The role of patents in promoting AI investment

Patents and technological development

 

Technological development is vital to achieving economic growth, as the recent history of the US economy clearly demonstrates. The patent system provides an incentive to invest in technological development, by giving innovators the right to prevent third parties from working their invention for a limited time (normally 20 years from the date of filing a patent application) in exchange for publishing details of how to work the invention.

 

Economists are divided over exactly how much of an incentive the patent system provides, compared to other tools. However, a 2007 study by three economists concluded that the patent premium for innovations that were patented was substantial, stating: “Firms earn on average a 50% premium over the no patenting case, ranging from 60% in the health related industries to about 40% in electronics.”

 

More recently, in January this year, the European Patent Office and EU Intellectual Property Office published a report which found that European companies that own patents have on average 28.7% higher revenue per employee and 43.3% higher wages than those that do not. The report was based on analysis of 119,000 European companies over a 10-year period.

 

In this context, it is notable that UK companies file fewer patents than their counterparts in other developed countries. The UK ranked seventh in terms of the number of PCT applications filed in 2023, behind China, the United States, Japan, Korea, Germany, India and France. German businesses filed over 133,000 PCT applications, while those in the UK filed less than 49,000.

 

Patents and AI

 

Commercialisation of AI is still at a relatively early stage, and it is perhaps too early to say to what extent the correlation between patents and economic growth will be evident in this sector. But there is no doubt that the patent system has a role to play (alongside other incentives) in de-risking both the R&D investment prior to an inventive concept being conceived and efforts to commercialise that inventive concept.

 

In particular, patent protection will be invaluable in incentivising investment in labour, equipment, etc., to develop and commercialise AI systems. In this regard, the economic arguments for patents are therefore as relevant for AI innovation as for any other area.

 

While patenting fundamental AI models is likely to remain difficult, because of the exclusion from patentability of certain categories of invention – including computer programs and mathematical methods, as such – g patenting novel and inventive applications of AI should be possible. Given that AI is likely to be transformative for many industries, from medicine, energy and transport to entertainment, finance and even law, the opportunity for patents to incentivise innovation is huge.

 

However, once again we see that the UK lags behind other countries in the patent race. According to a report in R&D World, Chinese companies were granted nearly 13,000 AI patents in 2024, and US companies more than 8,600. By contrast, there were just 369 AI patents granted to UK companies. 

 

A WIPO report on Generative AI (GenAI) patents published last year produced similar findings: during the period 2014-2023, China accounted for over 38,000 published patents, while the UK had just 714.

 

However, the WIPO report also found that the UK ranked third (behind China and the US) in terms of the number of GenAI scientific publications and the number of citations of GenAI scientific publications. This suggests that R&D is taking place, and the results are being published and noted – but it is not (yet) evident in patent applications.

 

If the UK is to generate success stories that compete with companies such as OpenAI in the US and DeepSeek in China (which recently became the number one app on the App Store), AI companies need to actively consider the patent system alongside other tools such as trade secret protection, brand building and first-mover advantage.

 

One positive example from the UK is Google DeepMind, which is based in London but since 2014 has been part of Alphabet. According to analysis published last year by GreyB, it had 972 patents globally, spanning 314 unique patent families. Of these, 833 patents were active and 291 had been granted.

 

Harnessing the power of patents

 

AI raises many interesting issues related to IP. For example, we have reported on the extensive litigation in the DABUS cases about whether an AI can be an inventor, and the UK Supreme Court recently agreed to hear a case concerning whether an application for a patent for an artificial neural network for file recommendation is excluded from patentability on the ground that it is a computer program as such (Emotional Perception AI Limited v Comptroller General of Patents). There is also litigation pending in the Getty Images case about alleged trade mark and copyright infringement by Stability Diffusion, an image generation tool.

 

We will continue to monitor these developments, but it is important not to lose sight of the bigger picture: there is historically a strong correlation between patent activity and economic growth. The patent system is therefore one of the most important tools a country has to incentivise technological innovation and the commercialisation of new technologies such as AI.

 

If the UK is serious about leading the AI technological revolution to promote economic growth, then we need to harness the power of the patent system.