In the first session, “Patenting Computer implemented inventions in Healthcare at the EPO”, Igor Dydenko provided the EPO view on patenting inventions in the Healthcare field which apparently is the top technical field with the most patent applications in the EPO (according to the EPO’s 2018 annual report).
Igor noted that the relevant provisions for healthcare related inventions were those relating to technicality (Art 52(2), (3) and 56 EPC) and medical methods (Art 53(c) EPC). A review of the two examination requirements (note the EPO moving away from the “hurdles” language*) of technicality and inventive step then followed which highlighted the Guidelines for Examination approach for mixed type inventions incorporating a mathematical method (see G-II, 3.3). In particular the discussion highlighted the requirement for inventive step to only be supported by those features of the claimed invention which provide a technical effect serving a technical purpose.
A number of examples highlighting different aspects of inventions in the medical field were presented including:
– Data processing using a mathematical method embodied in an AI system – T598/07 in which the identification of irregular heartbeats from an ECG was found to provide a technical effect
– Data exchange/management – two examples of the use of a smartwatch were discussed, one technical (the comparison of current measurements with historical measurements) and one not technical (allowing the downloading of data only if a premium profile is used by the user)
– Presentation of Information – T1779/14 in which the presentation of menu items on an insulin pump was found to be technical because the user was not required to scroll or switch menus on a limited display device.
With respect to the medical methods requirements under Art 53(c) EPC an example, based on T2438/11, was provided highlighting how a surgical method could be caught by the provision depending on how it was claimed. In the example discussed a method of aligning a surgical drill with a point was presented. In the first formulation of the claim the position of the drill was detected using a tracking system and then an aligning step aligned the drill with the point using a navigation system. The presence of the aligning step means that a surgical step is present and this claim would be caught by Art 53(c). It was noted however that if the claim was reformulated to a method for determining alignment of a surgical drill, comprising the same detecting step but with the aligning step replaced with a step of determining alignment of the distal end of the drill with the point, then this claim would not be caught by the Art 53(c) provisions as there is a functional/spatial separation between the claim and the insertion of the drill. There is no longer a surgical step.
A therapeutic method was also discussed. This example, which was based on T1680/08, related to a computer-implemented method for determining airway pressure of a lung ventilated by an artificial ventilator. In this case, even though there was no explicit step practised on the human body, the step of artificial ventilation was implicit and the claim was caught by the exception of Art 53(c) EPC.
The second talk, from Rachel Free, was on the subject of AI Inventions (“Assessing the Patentability of AI Inventions: Future Strategies”). The first part of the talk asked whether the concept of plausibility which is usually seen in the life sciences field might make its way into the AI/CII field. The starting point for this question appears to have been a recent EPO workshop on computer-implemented inventions where a discussion of plausibility arose in the context of speculative claiming.
Rachel highlighted the Supreme Court case Warner Lambert v Generics (UK) which dealt with plausibility. For those in the CII field Rachel explained that plausibility was introduced to prevent speculative claiming and ensure that the “patent bargain” (gaining a monopoly in return for disclosing the invention) was met. In the Warner Lambert case the invention related to a new use for a known drug and the traditional condition of sufficiency could be met simply by naming the drug in question. However, as the invention lay in the new purpose found for the drug and the patentee had not explained its contribution, the court found that the “purpose” of the sufficiency condition had not been met.
Potential parallels with the CII field were provided (e.g. the use, in a 1993 patent application, of a neural network to predict when an air bag should be activated and a patent application with a broad claim to using a machine learning system to predict when patients might develop a medical symptom).
It was however noted that there would be significant issues that would make it difficult importing a plausibility style test into the EPO approach for claims with a mixture of technical and non-technical features, namely defining a drug in a patent application might be easier than defining a machine learning algorithm and including test data in a CII patent might potentially limit the scope of the invention.
The second half of Rachel’s talk discussed the concept of AI inventors. The current EPO position is that the designation of inventor form needs to relate to an inventor rather than an AI system. The impact of AI systems, e.g. in the area of drug discovery, may however provide an argument that AI inventors need to be recognised. The EPO’s argument is that none of the current EPC states allow AI systems as inventors and identifying such systems as inventors would not be warranted, e.g. because such systems cannot own property. A draft CIPA position paper on the subject argues that only natural persons should be acknowledged as inventors and the inventions created by AI systems should place ownership and accountability with humans or human owned entities.
This is an area that’s had some public visibility recently with the filing of two patent applications naming the AI system Dabus as inventor (see BBC news story here). However, in this author’s view AI inventions can still be treated as discoveries by the humans who have set up the AI system which should be treated as an advanced tool. An alternative arrangement could be to recognise companies (who own AI assets) to be named as inventors.
Hugh Dunlop presented a session on “Claiming priority under the “same invention” test” and particularly looked at the field of standards setting where patent applications are often filed shortly before a submissions to a standards committee is made. Since the content of the submission is often posted to a standards server on the same day it is important that a later application can claim priority from the application filed shortly before the submission. Some examples with respect to three Samsung v Apple cases in 2013 ( EWHC 467 (Pat) and  EWHC 468 (Pat)) and a more recent case between Conversant and Huawei were provided.
In conclusion Hugh noted that where an interim disclosure has been made it is important that features claimed in a later publication do not depart from the wording of the priority application. Generalising a claim can lead to issues as the part of a claim that is not entitled to priority will need to be inventive over the disclosure.
Scott Roberts provided an entertaining look at the Enlarged Board of Appeal (EBOA) Referral in “G1/19 – Patentability of computer implemented simulations”. Scott pointed out that he had been responsible for producing the IP Federation amicus curiae brief that has been submitted to the EPO (see here). In Scott’s view the questions that have been sent to the EBOA are badly framed and can already be answered with respect to established caselaw. The referred questions are:
- 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?
- 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?
- 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?
In Scott’s view these can be answered as follows:
Question 1 – “yes” in accordance with T1173/97;
Question 3 – the answers to the first two questions are the same.
One interesting part of Scott’s talk was how he placed the invention in T489/14 into context within a product development process. The invention in T489/14 involves the simulation of pedestrians as part of the design process for a building structure. The invention in question resides in the early design work that occurs upstream in the development process that eventually results in a building. The invention under appeal is broadly similar to that decided in T1227/05 (Infineon), which is referenced in the Appeal.
The final session was presented by Phil Thorpe of the UKIPO (“CII decisions from the IPO”). A number of recent decisions relating to excluded subject matter were highlighted including BL O/148/19 (Fisher Rosemount), BL O/748/18 (Suunto), BL O/128/19 (Arris Enterprises – a decision on the papers) and BL O/0375/19 (Motorola) which were allowed and BL O/610/19 (Mastercard), BL O/800/18 (Motorola) and BL O/434/19 (Ocado) which were refused. The UKIPO’s practice of issuing communications that state that a search would serve no purpose was also briefly discussed. Apparently, the IPO issues such section 17(5)(b) search reports in only 2% of cases so if you’ve received one of these then you’re unlucky! Over 805 of such reports that are issued are in the business method and computer program fields (G06Q and G06F respectively)
Mark Richardson 14 November 2019
*(Note: the EPO is no longer referring to these as the two “patentability hurdles” as such language apparently implies a barrier)