Recently, the world of intellectual property has witnessed one of the most momentous decisions of the last decade in the United Kingdom. The Supreme Court’s judgment in the case of Emotional Perception AI Limited v. Comptroller General of Patents has marked the end of an era, abandoning the Aerotel test for the patentability of computer-implemented inventions, which has long been the standard for analysing inventions of this type. With this move, the UK aligns itself with the approach of the European Patent Office (EPO), consolidated following decision G1/19 of the Enlarged Board of Appeal.
The invention at the centre of the dispute, developed by Emotional Perception AI, is a system based on an Artificial Neural Network (ANN) designed to recommend multimedia files (such as music or video) that evoke an emotional response similar to a reference file. It does this by analysing the physical and objective properties of the files and learning, through a training process, to replicate human semantic and subjective perception. The UK Intellectual Property Office (UKIPO) had rejected the application, arguing that it was a “computer programme as such”, which is subject matter excluded from patentability.
The Past: The Age of the Aerotel Test
For nearly twenty years, the patentability of software-related developments in the UK was governed by the four-step test established in Aerotel Ltd. v. Telco Holdings Ltd. (2006). This approach focused on determining whether the actual “contribution” of the invention lay exclusively in excluded subject matter. The steps were:
- Correctly interpreting the claim.
- Identifying the real or substantial contribution.
- Asking whether said contribution falls solely on excluded subject matter.
- Checking whether the contribution is of a technical nature.
This test, although well established, generated a notable divergence from the practice of the European Patent Office (EPO). The main criticism of this analysis methodology was that the Aerotel test unduly mixed the requirements of “invention” with those of novelty and inventive step by focusing on the “contribution to the state of the art”.
The European Standard: The COMVIK Approach and Decision G 1/19
In turn, the European Patent Office (EPO) has consolidated its methodology over time through the COMVIK approach (originated in decision T641/00). This approach is now the standard for assessing the inventive step of inventions that combine technical and non-technical features.
The COMVIK approach posits that, for the analysis of inventive step, only those features that contribute to the technical character of the invention should be considered. In turn, non-technical features are not ignored, but are integrated into the formulation of the technical problem as a “constraint” that the technical expert must resolve.
In decision G1/19, the Enlarged Board of Appeal confirmed the COMVIK approach as the correct and universal methodology for all computer-implemented inventions. Therefore, this decision established a two-step methodology, where the first step is to determine whether the invention is technical in nature, which is satisfied if the invention has at least one technical element. Once this low threshold has been met, the crucial analysis moves on to the analysis of inventive step, where the COMVIK approach separates the technical part from the non-technical part.
The Judgment: UK Adopts the European Model
The UK Supreme Court, in its judgment, recognised this divergence as untenable. Explicitly citing decision G1/19 and the need for harmonisation with the European Patent Convention, the court invalidated the Aerotel test.
The new British doctrine adopts the approach of the European Patent Office (EPO), which can be summarised as follows:
- First hurdle (Is it an “invention”?): The question is whether the claim is technical in nature. If it involves at least one technical element, the answer is yes. This hurdle is deliberately low, and Emotional Perception AI’s invention was considered to easily clear it.
- Second hurdle (Novelty and Inventive Step): This is the core of the new approach. All the features of the invention must be identified and those that do not contribute to the technical character of the invention as a whole must be separated. Only features that contribute to the technical character (either because they are technical in themselves or because, even if they are non-technical, they interact with the technical elements of the invention to produce a technical effect) are considered in the analysis of inventive step.
The outcome for Emotional Perception AI was the reversal of the UKIPO’s decision. Its invention is not a “computer programme as such” and must be examined for novelty and inventive step under this new paradigm.
Conclusion: What is Patentable in the Field of Software?
This landmark decision brings much-needed clarity for companies and individuals engaged in innovation.
For inventions in the field of software, patentability no longer depends on an abstract “contribution” but rather on whether the invention provides a technical solution to a technical problem. The use of a computer is no longer merely a “disguise” for a business idea, but rather the means to execute a technical solution. Software that controls external hardware, improves the internal functioning of a computer, or processes data in a technically significant way to achieve a technical purpose is clearly patentable if it is new and inventive.
In particular, for inventions in the field of software and in the field of biotechnology, this new approach allows for the protection of new inventions in fields such as bioinformatics or AI-assisted medical diagnostics, such as those that make use of Artificial Neural Networks (ANN), where software processes biological data to produce a result with a clear and tangible technical purpose: a diagnosis, a risk prediction, a classification with medical relevance. In this way, solutions can be protected where the software does not merely present information, but transforms it to solve a technical problem.
The judgment in Emotional Perception AI solidifies that, both in the United Kingdom and at the European Patent Office (EPO), the value of an invention in the field of software does not lie in the elegance of its algorithm in the abstract, but in the technical effect it achieves in the real world. For the burgeoning biomedical software and AI in healthcare industry, this is the legal certainty they have been waiting for.
Written by: Dennis Cumba Chamorro. Patent Engineer. European Patent Attorney. Patent and Technological Development Department, PONS IP
