The End of an Era-otel
The Supreme Court has handed down their judgment on Emotional Perception AI Limited v Comptroller General of Patents, Designs and Trade Marks (‘EPAI’).
This marks the end of a long series of hearings that has made its way through the UK Courts. A quick summary of how we got here can be found in our previous article following this case.
Issue at hand – how is excluded subject matter assessed?
The EPAI case has largely revolved around deciding to what extent statutory restrictions on patenting a program for a computer as such applies to artificial neural networks (ANNs).
Section 1(2) of the UK Patents Act 1977 (UKPA) restricts the patentability of inventions in certain areas – this includes ‘programs for a computer’ – in so far as the invention relates to that area ‘as such’. The approach the UK Intellectual Property Office (UKIPO) took to assessing ‘as such’ was set out in Aerotel Ltd v Telco Holdings Ltd [2006] EWCA Civ 1371; [2007] RPC 7 (‘Aerotel’). Aerotel sets out a four-step test to decide whether a claimed invention goes beyond the excluded areas. Once it is decided that a claimed invention is beyond the statutory restrictions as such, then the novelty and inventive step of the claims can be analysed, in particular applying Windsurfing/Pozzoli to assess inventive step.
Similarly, the European Patent Office (EPO) excludes the protection of computer programs as such in Article 52(2) EPC. However, the assessment of exclusion is largely wrapped up in the assessment of inventive step. This approach was set out in the Board of Appeal’s decision in T 641/00 (‘Comvik’) which was confirmed by the later Enlarged Board of Appeal decision in G 1/19. While Comvik largely deals with inventive step, it first considers whether the claimed invention embodies or involves the use of a piece of physical hardware, however mundane (the so-called ‘any hardware’ step). Comvik then considers which features (be they technical or non-technical) contribute to the technical character of the invention as a whole and then asks if those features are inventive.
The Aerotel approach is at odds with the EPO’s approach. In the EPO decision T 154/04 (‘Duns’), the Board was of the view that ‘[Aerotel] is not consistent with a good-faith interpretation of the European Patent Convention’. This schism in turn contradicts section 130(7) UKPA which aims to bring UK patent law into conformity with the provisions of the European Patent Convention.
This broadens the issue at hand and asks whether the Aerotel guidance should no longer be followed.
Arguments for Aerotel
The Comptroller General argued threefold:
Firstly, that there is no evidence that Aerotel leads to a different outcome when compared to following the G 1/19 approach. The Court dismissed this argument saying that it is ‘impossible to be sure whether a departure from the Aerotel approach will or will not affect outcomes’. On the one hand if it does lead to different outcomes then justice to those affected requires necessary changes to be made. On the other hand, no harm is done ‘(beyond the need for some re-training)’ in updating the UK’s approach.
Secondly, the Comptroller General argued that since G 1/19 relates to inventive step, the adoption thereof would require the upheaval of well-established UK case law relating to inventive step ‘creating a Frankenstein monster’. The Court also dismissed this in a handwavy fashion saying they will address how they consider that consistency may be achieved.
Thirdly, the Comptroller General argued that the Comvik approach, and in particular the ‘any hardware’ test was really no more than ‘window-dressing’ and did not present any a barrier to patentability. This was also rejected by the Court.
The decision
In the Court’s own words in paragraph 48 of the decision: ‘Nonetheless the applicable principles are clear, and the fact that a change may be disruptive cannot by itself justify maintaining an idiosyncratic approach which is inconsistent with established jurisprudence of the EPO’.
As has been heavily foreshadowed by now, the Court decided that the Aerotel test should be scrapped and replaced with the test set out in G 1/19.
It was decided that the claim at hand clearly met the ‘any hardware’ step, but since none of the parties presented any arguments for or against the application applying G 1/19, the matter has been remitted to the UKIPO to apply the new test to the case.
In a comment that restores a certain amount of sanity to the history of this case, the Court confirmed the decision from the Court of Appeal that an ANN is a program for a computer.
What does this mean in practice?
The Court has provided some guidance on what the new approach should look like and in particular how it should interact with other aspects of prosecution (namely inventive step).
The Court provided a first stage – stage (1) – that corresponds to the ‘any hardware’ test, i.e., whether the claimed invention embodies or involves the use of a piece of physical hardware, however mundane. As submitted by the Comptroller General, any competent drafter can satisfy this stage by including some piece of hardware in the claim.
The second stage – stage (2) – then relates to determining the novelty and inventiveness of the claim. This, as decided by the Court, remains unchanged but ‘may require some modification’.
In between stages (1) and (2) lies the so-called intermediate step. This is where Comvik largely comes into play. The intermediate step recognises that a claim may legitimately consist of a mix of technical and non-technical features and asks which features contribute to the technical character of the invention when viewed as a whole. The Court viewed this as a stage which is concerned with filtering out features which do not contribute to the technical character of the invention before assessing inventive step.
Unhelpfully, the Court did not go into further detail in how the intermediate step interacts with the assessment of inventive step and have remitted the question of implementation details to the UKIPO. One speculative way of doing this would be to take the inventive concept of the claim as construed in step 2 of Windsurfing/Pozzoli as being the technical character of the claim as identified in the intermediate step.
A potential knock-on effect of this which will come as a relief to applicants is that there might be a reduction in the practice of the UKIPO not issuing searches on borderline excluded subject matter applications. This would mean that it will be possible for applicants to first file in the UK and be more likely to get an initial search for a relatively lower filing and search fee when compared to the EPO. This would make the UK a more patentee friendly jurisdiction especially for computer-implemented and AI-related inventions, though there still remain lots of unanswered questions relating to how the intermediate step will be applied in practice. However, how the new test will work in practice, and whether it will be more or less ‘patent friendly’ is yet to be seen.
The next step we are expecting is for the UKIPO to issue a practice note in the near future. This should set out how they plan to implement the G 1/19 test to assess excluded subject matter and how they plan for it to integrate with novelty and inventive step assessment. Until then it remains unclear how the UKIPO will assess excluded subject matter. We are unlikely to have any decision come out of the UKIPO until this guidance has been issued.
At Barker Brettell we keep up to date with the constantly evolving landscape that is patent law and are always looking at the best way to help our clients best protect their intellectual property.
If you would like to discuss this matter further, please do not hesitate to contact the authors, Francesco Di Lallo or David Combes, or your usual Barker Brettell patent attorney.
