15 November 2022

DABUS Down – High Court Dashes Hopes of ‘AI Inventor’ Advocates

Gavels have never been used in Australian courtsOn Friday 11 November 2022, three judges of the High Court of Australia (Gordon, Edelman and Gleeson JJ) refused – with costs – Stephen Thaler’s application for special leave to appeal a decision of the Full Court of the Federal Court of Australia determining that the (alleged) ‘AI inventor’ DABUS cannot be named as an inventor for the purposes of applying for a patent in Australia.  This brings to an end the efforts of Dr Thaler – and his ‘sponsor’, Artificial Inventor Project leader Professor Ryan Abbott – to achieve recognition for machine inventors under Australian law via the courts.  I have long argued that even if machines are capable of autonomous invention without human involvement (which remains, for the present, a dubious claim), it is far from obvious that granting patents for machine-generated inventions is economically or socially desirable.  The High Court’s refusal to weigh-in at this time places the issue back in the hands of the legislature to determine whether any change to the law is required.  However, any reforms should be preceded by a proper enquiry into the pros and cons, including consultation with experts, interested stakeholders, and the wider public.

To be fair, the High Court did not appear to completely shut the door on patents being granted for machine-generated inventions under the existing law.  The judges who heard Thaler’s application seemed open to the possibility that a human inventor might be identified in such cases, or perhaps that there might be no inventor but that someone might nonetheless be entitled to obtain a patent on the invention.  However, this particular case provides the Court with no factual basis to consider such alternative options, and the judges therefore did not consider it a suitable case in which to grant special leave to appeal.

Problems of a Test Case Based on ‘Agreed Facts’

It appears from the transcript of the special leave hearing (Thaler v Commissioner of Patents [2022] HCATrans 199) that the Court’s refusal to grant leave may have been due in no small part to the way in which Thaler’s case has been constructed from the outset.  As Sophie Goddard SC, appearing on behalf of the Commissioner of Patents, observed at the hearing:

The question whether an artificial intelligence system can be an inventor is one which has been self‑generated by this applicant worldwide, so far without much success. To create a test case, we would say, out of a theoretical question of law and promote that worldwide does not elevate the question to a matter of public importance…

This really goes straight to the heart of the matter: Thaler’s sole purpose (or, perhaps more correctly, the objective of Abbott’s Artificial Inventor Project) is to put to the test the legal theory that a patent can be validly granted from an application on which there is no human inventor, and the only named inventor is a machine.  To this end, the case was provoked in the Australian Patent Office, and subsequently run on the basis of certain ‘agreed facts’, i.e. matters that the parties stipulate should be assumed to be true for the purposes of the court proceedings.  Section 191 of the Evidence Act 1995, provides that such agreed facts are taken to be undisputed, need not be proven by evidence, and cannot be contradicted or qualified by evidence. 

In particular, the case has proceeded on the agreed basis that DABUS devised the inventions disclosed and claimed in Thaler’s patent application autonomously, and that it did so without human control or direction from Dr Thaler (or anybody else).  A consequence of this is that it has not been open to the courts to find that DABUS is not the inventor, or that there is any person who may be considered as inventor, either in place of, or jointly with, DABUS.

My reading of the transcript is that the High Court judges were not interested in taking up this question of machine inventorship in the abstract.  They appeared to take the view that, for the issue to be worthy of the Court’s consideration, they would want to be able to interrogate the underlying facts to examine, for example, whether it might be possible for the applicant (i.e. Dr Thaler) to be named as inventor, either in place of or in addition to DABUS.  As stated by Edelman J, addressing David Shavin QC appearing for Thaler, ‘the difficulty for this Court is that without having any submissions about the starting point, which is whether a natural person here could be the inventor, we are groping in the dark’. 

The Court’s preference for the case to be argued on the basis of competing theories of inventorship was made plain in the following exchange between Gleeson J and Ms Goddard:

GLEESON J: My question is really coming out of the fact that, as I understand it, Mr Shavin has made it fairly clear that [Thaler] does not wish to contend that he is the inventor. So, the proposition that he is the inventor is going to have to be propounded by someone. Would that be [the Commissioner of Patents]?

MS GODDARD: It could not be, your Honour, in this case, given the agreed facts – the way the matter has been run all the way, and the facts that were agreed from the very beginning ‑ ‑ ‑

In a subsequent exchange with Mr Shavin, Edelman J suggested that, with both Thaler and the Commissioner adhering to the agreed facts, the Court may wish to appoint a contradictor:

EDELMAN J: Your submission, Mr Shavin, ultimately comes down to the point that this Court, if it were to grant special leave to appeal, needs to consider the issue on the basis of an underlying legal assumption without a contradictor which may go to the heart of the question?

MR SHAVIN: Yes, because the question of law is, was the Commissioner entitled, merely by the appearance of the name of a non-natural person as the inventor, entitled to reject the application?

EDELMAN J: Yes, but that question of law may carry with it the question of whether or not it is possible to have an invention without an inventor. And your submission is that it is, and there is no contradictor to suggest that it might not be.

MR SHAVIN: Save as to this, your Honour. We say that the question is raised on the face without it being necessary for the Court to investigate what DABUS does. It is our respectful submission that the application having been rejected on the face without an inquiry into fact, the question is to whether that rejection, on the face, was lawful.

How Could ‘AI Inventorship’ be Tested Under Australian Law?

Further exchanges between the judges and counsel for the parties raised the question of whether, and in what circumstances, it might be possible to test ‘AI inventorship’ in Australia.  The difficulty in this case is that Thaler’s application lapsed as a result of his failure to satisfy a specific formality requirement under regulation 3.2C(2)(aa) of the Patents Regulations 1991 that the name of a (valid) inventor must be provided upon entry of a PCT application into the national phase in Australia.  The Commissioner’s determination that an inventor must be human, and that naming DABUS therefore failed to satisfy this formality, thus preceded any substantive examination of the application on its merits.

This difficulty was highlighted by the following exchange between Mr Shavin and Edelman J:

MR SHAVIN: The problem, your Honour, is this – that there is no other mechanism for getting the case to that point. Because the application has been rejected, effectively, by the Commissioner at the outset before there has been any consideration of the application, we can never get to the question of DABUS and the inventor and whether section 18 is otherwise satisfied.

EDELMAN J: There is an easy way the question could have been raised, which could have been if the applicant had listed himself as the inventor and the Commissioner and had rejected that on the basis that he was not the inventor but the artificial intelligence was the inventor, which would then have given rise to the prospect that nobody, for the purposes of section 15, was the inventor.

MR SHAVIN: The difficulty with that course … was that the inventor is encouraged to file an application nominating a fact he does not believe to be true.

If I understand correctly what his Honour is suggesting here, it would require Thaler effectively to confess to his subterfuge and amend the application to name DABUS as inventor in the course of examination (because otherwise the Commissioner would have no reason to question his claim of inventorship).  At that point the absence of a valid inventor is not a ground for objection to the application, although it would still be necessary for Thaler to demonstrate his entitlement to be granted a patent for the invention.  Under these circumstances, he might well be required to provide evidence of how the invention came into existence, and how he came to be in possession of knowledge of the invention.

I suspect that what Edelman J may have had in mind here is that such evidence might make it possible for the Court to identify a different approach that would recognise Thaler as inventor and/or entitle him to be granted a patent, e.g. via his ‘discovery’ of some interesting output generated by DABUS.  But the Court is here precluded from considering alternative options by the parties’ adherence to the agreed facts.

Conclusion – Not an ‘Appropriate Vehicle’

It might seem that the High Court was looking to address a case other than the one that was actually brought before it, but I think that it was reasonable for it to do so.  The Court’s is right to be concerned that if it were to accept the appeal constrained by the agreed facts, it would essentially be taking on a hypothetical question (i.e. one based on, in the words of Edelman J, ‘an underlying legal assumption’). 

Where the courts have previously addressed inventorship it has generally been in the context of a dispute, in which the legal question of who is rightfully an inventor is underpinned by the factual circumstances in which the invention was devised.  To accept that there is no inventor, or that the sole inventor is a machine, simply because Thaler asserts that to be so, ignores the possibility that the facts might lead to a different legal conclusion in relation to inventorship and/or entitlement.  To my mind, the Court is correct to question whether Thaler might be an inventor in this case, and to be concerned that answering the hypothetical question posed by the agreed facts may miss the real issues entirely.

Thaler has presented no evidence of precisely how DABUS came to devise the inventions disclosed and claimed in his patent application, or of what involvement he may or may not have had in that process.  And with the Commissioner accepting Thaler’s assertions at face value, there is no party to the proceedings arguing for any alternative view on the factual or legal implications of his claim that DABUS is the inventor or his denial that he is inventor or co-inventor.

The Full Federal Court had no option but to hear the appeal on the terms agreed by the parties.  The High Court, however, is under no such obligation.  As Gordon J stated, in denying Thaler’s application for special leave, ‘[t]he Court is of the opinion that this is not the appropriate vehicle to consider the questions of principle sought to be agitated by the applicant.’


Post a Comment

Copyright © 2014
Creative Commons License
The Patentology Blog by Dr Mark A Summerfield is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Australia License.