Artificial intelligence (AI) has recently not only been finding its way into more and more applications and areas of life, but also poses new questions and challenges for patent law.
One particular question currently being pushed forward by patent applicant Stephen Thaler in a number of jurisdictions is whether the respective patent law allows the “name” of an AI system to be designated as inventor instead of a natural person. Thaler is the developer and owner of an AI system he calls “DABUS” (“Device for the Autonomous Bootstrapping of Unified Sentience”), which is based on artificial neural networks. According to Thaler, the very processes that qualify a natural person as an inventor, namely the conceiving or devising of the invention, run automatically and without human intervention in his AI system, so that naming “DABUS” as inventor would be justified.
The only patent office so far that did not see a problem in naming “DABUS” as inventor and that granted a patent in the middle of this year is the Patent Office of South Africa. All other patent offices have rejected the applications due to a defective designation of the inventor.
However, the Australian Patent Office’s rejection decision was recently overturned by the Australian Federal Court on the grounds that the Australian Patents Act does not contain an exhaustive definition of the term “inventor” and that a broad interpretation would also be consistent with the Act’s objective of promoting technological innovation and its publication and dissemination irrespective of whether the invention was made by a human or a machine. However, an Australian patent has not yet been granted, as the Office has appealed the Federal Court decision.
In the US as well as in the UK the rejection decisions were recently upheld by the District Court For the Eastern District of Virginia and the London Court of Appeal, respectively. In the US case, the court focused on the fact that the plain wording of the law refers to the inventor as an “individual”, from which it could be concluded that only natural persons could be inventors within the meaning of the law. Moreover, it was not within the competence of the courts to interpret the wording of the law more broadly in order to take into account political considerations (such as strengthening incentives for innovation). Rather, it was up to the legislature to amend the wording of the law if it wished to take (greater) account of such policy considerations.
The matter was not quite so clear-cut in the UK, where the legal text merely states that inventor means “the actual deviser of the invention”, but without giving any indication that this could only be a natural person. Only two out of three judges voted to uphold the rejection decision. In both the USA and the UK, it is still open to the applicant to take the case to the next higher instance.
Appeal proceedings brought by Thaler are also currently pending before the German Federal Patent Court and the Boards of Appeal of the European Patent Office.
The question of whether an AI can be an inventor is interesting because it might also have an impact on substantive issues of patent law, in particular on the requirements for inventive step or also on questions regarding the ownership of the patent right.