Are digital twins as patentable as their physical counterparts?

We live in an increasingly digital world – not only do more of us communicate with each other and access information or content using ‘apps’ and the internet, but some of the work we may have previously conducted in the ‘real’ or physical world is now being performed using computer simulations or digital twins. For example, it is much more time- and cost-effective to simulate the behaviour and performance of a new design of wind turbine blade, than to construct the wind turbine blade and test it in a real wind tunnel. Similarly, it is safer for the public to simulate how an autonomous vehicle performs in a virtual street and use data collected from the simulation to train the artificial intelligence that controls the vehicle, than to let an autonomous vehicle loose in our cities to collect that data.

Many innovative companies in diverse technology areas use computer simulations, from those using software to design new cities, to those using digital twins to model the impact of changes to systems before implementing them in the real world, to those using software to identify and narrow-down potential new drugs for testing. In many cases, the simulations and software may help scientists and engineers to identify what not to do – in such cases, the output of the computer model may never be linked to a real-world product or method.

It is, therefore, unsurprising that companies want to protect their computer simulation inventions, as these are useful and important tools that can help them to design, test, and implement or manufacture. However, whether computer-implemented simulations are patentable at the European Patent Office (EPO) has, to date, depended on whether the claimed invention has any technical features that prevent the claim as a whole from being excluded from patentability. Generally speaking, methods of simulation, design or modelling may comprise some features that fall under the category of mathematical methods or methods for performing mental acts (which are excluded from patentability), but if all of the features of the claimed subject matter relate to these categories, the claimed subject matter is excluded from patentability. Thus, the EPO looks to see if the claimed subject matter contains any functional technical features, or whether the claimed subject matter has a specific and defined technical purpose.

However, last year, the Enlarged Board of Appeal (EBA) of the EPO was asked to consider whether a method of testing, by simulation, a modelled environment was a functional technical feature when there is no link in the claimed method to a physical entity or physical parameters (see here for a summary of the questions referred to the EBA; see here for the case itself). As noted above, this may be the case for many computer simulations or digital twin models, as well as in the field of training autonomous vehicles, finding and testing new drugs, designing and testing computer games, and so on.

On 15 July 2020, oral proceedings were held at the EPO to discuss the questions referred to the EBA (assigned case number G1/19), and for the EBA to hear the views of both the Appellant (patent applicant) and the EPO. More than twenty-three amicus curiae briefs were filed by third parties and stakeholders on this case, as the impact of the referral could be far-reaching. Shortly in advance of the oral proceedings, the EBA issued a communication informing the Appellant of some points which were likely to be addressed at the oral proceedings (see communication from the Registry of the Enlarged Board of Appeal dated 22 June 2020). Due to the ongoing COVID-19 pandemic, the EBA minimised the number of people who attended the oral proceedings in person but enabled interested parties to watch the proceedings as a livestream. Around 1600 people registered to do so, which indicates the interest in this case from the IP community. (It also supports the idea that such proceedings should be broadcast in the future, as this helps to improve and widen access to, and engagement with, the legal process.)

The oral proceedings began with a discussion on whether the questions referred to the EBA were themselves admissible. In other words, does the EBA need to answer the questions, or do existing statutes and case law provide the answers to these questions already? As a reminder, the questions referred to the EBA 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?

2. 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?

3. 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?

The EBA indicated that the referral was admissible in part. Specifically, they suggested that the first part of the second question (relating to providing criteria) was too broad or unacceptably general. The Appellant argued that guidance is needed on the criteria for assessing whether a simulation method is patentable, particularly for cases where the claim for that method does not include an input measurement step (to obtain parameters or data from the real, physical world), or an output control or manufacture step (to control a real world product/machine, or to manufacture/produce a real world product). In other words, is a simulation method that solves a technical problem patentable if there is no link to the real world, either at the input or the output of the method? Furthermore, it was argued by the Appellant that, even if the questions are not admissible in whole or in part, they would like the Board to confirm whether the approach outlined in T1227/05 (Infineon) for determining the patentability of computer-implemented simulations is the correct approach and should be used in the future. As noted in the article mentioned above, Infineon has provided, for over a decade, the test to determine the patentability of computer-implemented simulations at the EPO. In particular, Infineon states that a link to the real-world or technical parameters that are obtained, for example, using the laws of physics, moves a simulation method away from being purely a mental or mathematical method.

Representatives from the EPO stated that the referred questions are on points of law that are of fundamental importance, and certainty is required by the EPO as well as innovators, on how to handle the patentability of computer simulation inventions. Furthermore, they explained that a “direct link to physical reality” (a phrase which was used during another referral to the EBA, in G3/08) may not always be possible because, in some cases, the simulated environment may itself be a virtual environment (think computer games design). In addition, they questioned why the noise in an electric circuit is a “physical reality” (c.f. Infineon), but people’s movement within a building is not (c.f. the case in question).

The EPO went on to make very interesting arguments on whether it is necessary to distinguish between “potential technical effects” and “virtual technical effects”. Potential technical effects may arise when an invention is implemented in the real world, such as the control of an autonomous vehicle or the manufacture of a new product. Virtual technical effects may arise within a model or simulation of the real world, but which may never exist outside of that model or simulation. Thus, it may be more appropriate in some computer simulation cases to evaluate the virtual technical effect, instead of dismissing the case when there is no direct link to the real world or no real-world technical effect. For example, a computer simulation could use direct measurements from real sensors to solve a technical problem. Equally, a computer simulation could use virtual sensors placed in a simulated environment to solve a technical problem. The former would be considered patentable, but the latter may not. Similarly, engineers could build aircraft wings, test them in a wind tunnel and obtain data and parameters from these experiments to evaluate performance or improve the design of the wings. Alternatively, engineers could perform virtual experiments of the aircraft wings in a virtual wind tunnel and obtain data from these experiments for the same purpose. Again, the former would be considered patentable, the latter may not, but since both techniques are solving the same technical problem, should both techniques not be patentable? The EPO argued that, in our increasingly digital world, we are moving away from real-world experiments and data collection and towards virtual experiments, and that as a result, it is necessary to consider whether the technical purpose of a computer simulation to decide whether it is patentable (i.e. whether the claimed subject matter provides a technical solution to a technical problem).

After lunch, the EBA moved on to hearing arguments on the second and third referred questions. Shortly afterwards, the proceedings were closed by the EBA as they had heard from all parties and had no further questions themselves. The EBA will now provide their decision in writing in the next few weeks, and we will provide an update as soon as these are published.

In the meantime, we note that our impression of the oral proceedings is that the EBA will conclude that computer simulated inventions are no different to other computer implemented inventions and should not be treated differently. In other words, it is possible the EBA will conclude the questions are not admissible. Currently, computer-implemented inventions are generally patentable at the EPO if they provide a technical solution to a technical problem, and if they do not wholly relate to excluded subject matter (such as business methods, mathematical methods or mental acts). Computer simulations are a subset of computer-implemented inventions and therefore, should follow the same principle. However, in view of the points made above about some simulations or computer not necessarily having a direct link to a physical entity or a real-world technical effect, it would be useful for the EBA to provide some further guidance on what sorts of computer simulations are patentable. In particular, none of the arguments made during the oral proceedings referred to simulations of systems outside of engineering (such as simulations of the interaction of vaccines with viruses or of chemical reactions), probably due to the backgrounds of the representatives for both the Appellant and the President of the EPO. However, computer simulations must be interpreted broadly so as to include these non-engineering cases. Furthermore, if the EBA decides that virtual technical effects will be used to assess certain computer simulation inventions in the future, how would this virtual technical effect be ascertained, and how much detail of the simulation is required in the patent application? We look forward to reading the EBA’s thoughts.

If your business depends on computer-implemented inventions, we can secure the intellectual property (IP) protection you need, so you can maintain your exclusivity. In addition, we can define a clear path for you through the IP surrounding your technology, so you can continue and grow your business.

Innovation and IP in 2020

First published in Cambridge Catalyst, Appleyard Lees attorney Parminder Lally predicts what’s on the horizon for Cambridge-based innovators in 2020. This January I will have been working in

Read More »