UK Supreme Court dismantles the emotional rollercoaster

Today (11 February 2026) the UK Supreme Court delivered their long-awaited judgment in the Emotional Perception AI case. Applicants, innovators and patent attorneys working in the AI space can finally breathe a sigh of relief as it is now much clearer that AI innovations can be patented in the UK

We have already shared a summary of the case and the proceedings to-date, which you can read here. In essence, the case, which has been progressing through all the UK courts over the last few years, concerns whether artificial neural networks, ANNs, are considered “programs for computers…as such” and therefore are excluded from patentability in the UK.

The case is of importance to many AI innovators, but particularly those who are developing new AI and machine learning techniques and architectures. Many of these innovations are general and could be advantageously applied to many different application areas, but the UK IPO has historically frowned upon such innovations as being mathematical methods and/or programs for computers, despite the fact that these innovations have technical benefits such as smaller model size, easier and more efficient or privacy-aware training of models, on-device personalisation, and so on.

Today, the hearing began with a statement that UK law is intended to be in line with the European Patent Convention, and that determining whether an invention is a “program for a computer” has been difficult to determine in the UK and Europe for many years. The judge delivering the decision referred to how G1/19 in Europe had caused a change in the approach taken by the EPO (see our article on this case here: Patenting computer-implemented simulations - Appleyard Lees), and whether the UK’s Aerotel approach should continue to be followed or whether it should be replaced with the approach set out by the EPO’s Enlarged Board of Appeal in G1/19.

The Supreme Court has decided that Aerotel should be abandoned and replaced with G1/19. This is because the G1/19 approach cannot be described as “obviously wrong” and therefore should be followed by the UK.  They noted that adopting the G1/19 approach in the UK will not require any changes to the UK approach to novelty or inventive step.

The judge went on to say that the ANN of the claimed invention is not a “program for a computer as such” and that the UK IPO hearing officer was incorrect to reject the application on the basis of excluded subject matter.

The case will be referred back to the UK IPO hearing officer to apply the law to the facts of the case.

We will report further on the details of the judgment once we have carefully reviewed the written judgment and once the UK IPO releases their own practice note. In the meantime, if you have any immediate questions about the impact of this case on your UK patent portfolio, please get in touch with the Appleyard Lees AI team.