High Court Rules AI System with Neural Network Not Excluded as Patent, Highlights Technical Contribution in Patent Law

Citation: [2023] EWHC 2948 (Ch)
Judgment on


In the recent judgment of Emotional Perception AI Ltd v Comptroller-General of Patents, Designs and Trade Marks [2023] EWHC 2948 (Ch), the English High Court engaged with the intricacies of patent law in the context of an artificial intelligence system designed to recommend media files based on human perception. The case provides a pertinent exposition of the legal principles concerning the patentability of computer programs and the notion of technical contribution within the framework of the Patents Act 1977. This article critically analyses the judgment, highlighting the pertinent legal principles and the court’s application thereof to the facts at hand.

Key Facts

The appellant, Emotional Perception AI Ltd, sought to patent a system leveraging an Artificial Neural Network (ANN) to recommend media files, such as music, based on semantic similarity rather than genre. The ANN was trained through a method that allowed for the adjustment of output vectors to align with semantic distances in a semantic space.

During the trial, the court had to decide whether such usage of the ANN fell under the exclusion of “a program for a computer … as such” according to section 1(2)(c) of the Patents Act 1977. The UK Intellectual Property Office (UKIPO) had previously denied the patent, concluding that the invention was effectively a computer program.

The primary legal framework considered in this case was the statutory exclusion of computer programs from patentability under the Patents Act 1977. The High Court applied the four-step test from Aerotel Ltd v Telco Holdings Ltd [2007]:

  1. Construe the claim.
  2. Identify the (alleged) actual contribution.
  3. Assess whether it falls solely within the excluded subject matter.
  4. Determine whether the contribution is actually technical in nature.

Additionally, the court examined the five signposts from AT&T Knowledge Ventures v Comptroller of Patents [2009] to aid in the assessment of technical contribution.

A critical aspect of the legal analysis was determining the nature of the ANN, whether software emulation could be considered a computer, and if the ANN’s training process involved a computer program within the context of the exclusion.


The court concluded that:

  • An ANN, whether in hardware or as a software emulation, secures data in a way that is beyond mere software execution and emerges as a creation through self-learning; thus, it does not constitute a program for a computer for the purposes of the exclusion.
  • The ANN’s training involved some computer programming, but the claim did not pivot on this programming aspect, delineating a broader idea.
  • The patented system produced external technical effects, including sending improved file recommendations, that constituted a non-excluded technical contribution. Thus, the final product of the ANN’s output was deemed more than a mere subjective cognitive effect.
  • The system was not just outputting information but was actively selecting and transferring files based on learned technical criteria, taking the invention outside the realm of being a program for a computer “as such”.


The ruling in Emotional Perception AI Ltd v Comptroller-General of Patents, Designs and Trade Marks underscores the nuanced legal landscape for AI and computer-related inventions vis-à-vis patent law. It emphasizes the possibility of patent eligibility for AI systems that perform technical tasks ensuing in outputs extending beyond the conventional understanding of a computer’s internal operations. The decision presents a key development in supporting the intellectual property rights of innovations in the burgeoning AI field, delineating the boundaries where technical contributions from AI systems can be considered patent-worthy.

Related Summaries