If an artificially intelligent system invents a product, should patent authorities recognize it as the inventor of the innovation? Currently, a group of advocates, including scientists, legal experts and a professor, are trying to get the AI algorithm Dabus (device for the autonomous bootstrapping of unified sentience) recognized as the inventor for two patent applications filed in the UK, Europe, Israel and the US.
Traditionally, patents provide legal rights, like exclusive ownership, to a human inventor. Nonetheless, American engineer Stephen Thaler, the designer behind Dabus, has filed for patents in the name of Dabus AI, arguing that the algorithm deserves recognition and attribution for designing new products, thereby challenging the conventional patent systems around the world.
While Dabus relies on machine-learning methods that require vast amounts of data for its decision-making, Dabus can devise and develop new creative ideas with little human intervention, which is one of the requirements for an inventive act. Dabus’ first patent describes a food container that uses fractal designs to create pits and bulges to allow several containers to be fitted together more tightly. Dabus’ second patent describes a flashing device used for search and rescue for attracting human attention in emergencies. Dabus’ inventions possess the elements of novelty, utility and inventiveness that should satisfy the Canadian patent protection requirements.
Additionally, what is more remarkable is that “the machine invented in two different areas, neither of which its programmer had any background in”. The programmer’s lack of expertise in the subject matters of the inventions might demonstrate that Dabus is using a type of autonomous creative thinking to invent. However, the patent authorities are showing resistance in crediting AI as an inventor as there is a lack of expressly stated statutory definitions and supporting common law precedents. A spokeswoman from the European Patent Office has told BBC that it is hesitant to grant patents to AI because doing so would likely create unforeseen legal issues, “as any change would have implications reaching far beyond patent law, i.e. to author’s rights under copyright laws, civil liability and data protection”.
Currently, the Canadian Patent Act describes a patent as a government grant that gives the inventor and the patentee’s legal representatives the exclusive right, privilege and liberty of constructing and using the invention and selling it to others to be used, but it does not provide a clear definition of who, or what, could be an inventor. Regarding who can be an inventor, the UK’s Patent Act currently requires an inventor to be a person. In the U.S., the America Invents Act (AIA), 35 U.S.C. § 100(f) defines the term inventor as: “the individual or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention, excluding AI as an inventor.”
Thus, the legal and regulatory regimes need to be modernized to address the legal issues surrounding AI, which present unique challenges to the governments, companies and individuals that are developing, using or seeking to commercialize it. In particular, the intellectual property regulatory framework is still playing catch-up to the realities of the AI technology, despite the significant interest, investment and advancements in AI. In the near future, we will be seeing more AI inventors producing a variety of inventions or creative art work, as a total of 150,637 patents dealing with the development or use of AI were published in the last couple years, an impressive 175 percent growth between 2013 and 2016. Consequently, it will be very exciting to see the outcome of the Dabus case, as it could serve as a landmark test of patent law that could clarify the legal issues surrounding the statutory interpretation of “inventor.”
Written by Elif Babaoglu, Contributing IPilogue Editor and JD Candidate at Osgoode Hall Law School. Elif is also the co-director of events at the Osgoode Privacy Law Society.