Home > Insights > Patentability of simulations – Decision of the Enlarged Board of Appeal for G1/19
12 March, 2021

The thorny issue of the patentability (or otherwise) of inventions involving computer simulations at the EPO has been provided with an ending, of sorts, with the Enlarged Board of Appeal (EBoA) issuing their decision for referral G1/19. This is the culmination of the referral which began in 2019, with three questions on this issue having been referred to the EBoA by a Board of Appeal which had wanted, potentially, to depart from the then established practice. Whilst this decision by the EBoA does restore a degree of legal certainty, it seems to represent a course change (or at least a course correction), the details of which will likely be developed in the coming years.

As discussed in our previous bulletin, in decision T0489/14 from February 2019, an EPO Board of Appeal referred three questions, as referral G1/19, to the EBoA regarding the patentability of inventions involving computer-implemented simulations. The Board in T0489/14 made this referral as they appeared to disagree with the more established approach set out in T1227/05 (Infineon) for assessing patentability of simulations, and they also seemed minded that further criteria for patentability were necessary. A hearing was held in July 2020 (discussed in our bulletin of the time) which was notable for being the first EBoA hearing to be web-streamed – a development due in no small part to the extraordinary circumstances of 2020.

To set the scene, and as a brief reminder for those that have been following this referral, the questions referred to the EBoA were:

(Q1) In the assessment of inventive step, can the computer-implemented simulation of a technical system or process solve a technical problem by producing a technical effect which goes beyond the simulation’s implementation on a computer, if the computer-implemented simulation is claimed as such?

(Q2) If the answer to the first question is yes, what are the relevant criteria for assessing whether a computer-implemented simulation claimed as such solves a technical problem? In particular, is it a sufficient condition that the simulation is based, at least in part, on technical principles underlying the simulated system or process?

(Q3) What are the answers to the first and second questions if the computer-implemented simulation is claimed as part of a design process, in particular for verifying a design?

As always in referrals to the EBoA, the first issue is whether any of the questions are going to be answered at all – the EBoA will only consider questions aimed at ensuring uniform application of the law, or to answer a point of law of fundamental importance.

As seemed likely from the EBoA’s preliminary opinion and the course of the hearing last July, the EBoA considered both Q1 and Q3 as admissible, without much discussion.

The EBoA themselves had divided Q2 into a part (a), the first sentence, and a part (b), the second sentence. In the decision, the EBoA have been explicit that providing an exhaustive list of criteria for assessing whether a technical problem is solved is neither possible nor desirable. Indeed, throughout the reasoning the EBoA have been at pains to point out that what is considered “technical” may change over time “to accommodate new technical or scientific developments or to reflect societal changes”. Consequently, the EBoA decided that Q2(a) was not admissible, whereas the more specific Q2(b) perhaps was. In the end the EBoA reworded Q2 as:

(Q2) For the assessment of whether a computer-implemented simulation claimed as such solves a technical problem, is it a sufficient condition that the simulation is based, at least in part, on technical principles underlying the simulated system or process?

It was this reworded Q2 the EBoA chose to admit.

The answers
After a long and detailed discussion that forms a large part of the written decision, the EBoA came to the following overt answers to the admitted questions:

(A1) A computer-implemented simulation of a technical system or process that is claimed as such can, for the purpose of assessing inventive step, solve a technical problem by producing a technical effect going beyond the simulation’s implementation on a computer.

(A2) For that assessment it is not a sufficient condition that the simulation is based, in whole or in part, on technical principles underlying the simulated system or process.

(A3) The answers to the first and second questions are no different if the computer-implemented simulation is claimed as part of a design process, in particular for verifying a design.

On their face these answers do not seem particularly controversial, and indeed should not come as a surprise given the EPO’s mantra regarding computer-implemented inventions in recent years. The EPO’s settled view has been that a computer-implemented invention is a computer-implemented invention. There are no special cases or classes and they are all to be treated in the same way using the same framework, established in the COMVIK decision (T641/00) of 2002. This is the same whether the invention relates to the use of a blockchain for financial transactions, the use of artificial intelligence for classification of digital images, or the computer control of a nuclear power station.

In this spirit the EBoA here pointed out a number of times that it would not be correct to afford computer simulations a special status within the computer-implemented invention regime – they can neither “hold a privileged position within the wider group of computer-implemented inventions” nor be “discriminated against without any legal basis”. As such, with these answers, the EBoA has reaffirmed that the COMVIK approach (as outlined in the EPO’s Guidelines G-VII,5.4) must be used for assessing computer-implemented inventions involving simulations in the same way it is used for any other computer-implemented invention.

The end of the story?
From the answers given by the EBoA, one would be forgiven for thinking that nothing has changed – that existing practice has been confirmed by the EBoA and the planned divergence of the referring Board of Appeal in T0489/14 dismissed. Unfortunately, however, this is not the case and the discussions given by the EBoA preceding the answers contains some potentially significant changes in the field of simulations.

The COMVIK approach adopted by the EPO and the current EBoA requires that an invention provide a technical solution to a technical problem. When it comes to applying COMVIK to computer-implemented simulations, the leading case had been T1227/05 (Infineon), which held that such technicality could be conferred if the simulation was of “an adequately defined class of technical items” or for “an adequately defined technical purpose” provided that the claimed method was “functionally limited to that technical purpose”. In that case, the simulation of a circuit subject to 1/f noise was held to constitute such an adequately defined technical purpose, and was therefore patentable.

This had the effect that simulations of technical systems to calculate a value or property of said system may be patentable in and of themselves.

The referring Board of Appeal in T0489/14 had sought to do away with this, and instead re-introduce an older idea that, in order for the claimed solution to be technical, a “direct link with (external) physical reality” must be present. The EBoA perhaps wisely rejected this, indicating that it was an often sufficient, but not necessary criterion, thus avoiding the potentially far reaching consequences to hitherto accepted practice regarding other classes of computer-implemented inventions.

Nevertheless, the EBoA was clearly receptive to the idea that T1227/05 represented an exception rather than the rule. Whilst the EBoA stopped short of saying that the Board of Appeal in T1227/05 came to the wrong decision, since other factors for patentability were discussed in T1227/05, they have drastically limited the applicability of the decision of T1227/05 stating:

“In the Enlarged Board’s view, calculated numerical data reflecting the physical behaviour of a system modelled in a computer usually cannot establish the technical character of an invention in accordance with the COMVIK approach, even if the calculated behaviour adequately reflects the behaviour of a real system underlying the simulation.” (emphasis added)

The EBoA went further in stating that only “in exceptional cases” may such “calculated effects” be considered technical effects, such as if the use of such data is limited to an accepted technical purpose. The EBoA have stated that the criterion of T1227/05 listed above “should not be taken as a generally applicable criterion”. Thus, the act of calculating a value or property of a simulated system, in and of itself, is likely no longer be patentable – but this will very much to depend on the circumstances. In particular, the EBoA were keen to emphasise that this is regardless of whether the simulation is of a system that may be thought overtly technical such as a nuclear reactor, or overtly non-technical such as a marketing campaign – the key will often be how the simulation is used in the context of the rest of the claimed subject-matter.

On a more positive note the approach taken by the EBoA left relatively untouched the idea that an improvement in a simulation technique rendering it better suited for particular hardware could render the improved simulation technique a technical invention.

The concerns that a decision in G1/19 may have wider, perhaps unintended, consequences for computer-implemented inventions in general at the EPO do not appear to have been realized – and it would seem the decision is mainly limited to such inventions involving simulations.

For such inventions, however, the door that had been opened by T1227/05 has been to a large degree shut. It is no longer appropriate to be asking whether the simulation relates to “an adequately defined class of technical items” or is for “an adequately defined technical purpose”. Instead, unless the claimed invention is focussed on specifics of how a simulation is to be implemented (e.g. using specific techniques for specific hardware so as to improve processing efficiency), it seems that in many cases explicit (or implicit) steps of actually using the output of a simulation in some “technical” manner may be insisted upon to ensure the claim is allowable. The disadvantages brought about by including such steps will clearly be more of a concern in some applications than in others.

The EBoA may be said to have chosen a path somewhere between the positions of the referring Board of Appeal on the one side, and the appellant and the President of the EPO on the other. This necessarily leaves a degree of uncertainty that will only be lifted by future decisions of the Boards of Appeal. The legal certainty wanted by applicants of computer-implemented inventions involving simulations has, apparently, not arrived quite yet.

The extent to which national patent offices will follow the approach of G1/19 has, of course, yet to be seen. It may, therefore, be worthwhile considering broader filing strategies, including direct national patent applications, for important inventions centred on simulations.

A number of cases since 2019 touching on these issues have had their proceedings stayed in anticipation of this decision. These should now go ahead and will prove the first test cases for quite how the reasoning espoused by the EBoA will be applied in practice. However, whilst this area of law continues to develop, care should be taken when drafting new applications towards new simulation methods so that they are not reliant on the reasoning of T1227/05, and adequate detail is given regarding the use of simulation outputs. Of course, your usual Boult Wade Tennant advisor will be able to advise you regarding this.