04/11/2024
What is Generative AI?
Interest in generative artificial intelligence, or generative AI, has skyrocketed in recent years. Much of this can be attributed to the launch of ChatGPT – an AI chatbot powered by large language models – towards the end of 2022. Generative AI is disruptive in its ability to produce new content in a range of formats, including text and image, distinguishing it from ‘traditional’ AI whose functionality lies in the analysis of, as opposed to the creation of, data. Naturally, the attention that generative AI has commanded in recent years has led to consideration of to what extent it can help address challenges in a variety of sectors, one being the renewable energy sector.
Generative AI in Renewable Energy
One of the significant challenges faced in renewables is the unpredictability of sources such as wind and solar. Such unpredictability is a problem when it is critical that grid supply and demand correspond, forming a barrier to renewable energy grid integration. Much of the discussion surrounding generative AI in renewables looks to overcome this by exploiting its affinity for highly accurate forecasting that takes many relevant factors into account. In terms of actual implementation of generative AI in renewables, the National Renewable Energy Laboratory, based in the US, launched a model Sup3rCC which is able to output high-resolution future climate simulations which can be used to forecast renewable energy generation. At the intersection between generative AI and simulation, however, one should be mindful of the European Patent Office’s (EPO) approach to subject-matter that is excluded from patentability.
Patentability in Europe
The European Patent Convention (EPC) excludes mathematical methods from patentability. Fortunately, this only excludes inventions that are ‘purely abstract mathematical method[s]’. Although inventions based on simulation and modelling will, more often than not, include features that, alone, would fall under such an exclusion, the computer-implemented nature of such inventions adds a technical character that prevents them from falling foul to the exclusion as long as the claim-language reflects this. Whether such claims are ‘inventive’ , however, is determined based on features that ‘contribute to the technical character of the invention’ (the so-called ‘COMVIK’ approach).
Enlarged Board of Appeal decision G1/19 is relevant here, which centred on a patent application concerning simulating the movement of a crowd. This decision confirmed that the COMVIK approach is applicable to inventions related to computer-implemented simulations and further determined that:
- ‘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’; and
- ‘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’.
Consideration should also be had for the EPO’s guidance on artificial intelligence inventions. In particular, the Guidelines for Examination outline that although there is no general requirement to ‘disclose the specific training dataset itself’, ‘if the technical effect is dependent on particular characteristics of the training dataset used, those characteristics that are required to reproduce the technical effect must be disclosed unless the skilled person can determine them without undue burden using common general knowledge’.
Beyond the EPO, the patentability of AI-based inventions remains topical. Following a recent UK Court of Appeal ruling, the UK Intellectual Property Office issued updated guidance on examining patent applications involving artificial neural networks (ANNs). Notably, the guidance sets out that ‘patent examiners should treat ANN-implemented inventions like any other computer implemented invention’ within the context of subject-matter excluded from patentability.
While it may be too early to predict what the patent landscape will look like for generative AI in renewable energy over the coming years, a 2024 WIPO report looks at historical filing trends in different sectors. Promisingly, the number of patent families related to generative AI in energy management has experienced the most growth since 2018, beating applications such as transportation and agriculture.
How We Can Help
If you are an innovator working at the intersection between generative AI and renewable energy, our team of attorneys at Reddie & Grose have technical knowledge and expertise that can help you navigate the particulars of seeking patent protection for your inventions.
Read the previous article of our Digitalisation Mini-Series here: ‘Safeguarding the Future: The Critical Role of Cybersecurity in the Energy Transition’
This article is for general information only. Its content is not a statement of the law on any subject and does not constitute advice. Please contact Reddie & Grose LLP for advice before taking any action in reliance on it.
