Artificial intelligence (AI) is an increasingly prevalent phenomenon. Hundreds of millions of pounds are being invested into the development of AI-based technology, and patent applications have increased exponentially.

The framers of the European Patent Convention excluded ‘computer programs as such’ from patentability, however did not define the scope of this exclusion, leaving it to be determined in case law.

In the UK and Europe, whether subject matter is deemed patentable has come to turn on whether it is considered to be ‘technical’. The notion of technicality, which has no statutory or policy basis, has been interpreted differently by the European Patent Office (EPO) itself over time, and as between the EPO and UK.

In this discussion, case law developments will be considered with regard to their likely effect on the inherent patentability of AI-related subject matter in Europe as compared to the UK.

It will be argued that changeable notion of technicality does not provide a rational basis for determining whether subject matter is patentable or not.

An alternative approach to considering whether subject matter should be considered patentable will be proposed, in the context of AI-related subject matter, which is based on social utility.