When the UK Supreme Court issued its landmark ruling in Emotional Perception AI v Comptroller General, it scrapped the UK’s traditional, restrictive software patent test in favour of European-style rules. By establishing that artificial neural networks pass initial patent eligibility checks as long as they involve basic computer hardware, the ruling was widely celebrated as a major win for the tech sector. Much of the commentary at the time suggested the UK had opened the floodgates for AI patents. However, following the Supreme Court’s remittal of the case back to the UK Intellectual Property Office (UKIPO), the office has now formally refused the patent application—pouring cold water on that initial over-excitement. The decision can be accessed here for those so inclined.
To my mind, this outcome was always the logical result. The Supreme Court aligned UK practice directly with European Patent Office (EPO) jurisprudence, and the EPO has long been consistent in rejecting "subjective" or cognitive improvements. In fact, we highlighted the risk that this patent would be considered non-technical in our earlier commentary.
While getting an AI application past initial eligibility hurdles is now straightforward, an application will still be rejected if the underlying machine learning innovation is focused on subjective outcomes—such as recommending media based on human emotion—rather than solving a tangible technical problem.
Given the applicant’s history of vigorously contesting adverse decisions, I fully expect this ruling to be appealed to the High Court. However, they face a formidable climb. The Supreme Court set out to align UK law directly with established EPO practice, and the UKIPO’s decision merely carries that mandate to its logical conclusion. To overturn this ruling, the applicant will not just need to overturn the Hearing Officer—they would effectively have to convince the UK courts to break away from decades of settled European jurisprudence.
Subscribe to receive more articles like this here.

