22 December 2020

DABUS Down Under – Can an AI ‘Communicate’ an Invention to a Human Applicant?

Xmas robotBack in September I reported that an application naming an ‘artificial intelligence’ going by the name DABUS (‘Device for the Autonomous Bootstrapping of Unified Sentience’) as inventor had been filed in Australia.  The application – no. 2019363177 – was derived from an international application filed under the Patent Cooperation Treaty (PCT), and names Dr Stephen Thaler, the developer of DABUS, as applicant.  The Australian Patent Office issued a formalities report, objecting that the application does not comply with the Patents Regulations 1991, because ‘it is the Commissioner’s understanding that inventors for the purposes of the Patents Act 1990 must be natural persons.  As DABUS is not a natural person the Commissioner is of the view that no inventor has been supplied for the purposes of Regulation 3.2C(2)(aa).’  There has since been further correspondence between Dr Thaler’s Australian patent attorneys and the Patent Office.  Thus far, however, the initial objections have not been overcome, and it seems likely that the matter is headed for a hearing, which will result in a formal decision being issued.

On 19 November 2020, the applicant responded to the initial formalities report.  For some reason, this response is not currently available in the public ‘eDossier’ for the application, however the Patent Office’s reply of 25 November 2020 sets out the substance of the response:

You have indicated that there is no explicit definition of an inventor, which by implication means that an artificial intelligence machine can be an inventor. You have also stated that it is sufficient for the purposes of the Patents Act 1990 if the applicant derives title to the invention from the inventor under section 15(1)(c). In the present case this would require DABUS to transfer title to Stephen Thaler. It is not apparent that DABUS is capable of executing an assignment, or has a contract of employment (for instance).

Interestingly, however, the Patent Office reply goes further, by suggesting an alternative theory (not, it seems, put forward in the applicant’s response) under which the Dr Thaler might assert the right to be granted a patent:

It may be that you are suggesting that Dr Thaler should be regarded as the communicatee of DABUS (a category of applicants recognised under section 34(1)(e) of the 1952 Act, and amalgamated into the present section 15(1)(c) - JMVB Enterprises Pty Ltd v Camoflag Pty Ltd [2006] FCAFC 141). According to traditional principles the communication should be for the purpose of obtaining a patent (H's Application (1955) 73 RPC 197).

The significance of identifying a mechanism for Dr Thaler to derive rights to DABUS’ inventions is that, in the Office’s view, if DABUS ‘is not capable of transferring or communicating an invention to a person, it would appear to be inconsistent with the structure of the Act for the machine to be treated as an inventor for the purposes of the Act.’  However, quite why the Office should wish to make its own suggestions in this regard is something of a mystery.

The applicant, for his part, appears to want no part of this novel theory proposed by the Patent Office.  In a further reply dated 9 December 2020, his attorneys have argued simply that:

The common law confers ownership of anything produced by DABUS to its owner, the applicant. There is a general rule that the owner of a thing is the owner of the fruits of that thing, much like the owner of a fruit tree is entitled to the fruit produced by that tree. The principle of accession or first possession can apply.

I think that this is the right approach (even though it is unlikely to succeed).  A UK court has already ruled that DABUS is incapable of transferring intellectual property rights to Dr Thaler (or anybody else), and there is no reason to expect a different outcome in Australia.  Furthermore, an analysis of the historical basis for the category of ‘communicatee’ suggests that it may no longer exist under Australia’s current patent laws and that, in any event, it would not provide for Dr Thaler to obtain a patent on DABUS’ inventions.  His best bet is therefore to continue to push the ‘ownership’ line, even though it has so far failed in the UK, the US, and at the European Patent Office.  I believe that it will fail in Australia, also, under the existing law.  But my understanding is that the larger objective of the Artificial Inventor Project, which is backing Dr Thaler’s applications, is to lobby for changes in the law, to recognise AI inventors, and entitlement by way of ownership.

In any event, the Patent Office’s unusual proposition provides an opportunity to delve into this strange historical category of ‘communicatee’.  What is it?  Why did it exist?  Does it still exist somewhere in the modern patent law?  (Readers interested only in Dr Thaler’s prospects of success, and not the historical context, can skip ahead to the section entitled ‘Can Title to an Invention be Derived from an AI?’)

Entitlement to Apply for a Patent – ‘Traditional’ Principles

The persons who were permitted to apply for a patent, under subsection 34(1) of the Patents Act 1952, relevantly (for present purposes) included:

  1. the actual inventor;
  2. the assignee of the actual inventor; and
  3. a person to whom the invention has been communicated by the actual inventor, his legal representative or assignee (if the actual inventor, his legal representative or assignee is not resident in Australia).

I will turn to this strange third category of ‘communicatee’ shortly, but it is first useful to say a few words about the seemingly straightforward category of ‘the actual inventor’.  Nowadays, we would understand this to be a person (or persons) who has themselves independently devised the invention.  Under the Australian Patents Act 1990 it is not possible for anyone else to validly claim to be an inventor.  For one thing, prior publication or use of an invention, by anyone, anywhere in the world, will deprive it of novelty for the purposes of Australian patent law.  Furthermore, there are various provisions in the Act to prevent a valid patent from being granted to, or held by, ‘ineligible persons’ – including persons who derived or ‘obtained’ the invention from somebody else.

In the longer tradition of English and Australian patent law, however, it was considered perfectly legitimate for a person to be considered an inventor who had imported an invention from a foreign country.  Historically, this was thought to be sound policy: in 1955, a UK Assistant Comptroller of Patents stated the position as follows (H's Application (1955) 73 RPC 197, at 199):

The reason for granting a monopoly to a person who brings in a new manufacture, which he has not himself devised, from abroad is well understood. The historical background of this aspect of the law is well-known. It is easy to grant that a man who undertook serious financial and even personal risks by foreign travel in order to benefit the trade and industry of this country was deserving of reward.

Clearly, this cannot be a legitimate basis for awarding a patent to an ‘inventor by importation’ in the modern globalised and connected world.  Indeed, even in the 1950’s it was not enough merely to be the first to apply for a patent on an invention that had been developed elsewhere. H’s Application was itself a case in which the alleged inventor had acquired knowledge of the invention from the US manufacturer by way of samples sent via mail in response to an letter of inquiry about becoming a licensed manufacturer in the UK.  This relatively passive acquisition of information fell short of what was required to claim the status of an inventor by importation, with Lloyd-Jacob J finding (at 204):

In my judgment, a letter of request for information such as would disclose the alleged invention cannot of itself be said to establish possession of the alleged invention, and in so far as this Applicant ever was possessed of information with regard to this invention, he derived it, and derived it solely, from the information contained in Mr. Kuhn's reply which was dated 1st June, 1951.

As a result, the applicant was found to have ‘obtained’ the invention, i.e. to have illegitimately derived it, directly or indirectly, from the actual inventor, and thus not to have been entitled to apply for a patent.

An assignee of the inventor (whether an independent devisor, or a legitimate importer, of an invention) has also traditionally been entitled to apply for, and be granted, a patent, and this of course remains true today.  An assignment is a formal legal transfer of rights in the invention and/or any existing patent/application.

How is an Invention ‘Communicated’?

So what is this third category, of ‘a person to whom the invention has been communicated by the actual inventor’, and how does it differ from an assignee?  This provision of the 1952 Act has a couple of key characteristics, one explicit, and the other implicit.  The explicit requirement is that the communicator could not have been resident in Australia, i.e. the invention must have been communicated from overseas.  Additionally, it is implicit that the purpose of the communication should be for the making of a patent application, i.e. merely being informed about an invention for some other reason (such as a proposed commercial dealing) would not entitle someone to apply for a patent as a communicatee.  As the Assistant Comptroller observed in H’s Application, at 202-3:

An application may properly be made by a communicatee as the result of a communication made for that purpose, and if the patentee abuses the trust, actual or implied, on which the communication was made, no doubt redress must be sought outside the patent system. But it does not seem to me to follow necessarily from that, that a man who merely stays in London, invites a foreigner to disclose an invention to him for ordinary commercial purposes, and then waits in London until the foreigner, either in person or by post, brings it to him in his office there, has a right to apply as an inventor for a patent. His position is not the same as that of a communicatee who merely acts on behalf of his principal abroad.

This implies that a communicatee is (or was) a form of ‘agent’, although more in the nature of a trustee than an attorney.  The distinction appears to be that a communicatee is a form of local ‘representative’ of the inventor, or their assignee, who is empowered to act outside of the rules and conventions of a client-attorney relationship.  A communicatee nonetheless holds a position of trust, presumably requiring them to act in the best interests of, and having a fiduciary duty to, the ‘true’ inventor or assignee.  As such, a communicatee does not have the same status as an actual inventor or assignee.  Indeed, it is apparent from the decision in H’s Application that the UK application forms included a separate section to be completed when a patent application was made by a communicatee, ensuring that the nature and role of the applicant is such cases was clear, and that the actual inventor was also identified.

Historically, allowing for a communicatee to apply for a patent may have served a useful purpose, considering that correspondence with a foreign inventor or applicant may have been difficult, or at least involved long delays for delivery via surface mail.  Nowadays, however, such a role appears superfluous.  There is no longer any practical barrier to an applicant, wherever they may be located in the world, receiving timely information and advice, or instructing a local attorney via electronic communications.  In Australia, a foreign applicant can even apply directly for a patent online, so long as they can satisfy the formality of providing a local postal address.

A further reason why there may no longer be any role for a communicatee in Australian patent law is that the scheme for making applications under the 1990 Act differs materially from the 1952 Act.  Section 34 of the 1952 Act restricted those persons who could validly be granted a patent by restricting those who could apply in the first place.  A person who did not fall into one of the categories enumerated in subsection 34(1) simply could not apply, so the category of communicatee made it possible for a foreign inventor or assignee to appoint a local representative to apply on their behalf.  In contrast, the 1990 Act places no restrictions on who may apply.  Section 15 is instead concerned with prescribing the persons to whom a patent may ultimately be granted.  Anyone may file an application, but they must identify a ‘nominated person’ who is entitled under section 15 to have the patent granted to them.  By default, the applicant and the nominated person are assumed to be the same, but this need not be the case.  There is no impediment to any inventor or assignee, foreign or otherwise, appointing any representative they desire to handle filing of an Australian patent application on their behalf.

Does ‘Communication’ Still Exist in Australian Patent Law?

The provisions of the 1952 Act were designed for a world in which the ‘tyranny of distance’ was a real factor, importers of technologies new to the nation (but not to the world) could be treated as ‘actual inventors’, and global communications could be slow and unreliable.  Those days are gone – technologically, economically, and culturally.  One would think, therefore, that the role of communicatee should also be a thing of the past.

And yet, in JMVB Enterprises Pty Ltd v Camoflag Pty Ltd [2006] FCAFC 141 at [70], a unanimous Full Bench of the Federal Court of Australia observed that:

… the distinction between the assignee of the actual inventor and a person to whom the invention has been communicated by the actual inventor, which was made in s 34(1) of the 1952 Act, is not to be found in s 15(1). Those concepts have been amalgamated into the provision in s 15(1)(c), whereby a patent may be granted to a person who derives title to the invention from the inventor or a person who would, on the grant of a patent for the invention, be entitled to have the patent assigned to that person. (Emphasis in original.)

By itself, this passage might be understood to imply that the Full Court considered the role of communicatee, as it existed under the 1952 Act, to have survived as a sub-category of persons who ‘derive title’ to the invention.  If that is so, then I would respectfully suggest that the court was wrong.  As discussed above, a communicatee appears to have been something in the nature of a trustee, rather than a full beneficial owner of rights or title in the invention and/or any patent that might be granted.

In any event, the above observation of the Full Court is obiter dicta.  The outcome in JMBV did not turn on whether or not a communicatee (within the meaning of the 1952 Act) might fall within the terms of s 15(1)(c) of the 1990 Act, even though the court declined to rule out this possibility.  The essential requirement is rather that the nominated person, or patentee, must ‘derive title’ to the invention (at [72]):

… the language of s 15(1) is limited to a person who derives title to the invention from the inventor. Mere communication or importation, without anything further, is not sufficient to give rise to a title on the part of the communicatee or importer.  (Emphasis in original.)

Can Title to an Invention be Derived from an AI?

There are a number of well-established ways in which a person (including a corporation) may derive title to an invention from the actual inventor.  All involve the operation of law.  One is by way of express contract, i.e. an assignment of title in the invention along with all associated rights, including the right to apply for patents.  Another is by way of a employment relationship, either under common law or (preferably) through the express terms of an employment contract clearly specifying the employee’s duties and employer’s ownership of associated intellectual property.  As another example of transfer of title, if an inventor dies then ownership of their intellectual property passes to the estate, and ultimately to any heirs.

There is, however, no established legal basis for title in any invention made by an AI to be transferred to a person or corporation.  An AI has no relevant legal status.  It cannot own property in the first place, much less transfer legal title in any such property to another entity. 

With regard to the DABUS applications in particular, in my view the position in Australia is substantially the same as in the UK.  As the England and Wales High Court found in Thaler v The Comptroller-General of Patents, Designs And Trade Marks [2020] EWHC 2412 (Pat):

40. It is quite clear from the statutory scheme contained in the Patents Act 1977 that – whatever the meaning of the term "inventor" – a patent can only be granted to a person. I reach this conclusion explicitly without considering the meaning of the term inventor. …

46. As I have noted, I am proceeding on the basis that DABUS has "invented" the inventions the subject of the Applications. Nevertheless, I conclude that DABUS is not, and cannot be, an inventor within the meaning of the 1977 Act, simply because DABUS is not a person.

49. … Even if I accepted that DABUS was capable of being an "inventor" – which, for the reasons I have given, I do not – Dr Thaler's application would be hopeless, because DABUS would – by reason of its status as a thing and not a person – be incapable of conveying any property to Dr Thaler. In short, the ability to transfer, which DABUS lacks, is fatal to Dr Thaler's contentions. The same point can be put in a different way: because DABUS is a thing, it cannot even hold property, let alone transfer it. …

I do not believe that Dr Thaler’s position is assisted by the historical ‘communicatee’ category of applicant, even assuming (which I strongly doubt) that this is still relevant under the 1990 Act, at least because:

  1. a communicatee cannot take the place of the actual inventor who, under the scheme of the Patents Act 1990, must still exist and be a person;
  2. under the 1952 Act, communication of the invention was required to take place from a non-resident inventor or assignee to a resident of Australia, and for the specific purpose of applying for an Australian patent, none of which applies in the DABUS case because Dr Thaler is not an Australian resident, and DABUS is incapable of having purpose in communicating its inventions; and
  3. in any event, the key requirement of s 15(1)(c) – i.e. that the nominated person or patentee must derive title to the invention from the actual inventor – cannot be satisfied, since DABUS is incapable of holding property, or of legally transferring title to Dr Thaler.

Conclusion – AI Inventors Face Hurdles to Recognition in Australia

In my view, the suggestion by the Patent Office that DABUS may be capable of communicating its inventions to Dr Thaler, within the meaning of s 34(1)(e) of the 1952 Act, is fundamentally misconceived.  Furthermore, the decision of the England and Wales High Court makes clear that the established legal mechanisms by which an applicant derives title to an invention do not apply where the inventor is a ‘thing’, such as an AI.

Dr Thaler has rightly recognised that his prospects of success are not assisted by any of the principles upon which the right to apply for a patent have conventionally or traditionally been derived from an actual inventor.  Instead, he is presently relying upon the proposition that the common law recognises ‘a general rule that the owner of a thing is the owner of the fruits of that thing’.

I expect that the case will be referred to a hearing officer to determine the matter by way of a formal decision, in which all of the issues will receive due consideration.  However, Dr Thaler’s rather unconventional assertions that an AI can be an inventor, and that he is entitled to receive any resulting patents by virtue of ownership of the AI, are, in my view, unlikely to be accepted by the Patent Office.  Such substantive developments in the law would, at the very least, require judicial consideration or – more likely – legislative intervention.


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.