Patent-Eligible Invention Requirement Under the European Patent Convention and its Implications on Creations Involving Artificial Intelligence

Artificial Intelligence and its subfield, Machine Learning are areas of computer science; thus, they rely on algorithms, models, computer programs and software applicable in numerous areas. Since respective creations involve resources and shift from hardware to software, there is an incentive to protect them legally. Due to their dual nature, the algorithms, models, computer programs, and software might be too “technical” to avail copyright protection but not “technical” enough for a patent. Whereas trade secret protection might not be sufficient means of protection in all cases.The article explores the issues and, as its main argument, builds further on the academic proposals on the sui generis mechanism. It also suggests certification as the potential approach to avail the desired protection instead of diluting the existing protection frameworks. An alternative would be to lie on the complete availability or trade secret protection, none of which would be an adequate balance.