Patents owned by AI – much ado about nothing
Justice Beach of the Federal Court of Australia has decided that an artificial intelligence system can be an “inventor” for the purposes of the Patents Act. This is out of step with decisions in the US, UK and Europe. So, does this decision make Australia a world leading pioneer or an awkward outlier?
In Thaler v Commissioner of Patents (2021) AIPC 92-594;  FCA 879, the invention disclosed by the patent is described as “a container for food products including a series of fractal elements enabling multiple containers to be coupled together by inter-engagement, improving grip and heat transfer into and out of the container.”
Dr Thaler, the applicant in the patent proceedings, owns and operated a device known as DABUS and Dr Thaler owns the copyright in the source code presumably utilised by DABUS in its work that resulted in the invention. Dr Thaler named DABUS as the inventor of the invention for the purposes of patent protection but he remained the applicant for the patent.
The patent application was rejected by the Commissioner at the examination stage because an “inventor” must be a natural person and Dr Thaler appealed to the Federal Court. Some people have asked me if naming DABUS as the inventor was a mistake by Dr Thaler. While I do not know Dr Thaler and did not act in this matter, and on the basis that similar litigations of this kind are occurring across the world, I expect it was not a mistake but a test case.
In the Federal Court, Justice Beach was then forced to consider the meaning of “inventor” in the Patents Act and found that it is not defined there. The Commissioner of Patents put to the court that the meaning should be understood to be ‘inherently human’ and that the act implicitly then requires the inventor to be a natural person which is inconsistent with the inventor being a machine.
Finding that there was no specific provision in the Patents Act, Beach J agreed that AI could be an ‘inventor’, and that this understanding was economically beneficial and aligned with other tests of patent law that evolved over time to encompass new technologies. There does not seem to be a real concern about assignment of rights from DABUS to Dr Thaler on the basis that Dr Thaler owns the machine.
To be clear DABUS was the inventor here; not the applicant who is ultimately issued with the grant of a patent (not the inventor). Why? Because the patent applicant and ultimately the patentee must be human. And that’s the important part really.
While perhaps a fun, exercise for the hippocampus of the academically inclined amongst us, for the rest of us, the question is, how useful is this result and what does it all really mean? As discussed by Justice Beach, the decision is practically useful in industries like manufacturing and pharmaceutical research where a great deal of invention is created by artificial intelligence such that it would be incorrect and/or difficult to identify the person to whom the invention should be credited. As Justice Beach himself said, ‘If the output of an artificial intelligence system is said to be the invention, who is the inventor? And if a human is required, who? The programmer? The owner? The operator? The trainer? The person who provided input data? All of the above? None of the above? In my view, in some cases it may be none of the above. In some cases, the better analysis, which is consistent with the s2A object, is to say that the system itself is the inventor. That would reflect the reality. And you would avoid otherwise uncertainty.’
The same litigation is being run across the world with the UK and European Patent Offices having already rejected non-human inventors and it is straight out prohibited in the USA. So, it’s fair to say we are currently an awkward outlier but there is more to come so watch this space.