Mathys & Squire Partner Edd Cavanna has been featured in Practical Law with an article on the restrictive nature of the UK’s patent assessment framework in the field of AI and quantum computing.
Read the article below.
This article first appeared in the June 2025 issue of PLC Magazine.
In recent years, the IT sector has grown rapidly, with two strong drivers being AI and quantum computing (see box “What is quantum computing?”). In June 2017, PwC’s Global Artificial Intelligence Study estimated that AI could boost the world economy by $15 trillion by 2030 (www.pwc.com/gx/en/issues/analytics/assets/pwc-ai-analysis-sizing-the-prize-report.pdf). In April 2024, McKinsey & Company’s Quantum Technology Monitor reported that the quantum computing industry alone could add around $2 trillion to the world economy within a decade (www.mckinsey.com/capabilities/mckinsey-digital/our-insights/steady-progress-in-approaching-the-quantum-advantage).
While it is easy to view this growth solely through a financial lens, the implications reach much further. It is no exaggeration to say that the IT sector is on the cusp of a paradigm shift. The UK needs a robust strategy to capitalise on its strong academic sector. Driving innovation in the IT sector is important both for national security and to position the UK as a world leader. Part of this strategy will involve ensuring that innovators are able to corner the market in their respective fields.
While there are many social and political factors that may contribute to this strategy, this article focuses on the interplay between intellectual property law and the IT sector. In particular, the ability to patent technological inventions is a core factor in the attractiveness of both investment in, and the operation of, innovative technology solutions.
The current UK framework for patenting AI and quantum computing inventions risks falling out of date and making the UK less attractive for innovators. The UK must take steps to update the patent assessment framework in order to realise the benefits of these exciting technological advances.
A historical view
The present situation can be traced back to 1977, when the UK was heavily involved in negotiating the European Patent Convention (EPC), which is the legal text that provides for patents to be granted centrally by the European Patent Office (EPO) and validated across Europe. The EPC required each signatory to harmonise its law so that there could be no disagreement regarding the national validity of the centrally granted patents.
Patent law in the UK was last substantially updated by the Patents Act 1977 (1977 Act), a large part of which was intended to harmonise UK patent law with the EPC. It was also at this time that the question arose of whether, and to what extent, software should be patentable. Essentially, the 1977 Act sought to delimit generic software, which was intended to be excluded from patentability, from software that is directed to some technical purpose, which could be protected.
At that time, it was not possible to predict that the IT boom of the 80s and 90s was just around the corner, let alone foresee the more recent progress made in AI and quantum computing. Software was reasonably perfunctory and seemed to be adequately protected by copyright law. Nevertheless, the drafters of the EPC did consider that there might be future developments and included a proviso that software would only be excluded from patentability to the extent that the patent application related to software “as such” (see feature article “Artificial intelligence: navigating the IP challenges”, www.practicallaw.com/w-015-2044).
Over the following years, both the UK Intellectual Property Office (IPO) and the EPO developed case law to try to delineate the boundaries of the “as such” clause. While each body went through periods of greater or lesser enthusiasm for software patents, and each developed its own framework for assessing compliance, the two systems remained in remarkably close agreement on what should be allowed at any given time. This was crucial in order for patent holders and their competitors to have reasonable certainty in the validity of patents in the UK, irrespective of whether they were granted by the IPO or the EPO.
The current situation
From the perspective of patent attorneys using the system, it has felt like the IPO and the EPO have been diverging on their approach for the past few years. This divergence came into sharp focus with the explosion of innovation in the fields of AI and quantum computing.
When a new technology causes rapid developments, it is understandable that patent offices become more conservative, as has happened in this case. In short, they are concerned that patents will become too easy to obtain if patent offices become dazzled by shiny new technology which, as it achieves what was previously impossible, can seem more technical at the time of invention than it may later seem with hindsight.
As each new technology gains importance, it is vital that patent offices play their core role in encouraging innovation. Both the IPO and the EPO have, in essence, attempted to fit these new technologies into their existing frameworks for assessing software. They are necessarily bound not just by their (nominally harmonised) legislation, but also by the nearly 50 years of case law interpreting the legislation. In each case, the view appears to be that inventions in AI or quantum computing can be protected. However, the European framework has seemingly led to a more receptive system where it appears that, for the time being at least, there is better engagement with new technologies from the patent examiners.
This has meant that applicants are more attracted to the European patent system than the UK system. In addition, the European system allows for protection in up to 45 jurisdictions. Furthermore, as highly technical fields such as AI and quantum computing tend to rely on international talent, the conclusion is clear: in order to become a world leader in these technologies, the UK will need to persuade talent to come, and to stay. The UK must provide its businesses with legal certainty and protection.
During 2025, the Supreme Court is due to hear arguments about the patentability of software in the long-running proceedings in Comptroller-General of Patents, Designs and Trade Marks v Emotional Perception AI Ltd ([2024] EWCA Civ 825; see News brief “Artificial neural networks: Court of Appeal examines patentability”, www.practicallaw.com/w-044-2469). In this case, the IPO refused a patent application on the grounds that the relevant technology, which recommended music tracks based on an “emotional similarity” with other music tracks, was mere software.
Emotional Perception is interesting because the arguments presented cast doubt on the meaning of “a program for a computer” in section 1(2)(c) of the 1977 Act. In cases where the invention is, or is implemented in, a neural network, a part of the invention is encapsulated in the weights learned by the nodes of the network. It is not wholly apparent that this is best viewed as either software or hardware. It is encouraging that the IPO has engaged with the judicial process to help bring clarity to the situation.
While no corresponding case is working through the courts for quantum computing, in January 2025, the IPO updated its guidance to simply state that quantum computers should be treated in the same way as classical computers (www.gov.uk/government/publications/examining-patent-applications-relating-to-artificial-intelligence-ai-inventions).
While this may be true in a very broad sense, in that the quantum computing system is performing technical processes to qualify for patentability, it largely discounts the core features that make quantum computing so powerful, such as parallelism and superposition states. In practice, this could result in significant restrictions for patenting quantum computing software.
The difficulty of patenting quantum computing software is demonstrated by some recent Hearing Officer decisions at the IPO in which the special quantum processes were given short shrift (River Lane Research Ltd, BL O/130/22; Odyssey Therapeutics UK Limited, BL O/1193/23; 1QB Information Technologies Inc, BL O/935/22). This is not to say that the IPO is refusing to accept this new technology, but that the patent examiners are just as constrained by the assessment framework as the applicants.
It is not just the EU that the UK is competing with in this field. Major markets, such as the US and China, will also be forming their own pathway. For now, the US appears to be caught in a particular swing of the pendulum to being more restrictive about software in general, following the US Supreme Court decision in Alice Corp v CLS Bank International (573 US 208 (2014)). By contrast, China appears to be a welcoming market for these advanced technologies, at least in terms of patentability.
Where next?
The antidote to this competitive threat lies in ensuring that the IPO’s patent application guidelines are regularly updated in a pragmatic and fair manner. Training patent examiners to assess these new technologies, both from a technical and a legal point of view, is also critical. In addition, as Emotional Perception reaches its conclusion, some clarity should arrive regarding the patentability of software in general in the UK.
However, the UK should go further than this. It is currently operating within a framework in which software is assessed by rules that were created to deal with highly procedural classical software. The changes in the past 50 years in this field have been phenomenal and it is no surprise that the legal framework has struggled to keep up.
This naturally leads to the view that the UK should consider changing how it treats inventions in the fields of AI and quantum computing to make it easier to patent this technology. Such inventions are not mere software in line with the simplistic view that anything running on a computer is necessarily software; they show remarkable ingenuity by performing in genuinely new and technical ways. The existing UK framework is general enough to be interpreted to achieve this change without overturning any case law. The remaining question is whether it is brave enough to go first.
What is quantum computing?
Quantum computing is a technology that is based on the principles of quantum theory. It seeks to harness the subatomic physics of quantum mechanics where particles can occupy more than one value or state. In data processing, binary digits (bits) have one value (0 or 1) or occupy one of two states (on or off). A quantum computing bit (qubit) can simultaneously have more than one value or be in more than one state. Quantum computing increases the number of computations that can be processed concurrently, therefore enhancing the speed of computer processing.
