In this blog post, Gabriele Cifrodelli, PhD Candidate and RTA in CREATe, offers a (brief) comment on the latest DABUS Judgement in the UK, by arguing that, not surprisingly, once again a literal interpretation of patent law led to the non-recognition of AI as inventor. However, perhaps a more purposive interpretation should be considered, especially with regard to countries that are still in need of IP as a form of incentive to foster innovation.
On 20 December 2023, right before the holiday break, the Supreme Court of the United Kingdom delivered the judgement in the case Thaler v Comptroller-General of Patents, Designs and Trademarks  UKSC 49. It basically put an end (at least for now) to the issue whether an AI – such as DABUS – can be recognised as an inventor under the current patent system, without any human contribution. Not surprisingly, the Court confirmed that only a natural person (and not an AI) can be deemed inventor, according to Sections 7 and 13 of the UK Patents Act. This same outcome has already occured in other jurisdictions (to name a few: European Patent Convention, United States, Germany, Australia), since Dr Stephen Thaler – the human behind the DABUS machine – sought recognition of this AI inventorship all over the world. Based on the current patent legislation, there does not seem to be room for an inventor that is not human. The (legal) reasoning behind this outcome appears to be consistent with a literal interpretation of the law, and the UKSC itself – in the words of Lord Kitchin – agrees with the consideration made in the Court of Appeal’s judgement, according to which the current legislation was not originally drafted with AI inventors in mind and thus “[i]f patents are to be granted in respect of inventions made by machines, the 1977 Act will have to be amended”. However, after the results of the Consultation on AI and IP (copyright and patents) in 2022 (to which CREATe offered a contribution), for now the inclination seems to be towards not making any substantial changes to the law.
As in most other DABUS decisions, this strict adherence to the legal wording, without offering a more purposive interpretation, fails to consider what recognition of AI inventorship could mean in the current AI age from the perspective of incentives, that have always influenced the enactment and dissemination of intellectual property in different countries and industries. As I argued here and here, it is true that sofware developers often have non-economic incentives to innovate in the AI field (e.g. reputational motives, scientific curiosity, desire to exchange data with other researchers), and they do not look at patents as the main driver for innovation. It is also true that in other parts of the world these additional incentives are not often in place, simply because there are not enough AI developers to, for instance, exchange information with and collaboratively foster innovation. In this context, we should not forget that the South Africa Patent Office (SAPO) is the only one that decided in favour of inventorship for DABUS. Although many commentators argued that this was possible only because SAPO carries out a strict examination of the formalities in the patent application, without any substantial analysis of the patent requirements, probably there is something else at stake here. There is a pressing need for innovation, as evidenced by a purposive interpretation of the South African Patents Act 1978 and by the willingness of the South African government (expressed in a 2019 White Paper) to achieve a sustainable level of innovation through, among other ways, the reform of the IP system by considering the impact of AI. Therefore, recognition of AI-generated inventions can still constitute an important incentive to innovation. Confirming this, the Federal Court of Australia (before the decision was overtuned in appeal) stated – in the words of Beach J – that “no narrow view should be taken as to the concept of “inventor”. […] [T]o do so would inhibit innovation not just in the field of computer science but all other scientific fields which may benefit from the output of an artificial intelligence system”.
To conclude, will all this lead to a situation where AI is attributed inventorship? If not, would this mean, as mentioned by the UKSC in the judgment, that the inventor is still a human but they are supported by AI “as a highly sophisticated tool”? And, thus, would this (new) configuration require some sort of recognition? Only time and research will tell!