03 August 2022

Eligibility of Computer-Implemented Inventions Behind Unprecedented Numbers of Patent Office Rulings

image Once upon a time – not so very long ago, in fact – it was rare for the Australian Patent Office to issue a formal published ruling on the patent-eligibility of claims submitted for examination.  Indeed, ex parte decisions (i.e. those involving only the applicant and the Office) were generally in the minority, and most of those related to pharmaceutical extensions of patent term, allowability of amendments, and extensions of time to meet various deadlines.  Historically, the overwhelming majority of decisions have related to inter partes proceedings, such as patent oppositions.  That is, however, no longer true.  Following a peak in 2015, inter partes decisions have been falling, while ex parte decisions have been generally on the rise since 2011.  In 2021, ex parte decisions outnumbered inter partes decisions for the first time.  And in 2022 more ex parte decisions have been issued so far (i.e. up until the end of July) than in any past full year, with inter partes decisions once again lagging behind.

The sole driver of the growth in ex parte decisions has been patent-eligibility according to the ‘manner of manufacture’ test under Australian patent law.  The number of published Patent Office decisions relating to subject-matter eligibility of patent claims has risen from fewer than one per year in the 2000-2009 decade to 19 decisions in 2021 and 21 decisions in just the first seven months of 2022.  Since 2010 there have been 115 published decisions on eligibility, with the claims at issue being found ineligible in 100 of those cases.  Almost all of these have related to computer-implemented inventions.

The issue here is not that some subject matter is ineligible for patent protection.  That has always been, and will always be, true.  But the massive increase in published decisions is indicative of a more insidious problem.  Applicants rarely request hearings to appeal rejections by a patent examiner, and when they do it is usually because they genuinely believe that the examiner has got it wrong, and that the additional effort and expense is justified by good prospects of a better outcome when the matter is considered by an experienced hearing officer.  The unprecedented rise in applicants requesting hearings on patent-eligibility is due to a lack of clarity and coherence in the law.  And the fact that those applicants are so frequently wasting their time and money reflects the fact that the Patent Office has been championing an interpretation of the law that is at odds with the way in which it is being interpreted by the applicants and the patent attorneys advising them.

To be clear, the Patent Office on the one hand, and applicants and their attorneys on the other, are looking at exactly the same case law, and arriving at completely different conclusions as to how that law applies to particular claims.  Over the period during which this situation has arisen, there have been five Full Bench decisions of the Federal Court of Australia that should have served to clarify the law, but which appear to have had the opposite effect.  (For the record, those decisions are Research Affiliates LLC v Commissioner of Patents [2014] FCAFC 150, Commissioner of Patents v RPL Central Pty Ltd [2015] FCAFC 177, Encompass Corporation Pty Ltd v InfoTrack Pty Ltd [2019] FCAFC 161, Commissioner of Patents v Rokt Pte Ltd [2020] FCAFC 86, and Commissioner of Patents v Aristocrat Technologies Australia Pty Ltd [2021] FCAFC 202.)

The Aristocrat decision is currently before the High Court, where it was heard on 9 and 10 June 2022 (transcripts at Aristocrat Technologies Australia Pty Ltd v Commissioner of Patents [2022] HCATrans 103 and [2022] HCATrans 104).  While it might be hoped that the High Court will provide the clarity that has been lacking in the Federal Court decisions, I am not so optimistic.  Looking at what has been happening at the Patent Office over the past few years, I fear that whatever the High Court may have to say about the specific claims at issue in Aristocrat will be equally open to different interpretations when applied to different claims in other cases.

Inter Partes vs Ex Parte Decision Numbers

The following chart summarises the number of published Australian Patent Office decisions each year between 2000 and 2022 (year-to-date).  These are broken down into three classes of decisions: inter partes decisions (mostly substantive patent oppositions); ex parte decisions; and decisions relating to the Patent Office’s short-lived practice of divisional case management (‘Divisional CM’).

Number of Australian Patent Office Published Decisions, 2000-2022

A peak in inter partes decisions between 2014 and 2016 is due in large part to changes in patent opposition procedures brought in by the Raising the Bar reforms from 2013.  These reforms included new regulations governing extensions of time in opposition proceedings which resulted in new oppositions being completed more expeditiously, and accelerated a backlog of ongoing opposition proceedings.

The ‘divisional case management’ practice was designed to limit the ability of applicants to keep applications pending for extended periods by filing divisional applications, and then continuing to delay further active prosecution.  It was particularly relevant under older rules, which enabled applications to remain pending for over two years without any substantive action being taken by the applicant.  The practice was largely rendered obsolete by shortened deadlines introduced by the Raising the Bar reforms, and was discontinued, aside from exceptional cases, in 2013.

The rise in ex parte decisions from 2011 onwards is, as I have already noted, almost entirely due to applicants requesting to be heard in relation to examination objections on grounds of subject-matter ineligibility, i.e. lack of ‘manner of manufacture’.

The Rise and Rise of Patent-Eligibility Decisions

Throughout the 2000’s there were just four patent-eligibility decisions issued by the Patent Office.  These included the 2004 Stephen John Grant decision that was ultimately appealed to a Full Bench of the Federal Court of Australia.  The Full Court’s judgment in Grant v Commissioner of Patents [2006] FCAFC 120 confirmed the established position that abstract ideas, including business and legal schemes, do not constitute patentable subject matter, and that ‘[a] physical effect in the sense of a concrete effect or phenomenon or manifestation or transformation is required.’ 

For a while, the Grant decision appeared to have settled the law of patent-eligibility.  For the remainder of the decade there was just one further decision relating to patentable subject matter, and that was in relation to an application by a self-represented applicant asserting the discovery of radiation by antimatter which ‘is an intrinsic component of God's dust of creation, the ether and the binding force of the universe that is manifest in the form of an electrosubtronic field of infinitely high voltage…’.  This application was self-evidently deficient in numerous respects, though finding a lack of patentable subject matter in the attempt to claim any application of ‘a new theory or law of science’ was presumably the easiest – and kindest – way of disposing of the matter.

But patent attorneys working in the area of computer-implemented inventions were becoming aware by 2009 that something may have changed ‘behind the scenes’ at the Patent Office.  Claims that we would previously have expected to be found eligible were increasingly being rejected as not being for a ‘manner of manufacture.’  Back in 2017 I published a statistical analysis suggesting a change in examination practice may have occurred in November 2008.  However, since formal decisions are only issued when an applicant expressly requests a hearing following repeated attempts to overcome objections in examination, these early rejections did not begin to manifest as published decisions until 2010. 

The wave of decisions began innocuously enough with Invention Pathways Pty Ltd [2010] APO 10.  This case related to a business scheme for commercialising inventions in parallel with progressing international patent protection, which would traditionally have been regarded as plainly unpatentable, but for the use of a computer to record and monitor the steps in the claimed method.  By the end of the year, however, the Patent Office had also issued decisions in Iowa Lottery [2010] APO 25 and Research Affiliates, LLC [2010] APO 31Research Affiliates led, in November 2014, to the first in the recent series of Full Federal Court appeal decisions.

As the chart below illustrates, the Full Court’s decision in Research Affiliates, and its subsequent December 2015 decision in RPL Central, did not help to clarify the law.  On the contrary, these decisions preceded a veritable explosion in patent-eligibility decisions by the Patent Office, in which applicants were overwhelmingly unsuccessful.  And a second wave of Patent Office decisions has followed the Encompass, Rokt and Aristocrat Full Court decisions issued between 2019 and 2021.

Australian Patent Office Published Eligibility Decisions, 2000-2022

Can Recent Eligibility Decisions be Reconciled?

As time has passed the types of computer-implemented inventions that have been found to be ineligible have become increasingly sophisticated.  ‘Early’ cases, such as Invention Pathways and Research Affiliates, involved claims directed to otherwise ineligible business schemes in which ordinary general-purpose computer hardware and software was used merely to record and retrieve information used in the claimed method.  More recent cases, however, involve vastly more sophisticated uses of computer hardware and software technology, resulting outcomes that can appear difficult to reconcile with one another.

For example, Commonwealth Scientific and Industrial Research Organisation [2021] APO 43 (CSIRO) concerned an application for an invention that employs known smart contract and blockchain technology in a novel manner to monitor compliance with a manufacturing process involving multiple parties, without disclosing the overall process definition to all parties.  In this case, the Hearing Officer found that transforming a process definition into a numerical representation subject to homomorphic encryption rendered the claimed invention patent-eligible.  In reaching this conclusion, he rejected the examiner’s argument that the invention merely applied well-known techniques of creating numerical representations and homomorphic encryption, and stated that the examiner had ‘adopted an erroneous approach in dissecting the claims into its [sic] individual features and then assessing whether each of these features is well known in the art’ (at [35]-[36]).

On the other hand, a different conclusion was reached in Accenture Global Solutions Limited [2022] APO 49.  The invention in that case comprised a method employing multiple cryptographic steps in combination with distributed ledger (i.e. blockchain) technology to facilitate a secure transaction for transferring an asset between untrusted parties.  Just as in CSIRO, a variety of suitable cryptographic algorithms and distributed ledger technologies were admittedly known, and the ingenuity resided in combining these in a novel method to achieve the required result.  Here, the Hearing Officer concluded that ‘the claimed invention cannot be characterised as a patentable technological innovation’, but instead ‘as an unpatentable business innovation in the realm of electronic commerce’ that in substance ‘could be described as a business scheme’ (at [75]).

Having studied both of these decisions, I am unable to reconcile them.  There is nothing that is said – for or against patent-eligibility – in either decision that could not be applied to the other, if one were minded to do so.  It might be observed that CSIRO is Australia’s flagship national research organisation, whereas Accenture is a global business consultancy.  It might also be observed that CSIRO’s invention relates to managing a manufacturing process while preserving security of confidential information, while Accenture’s invention relates to managing an online transaction while protecting parties to the transaction from fraud.  But neither of these observations should be relevant to the outcome in either case.  Patent-eligibility does not depend on the identity of the applicant.  And while process management and transaction security could each be characterised as either ‘business’ or ‘technical’ problems, the question is not what type of problem is being solved, but rather whether the solution satisfies the legal requirement to be a ‘manner of manufacture’.

The same could be said of a number of other recent Patent Office decisions that are similarly difficult to reconcile.  This is a highly unsatisfactory situation.

Will the Aristocrat High Court Decision Help?

The unprecedented number of Patent Office decisions on patent-eligibility that have been issued over the past few years should be regarded as a sign that the system is not functioning as it should.  If the law were clear, and the guidance provided to examiners clearly articulated and consistently applied, then applicants would know where they stood, and examination outcomes would be predictable and rarely disputed.  That is clearly not the case at present.  Instead, applicants, the attorneys that advise them, patent examiners, and Patent Office hearing officers are looking at the same claims, and the same case law, and reaching diametrically-opposed conclusions on patent-eligibility.  The published decisions rarely favour applicants, and even when they do it is often difficult to identify consistent guiding principles distinguishing the few positive outcomes from the many unsuccessful cases.

Some responsibility for the current state of affairs lies squarely with the Australian Patent Office which, since around 2008/2009, has – rightly or wrongly – shifted the boundary dividing eligible from ineligible claims to computer-implemented inventions.  The Office perhaps feels vindicated by the outcomes in the handful of cases that have proceeded on appeal to the Federal Court, but should be concerned that the appeal decisions have failed to improve the situation.  With five cases going up to a Full Bench of the Court inside a decade – including an expanded panel of five judges in the Encompass case – it would not be unreasonable to expect that greater clarity would have been brought to the law.  This has, manifestly, not happened, with the number of disputed examination objections continuing to rise.  In its submissions to the High Court, Aristocrat has argued that, in its recent series of decisions, the Full Court has made matters worse by overlapping manner of manufacture with novelty and inventive step, creating an anomalous qualification on the principles governing patentability, placing an unjustifiable gloss on the statutory language, and fettering by contrived constraints the patent-eligibility of computer-implemented inventions.  I am inclined to agree with this analysis.

Many people are hoping that the High Court’s pending decision in the Aristocrat case will resolve the uncertainty around patent-eligibility of computer-implemented inventions, but I am sceptical.  The High Court has been presented with the specific case of an electronic gaming machine (EGM) that has been claimed as a combination of known hardware with novel programming that implements a new form of ‘feature game’.  While the Court will no doubt endeavour to contextualise this case within some more general principles of patent-eligibility, it cannot be expected to provide a bright-line rule that will suddenly clarify eligibility of all types of computer-implemented invention.  It is well-established that ‘manner of manufacture’ is to be treated as a concept for case-by-case development according to a common law methodology.  It cannot be replaced with an inflexible test or verbal formula, and the High Court will certainly decline to do so.

My fear, therefore, is that the High Court’s decision in Aristocrat – whichever way it goes – will be as open to differing interpretations and applications to other types of inventions as the recent series of Full Federal Court decisions.  Worse yet, the Court may issue multiple concurring opinions from which parties to disputes over patent-eligibility – including applicants and the Patent Office – will be able to cherry pick from different lines of reasoning to support their arguments.  This is already happening in eligibility cases based on the High Court’s multiple opinions in D'Arcy v Myriad Genetics Inc [2015] HCA 35 – in many published decisions the Patent Office relies upon particular statements in the minority opinion of Gageler and Nettle JJ, while applicants generally favour the opinion of the plurality comprising French CJ and Kiefel, Bell and Keane JJ.  Even if we can all agree on what the law is, or at least on where to find it, we will not necessarily agree on how it should be applied to new fact scenarios.

Conclusion – What is the Policy Behind Patent-Eligibility?

An issue that has not been addressed clearly in any of the Patent Office or Federal Court decisions finding various computer-implemented inventions to be unpatentable is this: what is the policy basis for excluding these inventions from patent protection?  If we believe that the patent system serves a useful purpose to foster innovation and protect investments in new technology in most fields of endeavour, then what is it about certain fields within the IT sector, such as secure digital transaction processing, that makes them exceptional?  Given that companies are clearly investing in research, development and commercialisation of these technologies, what makes these businesses less deserving of protection via the patent system than those operating in other areas?

In looking specifically at the patent-eligibility of Aristocrat’s claimed EGM, I am afraid that the High Court is unlikely to address these broader policy questions, leaving applicants – and the Patent Office – none the wiser as to the overarching principles to be applied in distinguishing between patentable and unpatentable computer-implemented inventions.


Post a Comment

Copyright © 2014
Creative Commons License
The Patentology Blog by Dr Mark A Summerfield is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Australia License.