The question of whether artificial intelligence (AI) can be named as an inventor on patent applications is sparking legal debates around the globe and made its way into the discussions of AIPPI World Congress this year.
The most notable case in this regard, involving AI system “DABUS” and its creator Dr. Stephen Thaler, continues to challenge traditional notions of inventorship across several jurisdictions. While some courts have upheld existing frameworks that require inventors to be human, others are reconsidering the rules considering evolving technology.
In 2021, South Africa became the first country to grant a patent listing DABUS, an AI, as the inventor. Since South Africa does not conduct substantive examinations, any rejections thus far have been based on formalities. The South African patent office accepted the application that identified AI as the inventor and issued the patent to Thaler, the owner of the system.
In a significant development soon after, a judge in the Federal Court of Australia issued a detailed ruling that an AI system could be recognized as an inventor under Australian patent law. This interpretation was based on a broad statutory analysis that did not strictly limit inventorship to humans. However, this decision was later overturned by the Federal Court, and the High Court of Australia chose not to hear the case.
Ryan Abbott, a professor of law and health studies at the University of Surrey, was one of the speakers at AIPPI’s morning discussion on October 20. He also led the legal team that represented Thaler in various jurisdictions. He explained that some jurisdictions like Israel and Austria don’t require an adventure to be designated at all. Other jurisdictions use all sorts of varied language about the importance of inventorship or what it means to be an inventor, “some of which use human centric terms and some of which don’t,” said Abbott.
“Most inventions aren’t owned by their inventors, but rather their employers,” Abbott said at a U.S. Senate Subcommittee hearing in 2023.
In contrast, the United States has held firm in its requirement that only human individuals can be inventors. In 2022, the U.S. Court of Appeals for the Federal Circuit ruled that the U.S. Patent Act defines an inventor as an “individual,” meaning a human being. The court referred to precedent, pointing out that an “individual” is a person under U.S. law unless otherwise specified.
Europe’s position has been equally nuanced. The European Patent Office (EPO) has rejected applications listing AI as the inventor, stating that an inventor must be a natural person. However, the legal framework allows member states some leeway in how they interpret inventorship. According to Abbott, a divisional application was filed in Europe, but the Formalities Division rejected it again on the basis that an inventor must be human. Abbott said a hearing of the formalities office is expected to take place in December. The UK has taken a similar approach. Although the case reached the UK Supreme Court, the ruling was clear: only a human can be listed as an inventor.
In China, CNIPA also ruled that inventors must be natural persons, due to the rights of identifying inventors falls under General Provisions of Civil Law, which includes natural persons, legal persons and unincorporated organizations. As DABUS, an AI system, does not fall into any of these categories, it therefore cannot be recognized as the inventor.
“The U.S. allows non-humans to be authors of things for copyright, for example, corporate authorship. And the argument is, patents are designed to do three things, encourage innovation, encourage disclosure of things that would otherwise keep as a trade secret, and encourage the commercialization of new products,” Abbott explained, he added that all of which would be facilitated by recognizing patents on AI generated inventions because it will encourage people to make and use inventive AI and R&D.
These cases underscore the broader challenge of reconciling traditional patent laws with the capabilities of AI. Patents, by design, are meant to encourage innovation, protect intellectual property, and promote the commercialization of new products. Many argue that excluding AI from the inventorship process undermines those goals, particularly in sectors like pharmaceuticals, where AI is already transforming research and development.
According to Abbott, AI is going to help society be better at R&D, and in a couple of decades may automate large fields of R&D. If companies can’t patent AI-generated inventions, they may instead keep them as trade secrets. This, he warned, would circumvent the purpose of the patent system, particularly in industries like pharmaceuticals where a large portion of investment occurs after the inventive stage, such as in clinical trials.
Recent developments in Germany further complicate the picture. A decision by the highest court in Germany, issued earlier this year, upheld the idea that AI-generated inventions could be protected, provided that a human – such as the owner or user of the AI – was listed as the inventor. The case has now been sent back to the German Patent Office, and preliminary indications suggest that the office may accept this approach.
“The AI did most of the work. You know, there are AI companies saying basically that someone with no training in AI or drug discoveries can use these platforms to make new drugs. So really, I think a lot of questions remain about how all of this is going to play out litigation. And my view of it is that this sort of thing really shouldn't be a barrier to accurately if AI can do a better job of inventing, driving a car or operating an elevated human being, that's something we want to encourage. It's something we want patent law to encourage,” said Abbott.
The ongoing discussions at forums such as the AIPPI World Congress highlight the urgent need to reconcile patent laws with technological advancements. As AI becomes increasingly integral to research and development, especially in fields like pharmaceuticals, addressing these legal uncertainties is crucial.
– Cathy Li, reporting from Hangzhou
