Patenting Algorithms in an Internet of Things and Artificial Intelligence World: Pathways to Harmonizing the Patentable Subject Matters and Evaluation of the Novelty Requirement

Constituting a disruptive technology, Artificial Intelligence (‘AI’) is impacting all industries. However, there are rising concerns that the patent system may not be fit for the future of innovation that is increasingly AI-related and intangible. Indeed, the execution of AI-related inventions requires some kind of computer implementation, thereby potentially reviving patentability issues related to computer-implemented inventions. Whilst patent offices around the world have found ways to adapt their patent systems to grant protection to software, difficulties remain in relation to algorithm-based inventions though they form a significant part of today’s innovation. Currently, algorithms themselves do not qualify as patentable inventions. Even if algorithms overcome this first hurdle, concerns arise in relation to the application of patentability requirements such as novelty where national differences remain. This research evaluates the adequacy of the novelty requirement in relation to AI-inventions where many of the underlying concepts and technologies are not novel. The ultimate aim is to evaluate the adequacy of the patent system by looking at inventions that utilize AI, with a particular focus on the excluded subject matters and the novelty requirement. To this end, the research adopts a comparative analysis of these concepts in Europe (EPC countries), Japan and the United States.