G1/19 – Computer-implemented simulation inventions are patentable at the EPO
24
Mar
2021
Keep simulating and “COMVIK” on

The Enlarged Board of Appeal (EBoA) has handed down its eagerly awaited decision (G1/19) concerning the patentability of computer-implemented (CI) simulations. The implications of this decision are far reaching since CI simulations are used to model systems found in a wide range of fields including engineering, chemistry, biology, and even human psychology.

Ultimately, the EBoA:

  • maintained the legal status quo; deciding that the patentability of CI simulations must be assessed on a case-by-case basis in the same way as any other computer-implemented inventions – the established “COMVIK” approach is to be used.
  • elected not to define “technical”, meaning technicality will also be decided on a case-by-case basis, as has been the practice for many years.
  • implied that AI inventions would be treated in the same way.

To understand the decision and its implications, a look at how computer-implemented inventions have been assessed historically at the EPO is needed.

 

Assessment of inventive step and the problem-solution approach 

The EPO’s longstanding practice for the assessment of inventive step, irrespective of technical field, uses the problem-solution approach (PSA). The PSA allows a simple question to be answered: do the claimed technical features solve a technical problem in a non-obvious way? Although, there is emphasis placed on the “technical” features, and the “technical” problem solved by the invention, technicality is assumed to exist with most inventions. However, computer-implemented inventions are not so straightforward.

Computer-implemented inventions, at a fundamental level, may merely involve computer programs. CI simulations go further and also usually involve mathematical algorithms. Both computer programs and mathematical algorithms are, on the face of it, non-technical, according to established EPO practice.  With no technical features, no technical problem is solved and an inventive step is missing.

However, the EPO has found that computer programs and mathematical algorithms may very well be considered technical, and thus solve a technical problem (T 0641/00 (COMVIK), T 0154/04, G3/08). Computer-implemented or CI simulation inventions have been found to go beyond merely a computer running a set of instructions, and thereby to solve a technical problem. The challenge has been in showing the EPO that this is what is happening.

The existing case law has avoided setting strict rules in favour of judging each case on its own merits. That being said, context is usually key.

For example, a CI simulation claimed as part of a method of manufacturing a product links to physical reality and goes beyond a computer running a set of instructions to solve a technical problem (T 453/91, T 1227/05). In this example, whilst the EPO can see that a technical problem is solved, the patentee of such a CI simulation invention may find a claim of this sort difficult to enforce. Primary infringement will only exist if a third party actually carries out both the simulation and manufacturing steps. In absence of this, the proprietor may need to rely on secondary or contributory infringement.

Striking a balance between claim scope and patentability of CI simulations is a dilemma for would-be patentees and the EPO alike. This tension may have made a referral to the EBoA inevitable.

 

The referral

The questions referred to the EBoA were:

  1. 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?
  1. 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?
  1. 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?

Question 1 asks a fundamental question: can a CI simulation go beyond the running of a set of instructions on a computer and in so doing solve a technical problem? Question 2 was reformulated by the board to ask two sub-questions: (2a) what is necessary for a CI simulation to go beyond the running of a set of instructions on a computer and solve a technical problem; (2b) is it enough that a CI simulation models technical principles of an underlying system/device? Question 3 asks if using the CI simulation as part of a design process makes a difference to the answers to questions 1 and 2.

The answers to the above questions could identify which contexts, if any, mean that a CI simulation in isolation goes beyond the running of a set of instructions and solves a technical problem.

 

Observations

The referral generated widespread interest and the EBoA received observations from industry and professional bodies on the referred questions.

Whilst some observations stressed the economic case in favour of obtaining patents for CI simulations in isolation, the EBoA was never likely to be, and indeed was not, swayed by this since it is limited to legal interpretation of the EPC.

There was a consensus amongst the observations that the established so-called “COMVIK” approach which applies the PSA to computer-implemented inventions was relevant for the assessment of CI simulations.

Unsurprisingly, whilst a number of the observers acknowledged that a CI simulation linked to physical reality conferred technicality, they were equally keen to stress that this was not the only way to confer technicality. Two main approaches were suggested, in the observations, to confer technicality without linking CI simulations to physical reality.

In the first it was suggested that it can be assumed that the CI simulation of a technical system or process corresponds to a real system or process. To the extent the real system or process is defined in terms of technical features, the CI simulation thereof is the “virtual” equivalent and is thus defined in terms of technical features so can solve a technical problem.

In the second it was suggested that simply tethering the CI simulation to a “specific technical purpose” is indicative of the technicality of the CI simulation, such that doing so results in solving a technical problem. It was also argued that a specific technical purpose may be using the CI simulation to improve an existing design.

If either of these approaches had been accepted, it would mean that modelling any system/process by CI simulation could go beyond a computer running a set of instructions to solve a technical problem. The bar for patentability would then have been lower, something desired by many interested parties.

 

The decision

The EBoA answered the questions posed to it as follows:

  1. 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.
  1. 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.
  1. 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.

By its answer to question 1, the EBoA has confirmed that a CI simulation can indeed be more than a computer running a set of instructions and can therefore solve a technical problem. A CI simulation can produce a technical effect outside of the running of the instructions on the computer.

The EBoA decided that question 2a was inadmissible in part because it wanted to avoid giving an exhaustive list of patentability criteria, and in part because it felt that this was not required in view of its answer to question 2b.  In answering question 2b it confirmed that it is neither sufficient nor necessary, in order to solve a technical problem, that the CI simulation produces a technical effect outside of the running of the instructions on a computer.

The EBoA rejected the two proposed approaches (the CI simulation having “virtual” technical features or tethering the CI simulation to a “specific technical purpose”) suggested by observers.

The answer to question 3 means that the use of a CI simulation as part of a design process makes no difference to determining whether the CI simulation goes beyond a computer running a set of instructions to solve a technical problem.

The EBoA favoured the legal status quo. It said that CI simulations must be examined on a case-by-case basis using the established COMVIK approach to determine whether a technical problem is solved, and indicated that boundaries (i.e. those set by the system or process to be simulated, or as the EBoA puts it “not unlike, for instance, a real forest for a landscape painting”) formed by the models underlying a CI simulation, whether technical or not in themselves, may contribute to technicality if they are “a reason for adapting the computer or its functioning” or if they “form the basis for a further technical use of the outcomes of the simulation (e.g. a use having an impact on physical reality)”. In this latter case, however, the further use must be at least implicitly claimed. For example, a CI simulation that optimises computing resources or provides data for controlling a technical device could indeed go beyond a computer running a set of instructions, and therefore solve a technical problem.

Those hoping for a move away from the case-by-case approach in favour of clear delineation of what is and what is not technical will be somewhat disappointed. The EBoA elected to “refrain from putting forward a definition for “technical”” and further commented “that any definition of a technical invention may have to be extended in the course of time in order to accommodate new technical or scientific developments or to reflect societal changes.” Although a relatively minor part of the decision, this means that the longstanding difficulty in determining whether a feature is or is not technical is here to stay. Practitioners have been at pains to provide an answer to the simple question: what is a technical feature? Obtaining a legally certain answer to this question could have simplified the assessment of patentability of computer-implemented inventions in Europe. Alas, no such certainty has been provided.

It is also worth noting that the EBoA declined to comment on judgments of both the UK and German courts where certain issues overlapped with those of the referral. Thus, it remains to be seen whether a patent granted by the EPO to a CI simulation invention would survive under equivalent national provisions.

 

Implications

CI simulation inventions are not automatically allowable or unallowable. Whilst proponents of patent protection for CI simulations will breathe a sigh of relief that CI simulations can solve a technical problem, they will no doubt bemoan the fact that merely pursuing a patent for the technical principles underlying the simulated system or process is not in itself sufficient to establish patentability.

Once again, context is key and EPO examiners will need to be persuaded that a CI simulation invention involves technical considerations.

One route, albeit only in exceptional circumstances, is to rely on an implied technical use of a CI simulation. Put simply, ask what the CI simulation is used for? Does the CI simulation allow a graphics card to render how light reflects off of complex objects, i.e. ray tracing in a video game? Does the output represent upcoming internet bandwidth spikes that can be managed proactively by an internet service provider? If so, this implied technical use may be enough to demonstrate technicality. However, care must be taken to ensure that a CI simulation invention embraces only the implied technical use. Considering the ray tracing example, if the graphics card was not implied by the claim, the output could also be said to model the physics of light reflections. In such a case, the claim also covers a non-technical use and does not solve a technical problem over its entire scope.  This will lead to a finding of lack of inventive step. Put another way, the output of the CI simulation invention should only have a “specific” and “limited” technical use. In practice, relying on a limited implied technical use will be challenging.

An alternative route to obtaining a patent for a CI simulation is to focus on how the simulation, when run, reduces computational expense. Does the computer run better when the CI simulation is run? Does the processor run at a reduced load? Is reading from and writing to the memory quicker? Is the processor used intelligently with a graphics card? Assuming that the steps of the CI simulation are tailored to modify the computer’s hardware in a non-standard way, the steps are likely to be technical. The steps when executed bring about a better running computer.

At a general level, inventions that concern trade-offs that the skilled person routinely deals with are considered obvious. Therefore, if the CI-simulation is merely an efficiency accuracy trade-off, it will likely be considered obvious. The flip side is that if that trade-off is removed and a CI simulation is more efficient and at least as accurate, if not more accurate than the prior art, such a simulation is likely to be found to go beyond mere computer implementation and solve a technical problem in a non-obvious way. A significant advantage of this route to technicality is that what is actually being simulated does not have to be technical, so long as the implementation of the CI simulation results in a better running computer. A CI simulation that models human behaviour could still result in a better running computer.  Therefore, a CI simulation patent could cover a single piece of software that is used in multiple technical and non-technical disciplines.

In either case, the detailed description of the CI simulation should emphasise the underlying technical considerations. If for example, the focus of the application is on a specific technical use or a better running computer, it will be much easier to succeed when arguing for patentability. Conversely, if the description has a statement that the CI simulation allows one to decide whether to manufacture a design, that is likely to be deemed a business consideration and will make it more difficult, if not impossible when arguing for patentability.

A wider practice point raised in the decision is one of sufficient disclosure under Article 83 EPC where care should be taken to ensure that the skilled person does not encounter undue burden when attempting to carry out the claimed invention. The takeaway here is that a CI simulation invention claimed in terms of functional steps will require the description to set out in detail how the simulation is implemented.

Whilst this decision is confined to CI simulation inventions, the EBoA stressed that there is no legal basis to treat CI simulation inventions differently from other computer-implemented inventions. The inference here is that all computer-implemented inventions will be assessed using COMVIK going forward. Due to the increased use of AI in many fields of technology, there have been some suggestions that the current EPO approach needs to change to accommodate this emerging technology. However, it can be inferred from the EBoA’s commentary in this decision that it feels all CI inventions, AI included, should be assessed using the same approach. Fundamentally, any AI invention will also have to be demonstrated to be more than a computer simply running instructions in order to solve a technical problem and be patentable. In any case, the above practice points should be considered when pursuing an AI invention, even if it does not use a CI simulation.

 

Practice points in summary

  • An implied technical use of a claimed CI simulation confers technicality. Any non‑technical uses should be excluded from the scope of the claims. Claims should be drafted to refer to the inputs, processes, and/or outputs associated with a limited technical use.
  • A better running computer when carrying out a CI simulation confers technicality. Claims should be drafted to refer to all of the CI simulation steps that, when executed, bring about a better computer.
  • Technical considerations should be set out in the description (i.e. implied technical use or getting a computer to run better). The description should be drafted to explain how technical challenges have been overcome. Any non-technical considerations, e.g. business considerations, should be avoided.
  • To avoid possible problems with insufficiency, the implementation of each step of a claimed CI simulation should be set out in detail in the description. Technical advantages of each step should also be indicated.

 

In summary

The EBoA has re-affirmed the developed case law. Applicants seeking patent protection at the EPO for CI simulation inventions as well as more general computer-implemented inventions should consider how their application will succeed under the COMVIK approach. There are proactive steps applicants can take to increase their chances of success.  In particular, emphasising the technical considerations of any such computer-implemented invention will serve applicants well.

View PDF

If you would like to hear more about this decision, our podcast on the topic can be accessed below: