Legal test for assessing AI patentability clarified in the UK

Mark Marfé and Carissa Kendall-Windless of Pinsent Masons were commenting after the UK Supreme Court decided (41-page / 383KB PDF) that the UK Intellectual Property Office (UKIPO) should reconsider a patent application filed by Emotional Perception AI Ltd (EPAI) for a system that it claims provides people with music and other media recommendations based on how they have been categorised by trained artificial neural networks (ANNs). The patent application was previously rejected by a UKIPO hearing officer.

Under the UK Patents Act 1977, inventions are patentable if they are new, involve an inventive step and are capable of industrial application. However, some things are potentially excluded from patentability under the legislation – including computer programs.

Whether something constitutes a computer program is a question that patent offices and the courts around Europe have grappled with for years. In England and Wales, the legal test for assessing whether something falls to be excluded from patentability as a computer program was set by the Court of Appeal in 2006 in a case involving Israeli company Aerotel. The four-step test it endorsed has been widely adopted by the UKIPO and courts in the UK since.

However, the Supreme Court has now said that the Aerotel test “at best jumbles up the test of an invention with the other requirements for patentability”. From now, the UKIPO and the UK courts will be expected to follow an alternative approach developed by the Board of Appeal at the European Patent Office (EPO). The Supreme Court said that approach better reflects the provisions in the European Patent Convention, from which the UK Patents Act 1977 is derived.

Having updated the correct legal test to be applied, the Supreme Court has reverted the case back to the UKIPO for it to assess whether EPAI’s patent application is novel and inventive.

Mark Marfé of Pinsent Masons said: “The UK Supreme's decision is another example of the UK courts leading the way in clarifying accountability for AI-based inventions.”

“It is encouraging to see the UK at the forefront of shaping how AI innovation is treated within the patent system,” Marfé said. “Much like the DABUS/machine inventor litigation, this case shows that the most important questions in AI and inventorship are being considered at the highest levels of the UK judiciary. That’s a positive sign for innovators looking for legal certainty.”

“Notably, the Supreme Court has sought to align its assessment of patentability for computer programs with the approach of the EPO. Harmonisation is welcome and assists greatly, as most companies in this space have international patent filing strategies,” he said.

Two ANNs operate in EPAI’s system. The first is a system that receives media files, together with natural language descriptions of how a human perceives those files – such as happy, sad, relaxing. That ANN analyses how the files have been described and, based on human instructions, makes assessments of how similar or different the respective media files are.

The second ANN analyses the media files, via parameters set by a human, to determine their characteristics – such as, in the case of music, its tone, speed, and loudness – and again makes assessments of how similar or different the respective media files are. The second ANN is also trained to cross-analyse the similarities and differences established by both ANNs and adjust its own internal assessment parameters – characterised by the courts as weights and balances – to make recommendations that more closely align with what it perceives a user likes.

A UKIPO hearing officer previously considered that the EPAI system as a whole was a computer program and that its provision of “semantically similar file recommendations” was “not technical in nature”, meaning he considered it was excluded from patentability under UK patent law. It is established case law in England and Wales that some computer programs can be lifted out of the exclusion from patentability if they can be shown to make a technical contribution to the ‘prior art’, that being the relevant technology or processes known at the time.

EPAI successfully appealed against the hearing officer’s decision before the High Court in 2023. The judge in that case, Sir Anthony Mann, described an ANN as an aspect of AI system and as being “a black box which is capable of being trained as how to process an input, learning by that training process, holding that learning within itself and then processing that input in a way derived from that training and learning”.

In his ruling, Sir Anthony Mann said a distinction between the two ANNs should be drawn, for the purposes of assessing the patentability of EPAI’s system. His view was that the first ANN – the hardware ANN – was the only one of the two that constituted a computer program. Even then, he considered it made a technical contribution that lifted it out of the exclusion from patentability.

In 2024, the Court of Appeal overturned the High Court’s decision. The Court of Appeal considered that both ANNs in question constituted computer programs. It also considered that the EPAI system did not make a technical contribution to lift it out of the exclusion from patentability applicable to computer programs.

The Supreme Court has now endorsed the Court of Appeal’s view that an ANN together with its weights and biases is a program for a computer.

“It is irrelevant whether an ANN is implemented on a machine in which the weights and biases are baked into the hardware and cannot be changed; or on a machine in which the network architecture is ‘hard-wired’ but the weights and biases can be altered; or on hardware which can be reconfigured; or on a conventional computer for which implementing an ANN is only one of many possible uses,” the Supreme Court said.

“In each case the ANN – irrespective of whether any of its features (and, if so which) are adjustable – represents a set of instructions to the machine to perform operations. The ANN as a whole is a computer program. That is so even in the extreme case of a trained ANN in which the weights and biases have been frozen and incorporated in hardware that can be reproduced by manufacturing identical copies of the same physical circuit of components, just as it is in the case of an ANN in which the weights and biases can be altered,” it said.

It is now for the UKIPO to fully assess whether EPAI’s system should nevertheless benefit from a UK patent.

Marfé said: “It is important to underline that even if an ANN and its weights are characterised as a computer program, that does not automatically make it unpatentable. It simply means the analysis then moves on to whether the statutory exclusion applies.”

Carissa Kendall-Windless of Pinsent Masons added: “Having adopted the new test, the Supreme Court considered it appropriate that the UKIPO – as the forum with day to day technical and examination expertise – carry out the detailed assessment of whether the EPAI system is patentable. That sends a positive signal about the IPO’s capability as an expert, reliable venue for resolving sophisticated patentability questions in emerging technologies.”

“The decision also provides meaningful guidance for AI innovators and their advisers. Going forward, claim drafting in patent applications may need to more closely tie machine learning advances to computational level technical effects – throughput, cache usage, numerical stability – or to physical world control,” she said.

“Viewed alongside the earlier DABUS litigation, this case confirms that the UK is testing the limits of AI patent law across multiple doctrinal fronts – from inventorship and entitlement to subject matter eligibility. The UK remains at the forefront of judicial engagement with the big questions that matter to AI developers. Even though the detailed assessment now returns to the IPO, the Supreme Court’s guidance brings valuable clarity, helping innovators understand how to protect their technology in the UK and strengthening the jurisdiction’s attractiveness for AI focused investment and international patent filing strategies,” she said.