Artificial intelligence as a tool for research and development in European patent law

Artificial intelligence (“AI”) is increasingly fundamental for research and development (“R&D”). Thanks to its powerful analytical and generative capabilities, AI is arguably changing how we invent. According to several scholars, this finding calls into question the core principles of European patent law—the field of law devoted to protecting inventions. In particular, the AI revolution might have an impact on the notions of “invention”, “inventor”, “inventive step”, and “skilled person”. The present dissertation examines how AI might affect each of those fundamental concepts. It concludes that European patent law is a flexible legal system capable of adapting to technological change, including the advent of AI. First, this work finds that “invention” is a purely objective notion. Inventions consist of technical subject-matter. Whether artificial intelligence had a role in developing the invention is therefore irrelevant as such. Nevertheless, de lege lata, the inventor is necessarily a natural person. There is no room for attributing inventorship to an AI system. In turn, the notion of “inventor” comprises whoever makes an intellectual contribution to the inventive concept. And patent law has always embraced “serendipitous” inventions—those that one stumbles upon by accident. Therefore, at a minimum, the natural person who recognizes an invention developed through AI would qualify as its inventor. Instead, lacking a human inventor, the right to the patent would not arise at all. Besides, the consensus among scholars is that, de facto, AI cannot invent “autonomously” at the current state of technology. The likelihood of an “invention without an inventor” is thus remote. AI is rather a tool for R&D, albeit a potentially sophisticated one. Coming to the “skilled person”, they are the average expert in the field that can rely on the standard tools for routine research and experimentation. Hence, this work finds that if and when AI becomes a “standard” research tool, it should be framed as part of the skilled person. Since AI is an umbrella term for a myriad of different technologies, the assessment of what is truly “standard” for the skilled person – and what would be considered inventive against that figure – demands a precise case-by-case analysis, which takes into account the different AI techniques that exist, the degree of human involvement and skill for using them, and the crucial relevance of data for many AI tools. However, while AI might cause increased complexities and require adaptations – especially to the inventive step assessment – the fundamental principles of European patent law stand the test of time.