Series
Blogs
Series
Blogs
The UK High Court has ruled that an AI invention comprising an artificial neural network (ANN) for providing media recommendations was not excluded from patentability. While section 1(2)(c) Patents Act 1977 excludes “a program for a computer…as such” from the scope of patentable subject matter, Sir Anthony Mann held that an ANN is not a computer program, and, even if it were, this one had sufficient technical effect to escape the exclusion. This is the first UK High Court case to consider patent protection for ANNs.
ANNs are machine learning models which can be trained to process inputs in a certain way, learn from that training, and then process inputs in a way derived from that training and learning. ANNs can exist as hardware ANNs or as a software-enabled computer emulation. In this case, Emotional Perception AI (EPAI) sought a patent for an ANN-based system which uses natural language processing and machine learning to provide media file recommendations to end users, in particular songs which are similar to those being listened to by the end user.
In a decision of 22 June 2022, the UK Intellectual Property Office held that EPAI’s ANN-based system had no “technical effect over and above the running of a program on a computer” and was therefore excluded from patentability under Section 1(2)(c) Patents Act 1977, which excludes “a program for a computer…as such” from the scope of patentable inventions. EPAI appealed to the UK High Court on the basis that its ANN-based system is not a computer program and that its invention makes a technical contribution in any event.
Sir Anthony Mann, sitting as a Judge of the High Court, considered whether an ANN is excluded from patentability as a computer program as such.
Sir Anthony agreed with EPAI that the ANN is not a computer program. While there is a computer program involved in training the ANN, there is a difference between that program and the fully trained ANN, which does not implement a series of instructions given to it by a human but operates according to something that it has learned itself. EPAI’s patent claims were broad enough to cover both hardware ANNs and software-enabled computer emulations, but Sir Anthony held that, even in the latter case, the emulated ANN operates, in substance, at a different level from the underlying software on the computer. In his own words, “[w]hat is said to be special is the idea of using pairs of files for training and setting the training objective and parameters accordingly. If that is right, and I consider it is, then the actual program is a subsidiary part of the claim and is not what is claimed”.
In case he was wrong in concluding that the ANN is not a computer program, Sir Anthony went on to consider whether the ANN had a technical contribution outside of what may constitute a computer program, such that the ANN fell outside the exclusion of a computer program “as such”. Sir Anthony acknowledged that the meaning of “technical” in this context is “elusive… and that boundaries are imprecise”. Nevertheless, following a review of the authorities which set out a number of ‘signposts’ for assessing whether a relevant technical contribution exists outside of a computer program, he concluded that the end result of EPAI’s system (i.e. the sending of a file recommendation to an end user) was an external technical effect which prevents the exclusion from applying. Therefore, even if EPAI’s ANN did constitute a computer program, it would still be patentable.
This landmark decision will be welcomed whole-heartedly by AI inventors and supports the UK’s ambitions to become a global leader in the field of AI governance. It also provides a helpful indication of how the UK courts will approach questions of patentability of ANNs going forwards. However, given the importance of the subject matter, we would not be surprised if the decision was appealed.
Watch this space…