acknowledging the rule of Shell Oil that a “disembodied idea” is indeed patentable if the claim is to a practical application of that idea.

The result of this approach is that CIPO has smuggled a prohibition on patenting of business methods into s 2. CIPO claims that this is not true:

[53] The Applicant submitted in the RPR at pages 7-8 that there is no Canadian jurisprudence that determines conclusively that a business method cannot be patentable business matter.

[54] We agree. However, our analysis above shows that the essential elements of the claims on file are directed to a scheme or rules involving mere calculations used to construct the anti-benchmark portfolio, not that they are directed to a business method. According to PN 2013-03, such schemes or rules are not patentable subject matter.

The argument is that the claims are unpatentable not because they are directed to a business method, but because they are directed to an abstract idea. (I’ll ignore the point that the prohibition of patenting “schemes or rules” has no particular basis in Canadian law, and appears to be borrowed from s 1(1) of the UK Act.) However, as Frankfurter J pointed out in his concurrence in Funk Bros 333 US 127, 135:

Everything that happens may be deemed “the work of nature,” and any patentable composite exemplifies in its properties “the laws of nature.” Arguments drawn from such terms for ascertaining patentability could fairly be employed to challenge almost every patent.

This is exactly what CIPO is doing. Because the inventive concept is always abstract, if CIPO wants to introduce a new substantive subject matter prohibition, it can focus on the inventive concept and concludes that it is abstract and therefore unpatentable. If it doesn’t want to create a new prohibition, it emphasizes the full claim wording and its practical applications.

Specifically, with PN 2013-03, CIPO has targeted computer-implemented inventions, which includes many business method patents, such as the application at issue in this case, and, with essentially the same logic, certain kinds of diagnostic methods.

I don’t know why CIPO has taken this approach. Perhaps CIPO has been influenced by the subject matter analysis provided by US examiners. Unfortunately, the US law of patentable subject matter is a train wreck of historic proportions, which is, moreover, incompatible with Canadian law: see my article, The Rule Against Abstract Claims: A Critical Perspective on US Jurisprudence, (2011) 27 CIPR 3. As Phelan J pointed out in, 2010 FC 1011 [32], “it can be useful to look to other regimes. However, this must be done mindfully.” He noted that:

 [33] The Commissioner relied heavily on foreign jurisprudence throughout her decision, particularly on that of the United Kingdom, Europe and the United States. This becomes troubling and even problematic when she ignores fundamental differences between the foreign and the domestic regimes, or ignores Canadian legal principles altogether. Specifically, her reliance on English and European systems does not take into account that both, in implementing the European Patent Convention (EPC), have fundamentally different legislation than Canada for determining patentable subject matter.

It would seem that the Commissioner did not take Phelan J’s words to heart.

Alternatively, maybe CIPO believes that by refusing to grant patents for business methods, they are helping promote innovation in Canada. And indeed, there is a plausible argument that business method patents are bad for innovation: see my article Business Method Patents and Patent Trolls, (2013) 54 Can Bus L J 38 for a review of the literature. But this is a dangerous game. The question of what subject matter should be patentable is a very difficult matter of policy, with no clear answer. Patents incentivize innovation, but that incentive comes at a price. That price is clearly not worth paying for things we already have; hence the “odious monopolies” of Elizabethan times on common articles such as salt and playing cards. But neither is the patent price worth paying for things we would have had even in the absence of the patent incentive, as a result of the ordinary march of technology. The “but for” theory of obviousness set out by the USSC in Graham v John Deere 383 US 1, 11 (1966) addresses this problem:

 [T]he underlying policy of the patent system [is] that "the things which are worth to the public the embarrassment of an exclusive patent" ... must outweigh the restrictive effect of the limited patent monopoly. The inherent problem was to develop some means of weeding out those inventions which would not be disclosed or devised but for the inducement of a patent.

If the obviousness threshold were such that patents were never granted except for inventions that would not have been developed without the lure of a patent, then patents would always be good, since having a useful product at a high price is better than not having it at all. However, the obviousness requirement in practice certainly does not have such perfect effect, both because the implicit counterfactual is far too difficult, and because the black letter law – would the invention have been obvious to a person skilled in the art – does not even purport to implement a “but for” requirement. Moreover, patent examiners do not have time to conduct an exhaustive examination of every application. This is not to criticize obviousness doctrine—the problem is inherently very difficult, and an imperfect solution is all we can aspire to—but a realistic policy analysis must recognize that the obviousness requirement is indeed imperfect as a means of prohibiting patents for inventions that we would have had in any event. Other doctrinal considerations, such as the granting of injunctive relief, and the cost of the litigation system, will also affect the costs and benefits of the patent system.

Another consideration is there are many non-patent incentives to innovate, such as a first-mover advantage, reputation, prizes, trade secrets, etc, which depend on the nature of the technology and the structure of the market and the industry. Whether we would have had a particular invention without the lure of a patent depends in part on how large the non-patent incentives are. Consequently, we know that the patent incentive is more important in some fields that others. Patents are generally considered to be very important in pharmaceuticals, chemical and medical devices, but less important in, for example, the tech sector. If the obviousness threshold is too low, so that many patents are granted for inventions that would have been developed in any event, and non-patent incentives are substantial, it might be a net boost to innovation to abolish patents entirely in some particular field. So, in Patent Failure, Bessen & Meurer argue, on the basis of the available empirical evidence, that patents work in the chemical and pharmaceutical fields, but “the average public firm outside the chemical and pharmaceutical industries would be better off if patents did not exist” (16). I won’t say that I agree with Bessen & Meurer, but I won’t disagree either; the problem is a difficult empirical one, and there is no easy answer. Consequently, it is quite possible that business method patents are a net drag on innovation; but it is also possible that they would instead provide a spur to innovation in that sector. Further, even if it is true that business method patents would be a net drag on business innovation, it is possible that the patents in the tech sector are just as bad. (For what it’s worth, my guess, and it is no more than that, is that business methods patents are about as bad, or as good, as patents in the tech sector.)

The important point from all of this is that the question of whether patents are a net benefit to innovation in a particular field depends on the interplay of a variety of doctrinal considerations, including obviousness doctrine, remedies, and procedure, with the non-patent incentives that turn on the nature of the product, the market and the industry. This is truly an extremely difficult problem. Even assuming that CIPO’s antipathy to business method patents is based on innovation policy considerations, it is clearly not within CIPO’s competence to set policy on this issue. As discussed here, CIPO uses the same kind of analysis to deny patents to diagnostic methods where the inventive concept is the discovery of a new correlation, while granting patents where the inventive concept is the discovery of a new analyte. Despite the difficulty, just discussed, of assessing whether patents are a net benefit in any particular field, I will go out on a limb and say while there is room to debate the need for business method patents, CIPO’s position on diagnostic methods is clearly just wrong from an innovation policy perspective; diagnostic methods belong with pharmaceuticals and chemicals as an area where the patent system is beneficial to innovation.

I don’t know whether CIPO’s position on computer-implemented inventions and diagnostic methods is a result of doctrinal confusion or policy confusion, but it is certainly entirely unsound on either basis.

I could go on and on, but I already have, in other places. For more on the issue, see the following (links are to draft versions):

Diagnostic Methods at CIPO

• The Rule Against Abstract Claims: A Critical Perspective on US Jurisprudence, (2011) 27 Canadian Intellectual Property Review 3 - 30

• The Rule Against Abstract Claims: History and Principles, (2011) 26 Canadian Intellectual Property Review 205 - 229

• The Structure of the Law of Patentable Subject Matter, (2011) 23 Intellectual Property Journal 169 - 204

No comments:

Post a Comment