The Supreme Court’s decision in Emotional Perception AI Ltd v Comptroller General of Patents marks the most significant change to UK computer-implemented invention jurisprudence in two decades. In adopting an approach aligned more closely with EPO jurisprudence and G 1/19, the Court held that the four-stage Aerotel test should no longer be followed. Instead, the analysis focuses on the technical character of the claimed invention, including an intermediate assessment of which features contribute to that technical character before substantive patentability requirements are considered.
As the UKIPO’s updated examination guidance recognises, the question is no longer whether a claimed contribution passes the Aerotel framework, but whether the invention, considered as a whole, contains features that contribute to technical character in a manner recognised by Article 52 EPC jurisprudence.
The obvious starting point for the new intermediate step, namely identifying features capable of contributing to the technical character of an invention, is the existing EPO Board of Appeal jurisprudence. In practice, the Supreme Court’s endorsement of G 1/19 appears likely to bring UK law much closer to the COMVIK approach applied by the EPO. The focus therefore shifts to identifying which features contribute to technical character and which instead relate to non-technical aims, such as commercial, administrative or cognitive requirements.
In general, EPO case law indicates that the following are considered “technical”:
- Processing measurements from the physical world, such as sensor, radar or medical imaging data (e.g. T 26/86)
- Improvements to the internal functioning of a computer, network or other technical infrastructure. (e.g. T 1351/04)
- Simulations used for a technical purpose, particularly in engineering and design. (e.g. T 1227/05)
- AI and mathematical methods when applied to a technical problem, such as image processing (e.g. T 208/84).
Conversely, EPO case law generally treats applications directed to the following matters as non-technical:
- Business, financial and commercial methods. (e.g. T 641/00)
- Administrative or organisational rules (e.g. T 154/04).
- Mathematical methods performed on abstract data.
- Recommendation engines, forecasting systems and decision-support tools whose output is information for a human user (e.g. T 306/10)
Notably, none of the EPO categories is technology-specific. AI, simulation, modelling and data analysis appear on both sides of the technical/non-technical divide. What matters is not the technology being used, but whether the claimed features contribute to a technical purpose. The same machine-learning technique may therefore be technical in one context and non-technical in another.
Under the COMVIK approach, the decisive question is not whether a claim contains technical features, which most computer-implemented inventions inevitably do, but whether the features relied upon for inventiveness provide a technical solution to a technical problem. It is against that background that the historic UKIPO decisions become particularly interesting.
Previous UKIPO decisions were decided using the Aerotel test, whereas the post-Emotional Perception analysis focuses on identifying features that contribute to the technical character of the claimed invention. Although the analytical framework has changed, the underlying enquiry is often directed at similar questions. In Emotional Perception itself, the Supreme Court noted that, although Aerotel was no longer to be followed, there was likely to be little change in what was and was not patentable, and no examples of a change in outcome had been identified to the Court. That observation provides an important reason for examining historic UKIPO decisions: while they were decided under a different legal framework, they may nevertheless offer useful evidence of the types of problems that the Office has historically regarded as technical.
Against that background, it is instructive to compare EPO jurisprudence with UKIPO decisions made under the Aerotel framework to determine whether similar patterns emerge.
Technologies that have been historically successful before the UKIPO
Industrial process control and engineering systems
Perhaps the clearest category of cases to succeed before the UKIPO comprises inventions operating directly within industrial environments.
An example is Fisher-Rosemount Systems (O/141/21), which concerned the management of maintenance activities within a process plant, including identification of suitably qualified personnel, provision of process diagrams and automated verification of completed work. In that case the hearing officer considered the interaction with the process plant itself to provide the necessary technical contribution.
Similarly, in General Electric (O/029/19), the invention analysed generator sensor data to identify faults and anomalies. The hearing officer concluded that the invention went beyond a computer program as such and remitted the application for further examination.
Signal processing, sensing and measurement
Applications involving the processing of measurements derived from the physical world also fare well.
In Accipiter Radar Technologies (O/390/17), the invention concerned a three-dimensional radar system capable of real-time tracking and analysis of airborne targets. The hearing officer found that the contribution lay not merely in the presentation of information but in the operation of the radar system and associated database architecture for generating previously unavailable information in operational timeframes.
Likewise, in Hitachi (O/809/18), a construction-site management system used laser-scanned three-dimensional representations of a site to identify congestion and safety issues. The hearing officer accepted that the claimed alerting system produced a technical contribution.
These decisions suggest that the processing of measurements obtained from physical systems remains one of the strongest indicators of technical character.
Engineering simulation and physical modelling
A further category of successful cases concerns simulation linked to physical systems.
In Landmark Graphics (O/112/18), seven applications concerning visualisation and manipulation of subterranean geology were found not to relate merely to excluded subject matter. The hearing officer considered that the inventions addressed technical issues arising in the modelling of subsurface formations.
A second group of applications from the same applicant in Landmark Graphics (O/138/18) concerning computer-implemented geological modelling similarly resulted largely in success, with three of four inventions being held non-excluded.
These decisions are notable because they demonstrate that software modelling can be technical where it is sufficiently connected to physical reality.
Technologies that have historically struggled before the UKIPO
Business and commercial systems
Business methods remain among the least successful categories.
In Google (O/0115/23), a privacy-preserving system for gathering store visit analytics was refused because the contribution was characterised as a business method implemented by conventional computing means.
Likewise, in Reaux-Savonte (O/0722/25), a distributed ledger system involving transaction validation and reward mechanisms was found to relate to a business method and computer program as such.
Notably, neither privacy features nor blockchain architecture were sufficient to transform the inventions into technical solutions in the eyes of the hearing officers.
Information analysis and decision support
The UKIPO also consistently rejected inventions directed primarily to deriving information for human decision-making.
In IBM (O/0528/25), an invention for identifying molecular dynamic states from large datasets, potentially useful in drug discovery, was refused because the hearing officer found no technical effect beyond analysis of information.
Similarly, healthcare decision-support systems predicting patient treatment requirements have struggled where the outcome is ultimately information for a clinician rather than an improvement in a technical process.
Borderline technologies
The most informative decisions are often those where similar technologies appear on both sides of the line.
Geological modelling
Landmark Graphics provides a striking example.
As noted above, in O/112/18 and O/138/18, geological modelling inventions were largely allowed because they addressed technical issues associated with modelling subterranean formations.
However, in a later case, O/1058/23, Landmark’s application for a seamless geological model spanning scales from planetary down to pore level was refused as a computer program as such.
The distinction appears to lie in whether the invention is directed to solving a technical problem associated with modelling physical reality, or merely providing a more sophisticated representation of information.
Medical technology
Medical inventions also divide sharply.
Where inventions relate to the processing of physical image data or operation of imaging systems, they tend to be viewed as technical. By contrast, where machine learning is used to predict outcomes, allocate resources or assist clinicians in decision-making, UKIPO decisions have often treated the contribution as lying in the provision of information to a clinician rather than in the solution of a technical problem.
Comment
Viewed through the post-Emotional Perception framework, the principal lesson from these UKIPO decisions is that the decisive issue was rarely the technology itself. AI, simulation, data analytics and modelling all appear in both successful and unsuccessful cases.
Instead, the recurring question even under the Aerotel test was whether the invention addresses a technical problem in a technical context.
Applications that improve the operation of industrial systems, process physical measurements, control equipment or solve engineering problems have always generally been successful, regardless of whether those applications are pursued before the UKIPO or the EPO. Conversely applications that organise information, support decision-making, optimise business activity or provide recommendations have generally failed.
UK practice will undoubtedly develop as the UKIPO applies the approach mandated by Emotional Perception. Nevertheless, given the lack of any significant change in subject matter which is considered patentable by the office, these older decisions may provide useful indicators of the types of features likely to be regarded as contributing to technical character at the new intermediate assessment stage. While Emotional Perception has changed the legal framework, it may not dramatically alter the types of inventions that succeed in practice. The historic UKIPO decisions discussed above suggest that inventions directed to technical systems, technical measurements and technical processes are likely to remain on the right side of the line, whereas inventions directed primarily to human decision-making, business activity and information processing may continue to face significant challenges.
