For the last two decades, companies seeking UK patent protection have had to contend with a patent system that often felt like it was designed to frustrate rather than foster innovation, in the software field at least.
Yesterday, February 11, 2026, that era officially ended. In its landmark decision Emotional Perception AI v Comptroller General of Patents, the UK Supreme Court didn’t just tweak patent law; it essentially admitted that the country had been on the wrong track for twenty years when it came to protecting computer-implemented inventions.
Better still, the Court went a long way to harmonising the UK approach with the EPO’s, giving much greater legal certainty to all users of the patent system across Europe, while still leaving some questions open for further development by the UKIPO and the lower courts.
For years, the UK's approach to patenting software was governed by the much-criticised Aerotel/Macrossan test established in 2006. This test set out a series of questions that appeared designed to categorise many innovative inventions as merely a “computer program as such” – or in other words “not a real invention and therefore unpatentable.”
The result of this approach? Applying for a patent via the UKIPO or via the EPO could lead to radically different outcomes, discouraging software inventors from applying for national UK protection, and even if a patent was granted at the EPO covering the UK, the validity of the patent before the UK courts was much less certain due to the courts applying a different and more stringent test than had been applied in the granting process.
While the UK continued to apply Macrossan, the European Patent Office (EPO) had spent years refining a much simpler “two-hurdle” approach.
In Decision G 1/19, the EPO’s Enlarged Board of Appeal noted that the first hurdle to be cleared by a software invention is often trivially low: if your claim mentions any piece of physical hardware—even a standard, off-the-shelf PC—it is no longer “just a program.” It has technical character and is officially an “invention.” The real fight then moves to the second hurdle: whether that invention is actually novel and inventive.
The Enlarged Board noted that the “two-hurdle” approach actually involved an intermediate step between the first hurdle (any hardware?) and second hurdle (novel and inventive?). The intermediate step filters out the claim features that do not make a contribution to the technical character of the invention. However it is clear that non-technical features that do make a contribution must be considered as part of the inventive step analysis.
The Supreme Court noted that the EPO’s Enlarged Board, while not having commented directly on the Aerotel/Macrossan test, explicitly followed certain decisions that loudly disapproved the Aerotel/Macrossan test as a “bad faith” implementation of the European Patent Convention. The criticism, whose logic resonated with the Supreme Court, was that the Aerotel/Macrossan test tried to decide if something was an “invention” based on how clever it was, rather than just what it was.
In the Emotional Perception ruling, the Supreme Court, to its credit, recognised that the EPO’s two-hurdle (plus intermediate filtering step) approach was more consistent with the legislation, and delivered a brutally honest critique of the previous legal framework. Led by Lord Briggs, the Court explicitly stated that the Aerotel/Macrossan test was “unsatisfactory,” “misleading,” and had created a “discrepancy” between UK and European patent law that needed urgent correction.
“The previous approach had led to a 'tortuous and artificial' distinction that was out of step with international standards and had served to disadvantage UK innovators.”
The Supreme Court has now brought the UK into alignment with the European Patent Office's “any hardware” approach. This means that if a system or method requires any physical hardware to operate, it is now fundamentally considered a patentable “invention.” The focus shifts from whether it's patentable to how novel and inventive it is.
There may still be some room for differences in approach between the EPO and UKIPO following the initial hardware determination. While the EPO decides inventive step using the problem-and-solution approach, the UK uses a different Pozzoli test for inventive step. The Supreme Court was explicit in saying that this could still be used, with the caveat that an intermediate step would be required (just as in the EPO) to “filter out features which do not contribute to the technical character of the invention viewed as a whole”. It suggested, without instructing the UKIPO outright, that one approach to “technical character” might be “to identity the alleged ‘inventive concept’ of the claim, as explained in Pozzoli.”
Emotional Perception’s application related to an Artificial Neural Network (ANN) providing media (e.g. song) recommendations based on emotional responses. It described how an ANN could be trained on songs characterised both by user tags describing their emotional effects and physical characteristics of the song (e.g. tempo, tone, volume and rhythm). The ANN would use machine learning to determine similarities in emotion and similarities in physical characteristics, and once trained, it would then be able to recommend new songs with emotional similarities to one another based on their physical characteristics.
The Supreme Court held that an ANN “constitutes, in essence, a set of instructions to manipulate data in a particular way so as to produce a desired result. In other words, an ANN is a program for a computer.” However, because the claim also involved other elements such as a database, a communications network and a user device, hardware was also required by the claim, so that the claim was not to a “computer program as such”. The case has been sent back to the UKIPO to evaluate the inventive step, with the caveat that there must be an intermediate step to decide which features (both technical and non-technical) are to be used in the determination of inventive step.
This isn't a free pass for every piece of code, but it dramatically lowers the initial hurdle. The gate that kept many software innovations out has been dismantled.
Yesterday's ruling isn't just a win for Emotional Perception AI; it's a victory for the UK tech ecosystem. It provides the legal certainty that has been sorely lacking, encouraging more investment, innovation, and ultimately, greater protection for the groundbreaking work happening in AI and software development across the country.
