The DABUS Affair

Abstract

The technical capabilities of Artificial Intelligence (AI) systems continue to be the subject of extensive analysis and research, especially for those attempting to verify the hypothesis of whether it would be possible to attribute inventorship/ authorship to an entity that lacks legal capacity and the justification for such attribution. There are both factual and legal questions that need to be answered first before recognising AI systems as inventors/authors under patent/copyright law .

To this extent, we propose to further analyse the DABUS case both through its shortcomings and the possibilities it may create for the evolution of the legal framework around regulating AI systems and potentially defining their legal status.

In the European Patent Office’s (EPO) experience, the patent system, in its current form, is fully equipped to deal with the legal issues around arising inventions involving AI. Because AI is considered a branch of computer science, inventions involving AI are considered computer implemented inventions in patents law terms. However:

i. Can AI be seen as an inventor?
ii. Can AI invent independently?

Following the patentability test, in the DABUS case, most countries, except for a few where this application was pending, have rejected the application based on the understanding that the inventor of a patented innovation needs to be a human being. This is closely linked to the concept that the patent is a right, and the society is not giving rights or obligations to machines or computer programs for the time being, regardless of how intelligent they might be. This is also related to the principle of human centric approach applied in Europe and reflected in the most recent accents and legislative proposals of the European Union (EU).

Keywords: Artificial Intelligence, AI-generated works, DABUS, patent law, copyright law, autonomy

JEL Classification: K13, K20, O3