The U.S. Patent and Trademark Office still has its hands full over the question of the Supreme Court’s views on life science patents, but the agency must now begin to grapple with an entirely new set of dilemmas. PTO recently announced it is seeking feedback on the status of artificial intelligence (AI) as an inventor and as a patent owner, two questions that may defy answer in the near term.
U.K. Researchers Argue for Change
Researchers at the University of Surrey in Surrey, U.K., announced at the beginning of August that they had filed for patent protection for two inventions autonomously created by an algorithm known as DABUS. The inventions are for relatively simple products – one is a beverage container with a design based on fractal geometry, while the other is a device for attracting the attention of would-be rescuers – but the team at Surrey has applied for patents for these items in the patent offices for the U.K., the U.S., and the European Union.
According to the statement, the U.K. Intellectual Property Office has already concluded that the inventions pass the preliminary test for inventiveness, although the agency has not yet taken up the question of whether non-human entities can be cited as an inventor. So far, this is as much as the researchers at Surrey are asking the agency to consider.
Ryan Abbott of the University of Surrey School of Law said in the statement, “there would be no question the AI was the only inventor if it was a natural person,” and argued that DABUS should be listed as an inventor. Abbott said developers of an algorithm should be designated the assignee or owner of any patents produced by that algorithm, arguing that such an outcome would reward innovation and “keep the patent system focused on promoting invention by encouraging the development of inventive AI, rather than on creating obstacles.”
The question is not an entirely novel one as indicated by a discussion of the question in a law journal last year, but the difficulty for patent offices is that the statutes under which these agencies work seem to leave them with little leeway. The statute in the U.S. states that a patent can be awarded to “whoever” invents something useful, while the law in the U.K. make reference to “persons” as inventors. Thus much of the debate is likely to center on a need for legislation.
PTO Broaches Ownership Question
The American patent office announced Aug. 27 that it seeks feedback on whether further guidance from the agency is needed to “promote the reliability and predictability” of patent applications filed on behalf of AI. PTO also poses the question of whether new forms of patent protection are needed, adding that some of the issues surrounding software inventions are relevant for the discussion of AI patents.
There is a need for clarity regarding terminology as the PTO noted that the term “AI inventions” is used to denote both inventions that utilize AI and those that are developed by AI. PTO also is inquiring into the circumstances in which humans might be designated co-inventor with the algorithm, as well as whether the laws and regulations are in need of revision to address the question of inventorship.
However, PTO went further, asking whether an algorithm can and should be allowed to be designated the owner of the invention. As intractable as such a problem may be, some of the more prosaic questions are likely to prove challenging as well, such as whether a change is needed in the written description requirement regarding the level of detail provided about the algorithm. PTO noted that this question could prove difficult to answer, given that some deep learning systems may have layers of functions that are obscured, and that some functions may evolve without human assistance or intervention. PTO is taking comment through Oct. 11, but it seems likely that the debate over AI inventorship and ownership has only just begun.