Artificial intelligence (AI)-generated inventions, or those produced entirely or partially by computer systems, are thought to be increasing in frequency as technology develops. In these creations, human intelligence is less evident, but the process of inventing is made simpler because the AI handles the majority of the mental labour. This situation, however, makes it more difficult to determine if the creation has an "inventive step," which is a requirement for patent eligibility that stipulates the innovation must be something that a competent individual would not know about (or difficult for a skilled person to invent). Indeed, a talented individual may not notice a particular AI-generated invention, but someone with access to a comparable AI is likely to see it. This study's main objective is to ascertain if patent laws are appropriate for determining if inventions produced by AI are patentable, specifically in relation to the creative action criterion. In order to achieve this, the study compares the creative phase (or lack of obviousness) requirements in the US, the EU, and Japan.
