Patents

An AI is not an inventor after all (or yet)

A strong Full Bench of the Federal Court of Australia has ruled that DABUS, an artificial intelligence, is not an inventor for the purposes of patent law. So, Dr Thaler’s application for DABUS’ patent has been rejected.[1] No doubt the robot will be back again[2] and we can expect that an application for special leave will be pending soon.

A dalek on display
By Moritz B. – Self-photographed, CC BY 2.5,

Dr Thaler had applied for a patent, No. 2019363177 entitled “Food container and devices and methods for attracting enhanced attention”, naming DABUS – an acronym for ‘device for the autonomous bootstrapping of unified sentience’ – as the inventor.

The Commissioner had rejected the application under reg. 3.2C for failure to identify the inventor. That rejection was overturned by Beach J on appeal from the Commissioner. And this was the decision on the Commissioner’s appeal.

Essentially, the Full Court ruled that an inventor for the purposes of patent law must be a natural person, not an artificial intelligence.

The Full Court held that identification of the “inventor” was central to the scheme of the Act. This is because, under s 15, only the inventor or someone claiming through the inventor is entitled to a patent.

Under the legislation before the 1990 Act, their Honours considered that an ‘actual inventor’ could be only a person with legal personality. At [98], their Honours summarised:

In each of these provisions, the ability of a person to make an application for a patent was predicated upon the existence of an “actual inventor” from whom the entitlement to the patent was directly or indirectly derived. Paragraphs (a), (c) and (e) describe the actual inventor as, respectively, a person, one that is deceased and has a legal representative (which must be a person), and one that is not resident in Australia. Paragraphs (b), (d), (f) and (fa) all contemplate an assignment happening between the patent applicant and the actual inventor. It is clear from these provisions that only a person with a legal personality could be the “actual inventor” under this legislative scheme.

This scheme, and its consequences, did not materially change under the 1990 Act.

Acknowledging that a none of the case law had to consider whether an AI could be an inventor, the Full Court noted that the ‘entitlement’ cases proceeded on the basis that ‘inventor’ meant the ‘actual inventor’. Their Honours considered the cases interpreting this expression were all premised on the ‘actual inventor’ – the person whose mind devised the claimed invention – being a natural person. At [105] and [106], their Honours explained:

None of the cases cited in the preceding five paragraphs confronted the question that arose before the primary judge of whether or not the “inventor” could include an artificial intelligence machine. We do not take the references in those cases to “person” to mean, definitively, that an inventor under the Patents Act and Regulations must be a human. However, it is plain from these cases that the law relating to the entitlement of a person to the grant of a patent is premised upon an invention for the purposes of the Patents Act arising from the mind of a natural person or persons. Those who contribute to, or supply, the inventive concept are entitled to the grant. The grant of a patent for an invention rewards their ingenuity.

Where s 15(1)(a) provides that a patent for an invention may only be granted to “a person who is an inventor”, the reference to “a person” emphasises, in context, that this is a natural person. …. (emphasis supplied)

Given that conclusion, and the structure of s 15, Dr Thaler’s argument that he was entitled on the basis of ownership of the output of DABUS’ efforts was to no avail. At [113]:

… having regard to the view that we have taken to the construction of s 15(1) and reg 3.2C(2)(aa) [i]t is not to the point that Dr Thaler may have rights to the output of DABUS. Only a natural person can be an inventor for the purposes of the Patents Act and Regulations. Such an inventor must be identified for any person to be entitled to a grant of a patent under ss 15(1)(b)-(d). (emphasis supplied)

The Full Court then drew support from the High Court’s reasoning in D’Arcy v Myriad esp. at [6] in which the majority emphasised that patentable subject matter had to be the product of “human action”.

Although not put in this way, it is apparent that policy considerations played a significant role in their Honours’ conclusion. At [119] to [120], their Honours pointed out:

in filing the application, Dr Thaler no doubt intended to provoke debate as to the role that artificial intelligence may take within the scheme of the Patents Act and Regulations. Such debate is important and worthwhile. However, in the present case it clouded consideration of the prosaic question before the primary judge, which concerned the proper construction of s 15 and reg 3.2C(2)(aa). In our view, there are many propositions that arise for consideration in the context of artificial intelligence and inventions. They include whether, as a matter of policy, a person who is an inventor should be redefined to include an artificial intelligence. If so, to whom should a patent be granted in respect of its output? The options include one or more of: the owner of the machine upon which the artificial intelligence software runs, the developer of the artificial intelligence software, the owner of the copyright in its source code, the person who inputs the data used by the artificial intelligence to develop its output, and no doubt others. If an artificial intelligence is capable of being recognised as an inventor, should the standard of inventive step be recalibrated such that it is no longer judged by reference to the knowledge and thought processes of the hypothetical uninventive skilled worker in the field? If so, how? What continuing role might the ground of revocation for false suggestion or misrepresentation have, in circumstances where the inventor is a machine?

Those questions and many more require consideration. Having regard to the agreed facts in the present case, it would appear that this should be attended to with some urgency. However, the Court must be cautious about approaching the task of statutory construction by reference to what it might regard as desirable policy, imputing that policy to the legislation, and then characterising that as the purpose of the legislation …. (emphasis supplied)

Finally, in this quick reaction, it can be noted that the Full Court recognised that their Honours’ decision was consistent with the English Court of Appeal’s decision on the counterpart application. Their Honours considered, however, there were sufficient differences in the legislative schemes that a wholly autocthonous solution should be essayed.

Commissioner of Patents v Thaler [2022] FCAFC 62 (Allsop CJ, Nicholas, Yates, Moshinsky And Burley JJ)


  1. Patent application No. 2019363177 entitled “Food container and devices and methods for attracting enhanced attention”  ?
  2. With apologies to you know who.  ?

Aristocrat gets special leave

The High Court has granted Aristocrat special leave to appeal the Full Federal Court’s ruling that Aristocrat’s application for an electronic gaming machine (EGM or “pokie”) was not patentable subject matter.[1]

The patent application

Aristocrat’s application is entitled ‘A system and method for providing a feature game’ – App. No. 2016101967; yet another problematic “innovation” patent.

The Commissioner and Aristocrat were in agreement that the case rose or fell on the patentability of claim 1:

(1) A gaming machine comprising:

(1.1) a display;

(1.2) a credit input mechanism operable to establish credits on the gaming machine, the credit input mechanism including at least one of a coin input chute, a bill collector, a card reader and a ticket reader;

(1.3) meters configured for monitoring credits established via the credit input mechanism and changes to the established credits due to play of the gaming machine, the meters including a credit meter to which credit input via the credit input mechanism is added and a win meter;

(1.4) a random number generator;

(1.5) a game play mechanism including a plurality of buttons configured for operation by a player to input a wager from the established credits and to initiate a play of a game; and

(1.6) a game controller comprising a processor and memory storing (i) game program code, and (ii) symbol data defining reels, and wherein the game controller is operable to assign prize values to configurable symbols as required during play of the game,

(1.7) the game controller executing the game program code stored in the memory and responsive to initiation of the play of the game with the game play mechanism to:

(1.8) select a plurality of symbols from a first set of reels defined by the symbol data using the random number generator;

(1.9) control the display to display the selected symbols in a plurality of columns of display positions during play of a base game;

(1.10) monitor play of the base game and trigger a feature game comprising free games in response to a trigger event occurring in play of the base game,

(1.11) conduct the free games on the display by, for each free game, (a) retaining configurable symbols on the display, (b) replacing non-configurable symbols by selecting, using the random number generator, symbols from a second set of reels defined by the symbol data for symbol positions not occupied by configurable symbols, and (c) controlling the display to display the symbols selected from the second set of reels, each of the second reels comprising a plurality of non-configurable symbols and a plurality of configurable symbols, and

(1.12) when the free games end, make an award of credits to the win meter or the credit meter based on a total of prize values assigned to collected configurable symbols.

(emphasis supplied by Middleton and Perram JJ).

The new or innovative feature lay in the feature comprising the free game integer – integers 1.10 to 1.12. It was apparently common ground that the other features were part of the common general knowledge for electronic gaming machines.

In the Full Court, Middleton and Perram JJ at [3] and [4] explained:

3 It is not suggested that there is anything inventive about Claim 1’s EGM except for its feature game and it is in all other respects an unremarkable EGM. (Because the 967 Patent is an innovation patent strictly the question is whether there is anything innovative about it, but nothing turns on the distinction between inventive and innovative for present purposes). A feature game is a secondary game awarded to a player on the occurrence of a defined event in the ordinary or ‘base’ game of spinning reels, termed a ‘trigger event’. Once the feature game is enlivened by the trigger event the feature game appears and the player is able to play it and potentially to win further prizes. When the feature game is completed the EGM reverts to the base game.

4 The point of feature games is to encourage players to keep wagering on the EGM by making it more interesting to do so. Since the revenue generated by an EGM is a function of the amount wagered upon it – in New South Wales up to 15% in the long run – the more successful a feature game is in keeping the player wagering, the more lucrative the EGM is for its operator. A successful feature game is therefore commercially valuable both from the perspective of the class of persons who operate EGMs and from the perspective of those who manufacture and distribute them to that class. The Respondent (‘Aristocrat’) is part of a world-wide group of companies engaged in the manufacture and distribution of EGMs and is the particular member of the group which owns the 967 Patent.

Their Honours then explained how the free game feature worked at [11] – [12]:

11 The game defined by integers 1.10–1.12 is in fact not a single game at all but rather a family of games with particular common attributes. On the occurrence of a trigger event (integer 1.10) the player is awarded one or more free games of the feature game (integer 1.11). The feature game (integer 1.11) consists of a second set of reels. Amongst the symbols on these reels are ‘configurable’ symbols. The patent does not define a configurable symbol but it does provide for them to be assigned prize values by the computer on which the game is played, which is known as the game controller (integer 1.6). A preferred embodiment of the invention suggests that the configurable symbol may be overlaid with the amount of the prize which has been assigned to it (although this is not a necessary feature of integer 1.11 and any symbol will do). In that preferred embodiment, the configurable symbol is an image of a pearl and it is configured by the overlaying on that image of different prize amounts, e.g., some pearls appear with ‘250’ and others with ‘1000’, where those figures represent credits.

12 Returning to the feature game, each time a configurable symbol appears in the display grid at the end of the free game that particular symbol position on the relevant reel stops spinning for any remaining free games and the configurable symbol remains locked in place in any subsequent play of the feature game (i.e. if the player still has any free games left). When the player eventually runs out of free games in the feature game a prize is awarded related to the number of configurable symbols which have been locked in place (integer 1.12). In the preferred embodiment the prize is the sum of the assigned values on the pearls which have been frozen on the display grid but integer 1.12 is consistent with the prize being calculated in some other way.

Burley J

At first instance (on appeal from the Commissioner), Burley J considered that the case law required a two stage assessment of patentable subject matter:

  1. The initial inquiry was whether or not the claim was for a mere scheme or business method of the type that was not the proper subject matter for a grant of a patent.
  2. If so, a second inquiry arose: whether or not the claim involved the creation of an artificial state of affairs where the computer was integral to the invention, rather than a mere tool in which it was performed. That is, was there invention in the computerisation of the claimed method?

Each step was to be undertaken as a matter of substance rather than mere form.

Applying that methodology, Burley J avoided the whole mere scheme or business method controversy by holding that the claim was for a mechanism of a particular construction where the integers interacted to produce a particular product – an EGM.

While this blogger welcomed the result, it did raise a rather awkward question: why did putting the integers in a box create patentable subject matter when essentially the same functionality could also be supplied over a network including, dare one say it, the Internet.

The Commissioner appealed.

The Full Court

All three judges (Middleton, Perram and Nicholas JJ) unanimously allowed the appeal. Middleton and Perram JJ delivered the main opinion and Nicholas J delivered a separate concurring opinion.

All three judges accepted unreservedly that a mere scheme or abstract idea was not patentable subject matter.

All three judges also accepted that whether something is patentable subject matter was to be determined as a matter of substance rather than mere form.

Middleton and Perram JJ

At [14] – [15], Middleton and Perram JJ considered that the feature game itself defined by integers 1.10 to 1.12 was a mere abstract idea for purposes of patent law. This was because either it was the definition of a family of games with common attributes and so akin to the rules of a game. Or it was because it was a method of increasing player interest in the EGM and so increasing the operator’s gaming revenue. On that view, it was just a business scheme or scheme.

At [16], their Honours accepted that an invention which physically embodied an abstract idea and gave it some practical application could be patentable subject matter. Thus, a mechanical poker machine which allowed a game (the abstract idea) to be played could be patentable. In such a case, however, the patent would protect the physical embodiment and not the abstract idea:

But the patent protects the invention which is the poker machine and not the abstract idea consisting of the game which it plays. This is consistent with decisions on board or card games to the effect that the game itself, no matter its ingenuity, does not comprise patentable subject matter but the physical apparatus used for playing the game (such as cards or the board) may do so ….

Crucially, their Honours then held that the implementation of the game by means of an unspecified computer program could not be a manner of manufacture unless the implementation resulted in some development of computer technology rather than its utilisation. At [18], their Honours said:

The implementation of an abstraction such as that disclosed by integers 1.10–1.12 by means of an unspecified computer program to be executed on the computer which is the game controller will not give rise to a patentable invention unless the implementation itself can be seen as pertaining to the development of computer technology rather than to its utilisation …. [2]

Middleton and Perram JJ next accepted that Burley J’s approach could be workable where the subject matter of the claimed patent was obviously implemented in a computer. In other cases, however, it had the potential to lead to a wrong result especially where “for example, whether a claimed physical apparatus such as an EGM is, in truth, no more than a particular kind of computer.”

Instead, Middleton and Perram JJ proposed at [26] a different two-step analysis:

(a) Is the invention claimed a computer-implemented invention?

(b) If so, can the invention claimed broadly be described as an advance in computer technology?

In the present case, the claimed invention was a computer-implemented invention.

First, although the apparatus claimed was an EGM (or poker machine), it was in substance a computer. At [32] – [34], their Honours found that the EGM was a game control computer (integer 1.6) attached to a random number generator (integer 1.4, another computer) with some input and output devices (respectively, integers 1.2, 1.3 and 1.5 (input) and 1.1 (output)) with associated software instructions (integers 1.7, 1.8 and 1.9 and the feature game 1.10 – 1.12).

Secondly, to avoid the (erroneous) conclusion that the claimed invention was for the practical implementation of an abstract idea in a device, at [50] their Honours considered it was necessary to distinguish between a claim for an invention which was a computer and a claim for an invention implemented in a computer.

As a claim for either type of invention could appear to be a claim for a computer, it was necessary to identify what was in fact the substance of the claimed invention.

Here, the correct characterisation of the claim had to take into account the two elements of the claim: one element being the EGM (in effect, the computer) and the other element being the feature game and, importantly, the feature game was the only innovative feature of the claim.

As the feature game was to be executed in the computer, at [56] Middleton and Perram JJ considered the relationship of the feature game elements to the computer elements was one of implementation. Accordingly, the claim was to a computer-implemented invention.

Turning to the second question (b), Middleton and Perram JJ held that the claimed invention was not directed to an advance in computer technology. At [63], their Honours said:

Because the invention is the implementation of a feature game on the computer which is an EGM, the next question is whether what is put forward as inventive (or innovative) about Claim 1 pertains to the development of computer technology or merely its use. The fact that integers 1.10–1.12 leave it entirely up to the person designing the EGM to do the programming which gives effect to the family of games which those integers define inevitably necessitates the conclusion that Claim 1 pertains only to the use of a computer. Indeed, it purports to do nothing else. Claim 1 is silent on the topic of computer technology beyond that the person implementing the invention should use some.

It did not matter that the claim improved player engagement or improved subjective satisfaction. That had nothing to do with developing or advancing computer technology. Similarly, the use of configurable symbols for prizes was of no assistance. That might advance gaming technology but was not an advance in computer technology.

Nicholas J

Nicholas J reached the same conclusion but by a somewhat different route.

Like Middleton and Perram JJ, his Honour started from the proposition that mere schemes and abstract ideas are not patentable. At [115] Nicholas J considered that to be patentable the case law required a claimed invention to relate to some technological innovation:[3]

The Full Court noted at [100] the distinction drawn in Research Affiliates at [94] between a technological innovation which is patentable and a business innovation which is not. The use of the expression “technological innovation” emphasises the need to identify a technological contribution in a field of technology. The desirability of providing patent protection to technological innovations is reflected in Art 27(1) of the Agreement on Trade Related Aspects of Intellectual Property Rights which relevantly provides that “… patents shall be available for any inventions, whether products or processes, in all fields of technology, provided that they are new, involve an inventive step and are capable of industrial application …”. The language of “technological innovation” has now been adopted in s 2A of the Act which refers to the promotion of economic wellbeing through technological innovation as an object of the Act.

An interesting invocation of the new Objects clause in the Act.

At [116], Nicholas J considered it was important to avoid an excessively rigid or formulaic approach to this issue. This was because the technological advance may lie in the field of computer technology. However, the required technological advance could also lie in a field of technology outside the computer:

… it is not appropriate to adopt an excessively rigid or formulaic approach to the question whether a computer implemented scheme is a manner of manufacture. This is especially true in situations where there may be no clear distinction between the field to which the invention belongs, and the field of computer technology. There may well be a technological innovation in the field of technology to which the invention belongs even though it cannot be said that there has been some technological innovation in the field of computers. The field of the invention may encompass different fields of technology that have their own technical problems that lie “outside the computer”. Moreover, the solutions to these problems may necessarily rely upon generic computing technology for their implementation. That does not necessarily render such solutions unpatentable.

His Honour gave as an example at [120] a computer-implemented invention for running a refrigerator in a more energy-efficient way. This might not involve an advance in computer technology but could well be patentable where the invention lay in the field of refrigeration technology. So, there could also be patentable subject matter in the way a gaming system or machine functioned even if there had not been an advance in computer technology.

Turning to Aristocrat’s patent application, Nicholas J considered at [135] that Burley J’s two-step test did not adequately address the Commissioner’s submission that the claim was for nothing more than the (unpatentable) rules of a game implemented in generic computer technology for its well-known and well-understood functions.

Citing RPL at [96], the fact that the feature game instructions were embodied in a computer was not sufficient to qualify as patentable subject matter. While the purpose of the invention was to provide a different and more enjoyable playing experience, the claim was not directed to overcoming any technological problem. At [141], his Honour explained:

The specification does not identify any technological problem to which the patent purports to provide a solution. Nor did the expert evidence (insofar as it was made available to us) suggest that the invention described and claimed in the specification was directed to any technical problem in the field of gaming machines or gaming systems. Rather, as the specification makes clear, the purpose of the invention is to create a new game that includes a feature game giving players the opportunity to win prizes that could not be won in the base game. Ultimately, the purpose of the invention is to provide players with a different and more enjoyable playing experience. The invention is not directed to a technological problem residing either inside or outside the computer.

Generic computer technology / software

While acknowledging that the phraseology had been used in earlier Full Court decisions, Middleton and Perram JJ considered at [35] to [42] that testing whether the claimed invention merely involved a generic computer or generic software was not “especially helpful”. It was preferable to focus on whether the claimed invention related to an advance in computer technology.

In contrast, Nicholas J (who had participated in the earlier cases) was not so troubled. At [112], his Honour considered it could be a useful signpost to patentability (or not) to ascertain whether the computer or software was just conventional computer technology being used for its well-known and well-understood effects.

Where to now

The law relating to “manner of manufacture” is a mess.

Following NRDC and before about the mid–2000s, the issue hardly, if ever, came up. Since the mid–2000s, there have been numerous cases; it is perhaps no exaggeration to say there are more, indeed way more, cases each year than in the previous 50 years.

And, unless one is prepared to say that it’s a computer-related invention and so it is not patentable subject matter (which the Courts repeatedly do not say), it is very hard to say what will pass the threshold and what will not.

One issue is that the cases are replete with comments like there is nothing new or inventive about that. That suggests that what is really the issue is lack of novelty or inventive step.

That though gives rise to a whole set of sub-issues. First, the whole situation is exacerbated by the abominations called “innovation patents”, which don’t actually require any invention, just that what is claimed (in effect) be new.

Secondly, proving lack of inventive step is a complicated, expensive and risky gamble – especially under our law before the Raising the Bar amendments.

Thirdly, while cases like CCOM, IBM and Welcome-Catuity could provide a principled approach to this issue, that could well end up with Australian law granting patents in circumstances where the USA and the EU would not. Not a situation the Productivity Commission would favour. And, it is far from clear that any clear or consistent approach has emerged in either jurisdiction, especially the USA.

The reason, or at least one of the reasons, why NRDC was a “watershed” in patent law was that it got rid of the artificial pigeon-holes or categories of “vendible product” which, as the NRDC judgment so tellingly demonstrated, had resulted in such inconsistent and unpredictable results to the test of manner of manufacture. Instead, it adopted an open-ended, flexible approach directed to achieving the objects of patent law.

Against that background, one might argue that Nicholas J’s approach, with respect, allows a degree of flexibility and forward-thinking which may not follow from the approach taken by Middleton and Perram JJ. Whether his Honour’s approach provides any more certainty may be debated but at least it would avoid a narrow categorisation. Whether that is an approach that finds favour with the High Court, or what direction it might take, remains to be seen.

[Some typos were corrected on 24 June 2022]


  1. That is, not a ‘manner of manufacture’ for the purposes of Patents Act 1990 s 18(1)(a) and 18(1A)(a). Exceptionally, special leave was granted on the papers.  ?
  2. Citing Commissioner of Patents v RPL Central Pty Ltd [2015] FCAFC 177; 238 FCR 27 at [96] and [102] (Kenny, Bennett and Nicholas JJ).  ?
  3. Nicholas J considered that Research Affiliates at [115] – [119] (and RPL Central applying it) broadly equated the requirement for a technical contribution or technological innovation to the “artificially created state of affairs” required under the NRDC test.  ?

The English Court of Appeal rejects AI too

The Court of Appeal has dismissed by a 2:1 majority Dr Thaler’s appeal from Marcus Smith J’s decision rejecting the patent applications on the grounds that DABUS is not an inventor.

A recap

In late 2018, Dr Thaler applied for the grant of two patents, GB18116909.4 entitled “Food Container” concerning the shape of parts of packaging for food and and GB1818161.0 entitled “Devices and Methods for Attracting Enhanced Attention” for a form of flashing light.

In his application forms, Dr Thaler named DABUS as the inventor and stated that his entitlement arose “by ownership of the creativity machine ‘DABUS’”.

In later, amended forms, Dr Thaler explained “the applicant identified no person or persons whom he believes to be an inventor as the invention was entirely and solely conceived by DABUS”.

Patents Act 1977 (UK) ss 7 and 13

Section 7 of the UK’s Patents Act 1977 defines who is entitled to apply for and obtain a patent:

7 Right to apply for and obtain a patent.

(1)Any person may make an application for a patent either alone or jointly with another.

(2)A patent for an invention may be granted—

(a)primarily to the inventor or joint inventors;

(b)in preference to the foregoing, to any person or persons who, by virtue of any enactment or rule of law, or any foreign law or treaty or international convention, or by virtue of an enforceable term of any agreement entered into with the inventor before the making of the invention, was or were at the time of the making of the invention entitled to the whole of the property in it (other than equitable interests) in the United Kingdom;

(c)in any event, to the successor or successors in title of any person or persons mentioned in paragraph (a) or (b) above or any person so mentioned and the successor or successors in title of another person so mentioned;

and to no other person.

(3)In this Act “inventor” in relation to an invention means the actual deviser of the invention and “joint inventor” shall be construed accordingly.

(4)Except so far as the contrary is established, a person who makes an application for a patent shall be taken to be the person who is entitled under subsection (2) above to be granted a patent and two or more persons who make such an application jointly shall be taken to be the persons so entitled.

Although the wording is somewhat different, ss 13(1) and (2) of the UK Act appear to be similar in effect as, respectively, s 29 and s 15 of the Australian Act.[1]

Section 13 of the UK Act further provides:

13 Mention of inventor.

(1)The inventor or joint inventors of an invention shall have a right to be mentioned as such in any patent granted for the invention and shall also have a right to be so mentioned if possible in any published application for a patent for the invention and, if not so mentioned, a right to be so mentioned in accordance with rules in a prescribed document.

(2)Unless he has already given the Patent Office the information hereinafter mentioned, an applicant for a patent shall within the prescribed period file with the Patent Office a statement—

(a)identifying the person or persons whom he believes to be the inventor or inventors; and

(b)where the applicant is not the sole inventor or the applicants are not the joint inventors, indicating the derivation of his or their right to be granted the patent;

and, if he fails to do so, the application shall be taken to be withdrawn.

(3)Where a person has been mentioned as sole or joint inventor in pursuance of this section, any other person who alleges that the former ought not to have been so mentioned may at any time apply to the comptroller for a certificate to that effect, and the comptroller may issue such a certificate; and if he does so, he shall accordingly rectify any undistributed copies of the patent and of any documents prescribed for the purposes of subsection (1) above.[2]

Can an AI be an inventor?

All three judges were in agreement that DABUS, as an AI, cannot be an inventor because an inventor must be a person.

At [116] Arnold LJ, who gave the lead judgment for the majority, said:

In my judgment it is clear that, upon a systematic interpretation of the 1977 Act, only a person can be an “inventor”. The starting point is section 130(1) which provides that “‘inventor’ has the meaning assigned to it by section 7 above”. Section 7(3) provides that “‘inventor’ in relation to an invention means the actual deviser of the invention”. A dictionary definition of “deviser” is “a person who devises; a contriver, a planner, an inventor” (Shorter Oxford English Dictionary, 5th edition, Oxford University Press, 2002). Section 7(2) provides that a patent may be granted (a) “primarily to the inventor or joint inventors”, (b) “to any person or persons who …”, (c) “the successor or successors in title of any person or persons mentioned in paragraph (a) or (b) above”, but “to no other person”. As Lord Hoffmann explained in Yeda Research and Development Company Ltd v. Rhone-Poulenc Rorer International Holdings [2007] UKHL 43, [2007] Bus LR 1796 at [20], this is “an exhaustive code”. It is clear from this code that category (a) must consist of a person or persons, just as much as categories (b) and (c) do. Section 7(4) creates a presumption that “a person who makes an application for a patent shall be taken to be the person who is entitled under subsection (2) above to be granted a patent”. Again, it is plain that only a person can be entitled under section 7(2), and thus only a person can fall within paragraph (a).

Similarly, at [50] – [51], Birss LJ said:

50 Second, a purpose of the definition of inventor in section 7(3) was to change the law, from the old law going back to the Statute of Monopolies, and to abolish the idea of a true and first inventor who could be anyone other than the actual deviser of the invention. The concept of the actual deviser of the invention was already known to United Kingdom patent law from the 1949 Act (s16). It was the person who actually devised the invention. The contrast was between that person and others who had not done so but were regarded as the true and first inventor, e.g. importers. So in Yeda paragraph [18] the contrast was drawn between the actual deviser and a pretended or deemed deviser.

51 The rest of the 1977 Act is drafted on the footing that the inventor is a person. For example s7(2)(c) of the 1977 Act refers to “person or persons mentioned in paragraph (a) or (b)” and s13 of the Act require an applicant to identify the “person or persons whom he believes to be the inventor or inventors”. However, for what it is worth, I would hold that the mechanism by which the inventor is a person in the scheme of the 1977 Act is because the “actual deviser” is a person and so, by definition in s7(3), is the inventor.

and at [54] – [55]:

54 I conclude that the answer to the first question is a simple one. Within the meaning of the 1977 Act the “inventor” is the person who actually devised the invention.

55 That conclusion is arrived at without any need to examine the policy arguments raised by both parties. Machines are not persons. The fact that machines can now create inventions, which is what Dr Thaler says happened in this case, would not mean that machines are inventors within the meaning of the Act. Assuming the machine is the entity which actually created these inventions, it has no right to be mentioned as the inventor and no right to employee’s compensation under s39 (which no doubt it never had anyway).

At [102], Lady Justice Elisabeth Laing explained:

Both Birss LJ and Arnold LJ agree that the clear effect of the relevant provisions of the 1977 Act is that an inventor must be a person. I also agree with that conclusion. Rights are a consistent theme which runs through section 7. Only a person can have rights. A machine cannot. The premise of section 7 is that an inventor is and can only be a person. A patent can be granted ‘primarily to the inventor’, and only to someone else in the circumstances described in section 7(2)(c) and (d). A patent is a statutory right and it can only be granted to a person. That means that the effect of section 7(2)(a) is that the inventor must be a person. Only a person can make, before the invention is made, an enforceable agreement by which he is entitled to the whole of the property in the invention (other than equitable interests) (section 7(2)(b)). Such an agreement can only be made with another person. Only a person can have a successor in title (section 7(2)(c)). It follows that, absent a statutory deeming provision, it is simply not possible, as a matter of law, for Dabus to be an ‘inventor’ for the purposes of section 7. Nor, for the reasons given by Arnold LJ, has Dr Thaler identified any enactment of rule of law by which he is entitled to that property (even if he could, that would not help his case, because it would not overcome the hurdle that Dabus in not an inventor for the purposes of the Act). Only a person has a right to be named as an inventor (section 13(1)). Section 13(2)(a) also assumes that the inventor is a person. Arnold LJ gives many other examples of provisions in the 1977 Act which support that construction.

Dr Thaler was not entitled to the inventions

In the Court of Appeal, Dr Thaler appears to have shifted ground from the position taken before Marcus Smith J.

Dr Thaler accepted he was not entitled to the inventions under s 7 as the inventor. At [124], Arnold LJ recorded that Dr Thaler also abandoned his claim to entitlement on the basis of an assignment from DABUS.[3] Instead, Dr Thaler based his entitlement on the doctine of accession. Dr Thaler’s argument was that as the owner of DABUS he was the owner of the inventions it produced just as the owner of a cow is the owner of its calves.

At [130], Arnold LJ cited Blackstone for the common law definition of the doctrine of accession on which Dr Thaler relied:

“The doctrine of property arising from accession is also grounded on the right of occupancy. By the Roman law, if any given corporeal substance received afterwards an accession by natural or by artificial means, as the growth of vegetables, the pregnancy of animals, the embroidering of cloth, or the conversion of wood or metal into vessels and utensils, the original owner of the thing was intitled by right of possession to the property of it under such it’s statement of improvement; but of the thing itself, by such operation, was changed into a different species, as by making wine, oil, or bread, out of another’s grapes, olives, or wheat, it belonged to the new operator; who was only to make a satisfaction to the former proprietor for the materials, which he had converted. And these doctrines are implicitly copied and adopted by our Bracton, in the reign of king Henry III; and have since been confirmed by many resolutions of the courts.”

Arnold LJ explained that the doctrine of accession was rooted in the principle of “dominion” – exclusive possession: a person who has exclusive possession of a tangible (the cow) which produces a tangible (a calf) will generally also have exclusive possession of the tangible produced.

However, an invention is information, an intangible. As an intangible, Arnold LJ noted at [133] it was distinguishable from a tangible on the basis that intangibles are non-rivalrous goods: they were capable of simultaneous consumption by more than one person. Accordingly, a new intangible is not capable of exclusive possession.

Furthermore, Dr Thaler was forced to accept that the doctrine of accession did not apply to at least some situations where an intangible was produced by a tangible. At [135], Arnold LJ gave the example of a photograph taken by A using a camera owned by B: subject to an agreement to the contrary, the copyright in the photograph would be owned by the photographer, not the owner of the camera.[4]

Arnold LJ considered that Dr Thaler’s arguments were more about what the law should be rather than what it is. As to what it is, his Lordship said at [136] it ran into two obstacles:

…. As matters stand, it seems to me that the argument faces two obstacles. The first is that it pre-supposes that a machine can be an inventor for the purposes of the 1977 Act. The second is that I cannot see any basis in current law for a person to have a legal right to stand in the place of a machine with respect to the right to apply for a patent, because that pre-supposes that the machine would otherwise have that right, but as noted above machines do not have rights. A point which underlies both these obstacles is that modern patent law is almost entirely a creature of statute.

Accordingly, Dr Thaler’s applications should be rejected.

Birss LJ dissented

Despite agreeing that an AI was not an inventor for the purposes of the Patents Act 1977, Birss LJ would have allowed the appeal. For his Lordship, it was no part of the Comptroller’s role under section 13 to inquire into whether DABUS was in fact an inventor.

To reach this conclusion, Birss LJ examined the history of s 13 and its predecessor under the Patents Act 1949, s 16. Section 16 of the 1949 Act wrought a number of changes. One of those was to permit assignment of the right to apply for a patent which in turn brought in the requirement that the assignee applicant provide an assent from the inventor.

The Banks Committee, however, reported that these arrangements had led to a number of problems. One problem was the difficulty in confidently identifying the inventors.[5] A second problem was that at the examination stage the Comptroller was in no position to check the accuracy of claims to inventorship or assignment. At [44], Birss LJ explained that the Banks Committee had recommended abolishing declarations of inventorship and assent:

This, we think, would simplify the procedure for applicants, protect patentees against invalidation of their patents through inadvertent error in naming inventors, give recognition to the fact that the Patent Office is in no position to check the completeness of declarations of assent and inventorship at present required (which therefore provide no real safeguard against wrongful obtaining of patents), and abolish the anomaly of “communicated” invention. We realise, however, that safeguards are necessary to protect the interests of inventors, and would stress that making of any application by a person other than the inventor should not imply in any way that the inventor has assented to the making of the application or acknowledged the right of the applicant to make it. Furthermore, since we believe that it is most important that the contribution of the inventor, where he is not the applicant, should not be overlooked, we recommend that the applicant, both in Convention and in non-Convention applications, should be required to name the person(s) believed to be the inventor(s), who would then be named in the published specification, but that the fulfilling of this requirement should not prejudice the right of any other person to apply under section 16 of the [1949] Act to be mentioned in the patent, nor should a bona fide error in the naming of inventors invalidate a patent. Where disputes arise over ownership of applications or patents, a new procedure will be necessary for their resolution, and this we recommend in Chapter 13.

Having regard to this history, Birss LJ considered at [58] – [59] that the requirements of s 13(2)(a) would be satisfied so long as the applicant identified who the applicant believed was the inventor. Correspondingly, s 13(2)(b) merely required the applicant to identify the basis on which the applicant claimed derivation from the inventor.

It was no part of the Comptroller’s role, at this stage, to investigate whether or not the identified inventor was in fact legally an inventor or the identified basis of derivation was legally effective. Any other approach would resurrect the very problems encountered under the 1949 Act [s 16][49uk].

In light of this conclusion, his Lordship did not consider Dr Thaler’s arguments based on the doctrine of accession.

With respect, a difficulty with Birss LJ’s approach is that an application would proceed to grant and could be asserted against third parties, or block subsequent applications, unless someone came forward to challenge the grant. And under the UK Act it seems objections on the ground of entitlement can be brought only by someone else claiming entitlement: See Birss LJ at [70] citing s 72.

Arnold LJ (with the agreement of Lady Justice Elisabeth Laing) accepted that the role of the Comptroller under s 13 was limited. His Lordship considered, however, rejection of Dr Thaler’s application did not involve such issues. Rather, the Comptroller took the information provided by Dr Thaler at face value and on its face that information failed to identify the person who was the inventor or any basis operative to transfer the rights of an inventor to Dr Thaler.[6]

As the applications failed to comply with two important statutory requirements, therefore, Arnold LJ held at [148] that the Comptroller was right to reject them.

A moral right

Before concluding, it is worth also noting that the Court of Appeal recognised that the requirement to identify the inventor is really a species of moral right. As Arnold LJ noted at [121], a paternity right: the right to be identified as the creator. At [44], Birss LJ noted the Banks Committee’s recognition that:

it is most important that the contribution of the inventor, where he is not the applicant, should not be overlooked, we recommend that the applicant, both in Convention and in non-Convention applications, should be required to name the person(s) believed to be the inventor(s), who would then be named in the published specification….

Thaler v Comptroller General of Patents Trade Marks And Designs [2021] EWCA Civ 1374 (Arnold and Elisabeth Laing LJJ, Birss LJ dissenting)


  1. See Foster’s Australia Limited v Cash’s (Australia) Pty Ltd [2013] FCA 527. Section 31 extends s 29 to joint applicants.  ?
  2. Section 13 of the UK Act does not have a counterpart in the Australian Act. However, ss 29(4A) and (4B) require applications to be in the approved form – which requires the inventor to be identified and reg. 3.2C(2)(aa) requires applications under the PCT to identify the inventor.  ?
  3. Before Beach J in Australia, Dr Thaler also accepted that, as an artificial intelligence system is not a legal person, DABUS could not legally assign the invention: see e.g. [115].  ?
  4. In Australia, see Copyright Act 1968 s 35, bearing in mind there are exceptions for works made in the course of employment, taking of photographs for a private or domestic purpose or the exception for photographs taken by journalists for newspapers etc.  ?
  5. A problem finally addressed in the Patents Act 1990 (Cth) by the enactment in the “Raising the Bar” Act of s 22A and s 138(4).  ?
  6. At [108], Lady Justice Elisabeth Laing explained “It will be clear from what I have already said that I do not agree that section 13(2)(a) only requires an applicant to ‘state their genuine belief about who the inventor was’ (judgment, paragraph 60). Rather, it expressly requires the applicant to identify the person who, he believes, is the inventor. That is a different requirement, and it is not met by a statement that the applicant genuinely believes that the invention was devised by a machine ….” (emphasis supplied)  ?

Should AI systems be classifiable as patent inventors?

Last month, Melbourne Uni. held a seminar in which Professors Gans (Toronto), Paterson (Melbourne) and Weatherall debated whether AI systems could or should be considered inventors for the purposes of patent law.

A video recording of their discussion is now available here.

Meanwhile, the Max Planck Institute in Germany has published:

More on DABUS – this time in Old Blighty

Ann Dufty, whom many of you will know, has pointed out to me that Dr Thaler’s lack of success in the UK was not limited “merely” to rejection by UKIPO. Dr Thaler’s patent application has been rejected by Marcus Smith J sitting in the High Court of England and Wales.

Under the Patents Act 1977 (UK) s 7(2) is in similar terms and, one might say, to the same effect as s 15 of the Australian Act:

(1) Any person may make an application for a patent either alone or jointly with another.

(2) A patent for an invention may be granted –

(a) primarily to the inventor or joint inventors; 
(b) in preference to the foregoing, to any person or persons who, by virtue of any enactment or rule of law, or any foreign law or treaty or international convention, or by virtue of an enforceable term of any agreement entered into with the inventor before the making of the invention, was or were at the time of the making of the invention entitled to the whole of the property in it (other than equitable interests) in the United Kingdom; 

(c) in any event, to the successor or successors in title of any person or persons mentioned in paragraph (a) or (b) above or any person so mentioned and the successor or successors in title of another person so mentioned; and to no other person.

(3) In this Act ‘inventor’ in relation to an invention means the actual deviser of the invention and ‘joint inventor’ shall be construed accordingly.

(4) Except so far as the contrary is established, a person who makes an application for a patent shall be taken to be the person who is entitled under subsection (2) above to be granted a patent and two or more persons who make such an application jointly shall be taken to be the persons so entitled.”

Sub-section 7(3) does not have an obvious counterpart in the Australian legislation but, DABUS aside, one would think Australian law is to the same effect. Putting to one side the question whether or not an AI can be an inventor, Australian law would consider the person who made the invention the inventor.[1]

In the English case, Dr Thaler’s substantive arguments were essentially the same as those advanced in Australia.

As to the meaning of “inventor”, Marcus Smith J at [45(3)] first noted that Lord Hoffmann in Yeda had agreed with Laddie J’s view that an inventor was a “natural person”:

The inventor is defined in section 7(3) as “the actual deviser of the invention”. The word “actual” denotes a contrast with a deemed or pretended deviser of the invention; it means, as Laddie J said in University of Southampton’s Applications [2005] RPC 220, 234, the natural person who “came up with the inventive concept.” It is not enough that someone contributed to the claims, because they may include non-patentable integers derived from prior art: see Henry Brothers (Magherafelt) Ltd v Ministry of Defence [1997] RPC 693, 706; [1999] RPC 442. As Laddie J said in the University of Southampton case, the “contribution must be to the formulation of the inventive concept”. Deciding upon inventorship will therefore involve assessing the evidence adduced by the parties as to the nature of the inventive concept and who contributed to it. In some cases this may be quite complex because the inventive concept is a relationship of discontinuity between the claimed invention and the rior art. Inventors themselves will often not know exactly where it lies.

Marcus Smith J then noted that he had not been cited any authority which explained why inventors were confined to natural persons only. His Lordship said at (a):

There is no authority to which I was referred or which I have myself been able to find which explains why the inventor is limited to natural persons only, as opposed to including also legal persons. Whilst one can see the need to limit Class (a) and so the term “inventor” to someone having personality, the exclusion of legal persons from the definition seems less clear-cut. The 1977 Act could, after all, have explicitly referred to “natural persons” rather than just the “inventor”.

Next his Lordship pointed out that DABUS was not, on any view, a person. Then, in contrast to Beach J at [135] – [145], Marcus Smith J considered that the requirement for a valid patent to have an inventive step was decisive:

It seems to me that, when once the notion of an “inventive step” is factored in, the restriction of the term “inventor” to natural person becomes inevitable. An “invention” by definition[29] must involve an “inventive step”, which is something “not obvious to a person skilled in the art”.[30] It is difficult to see how an inventive step can conceived of by a corporation – which must act through agents – without also striking one of those agents. In other words, the inventive step in the mind of a natural person is attributed to the corporation, which only has the inventive step in its “mind” by virtue of such attribution.[31] There is some sense in keeping the definition of inventor close to that which must arise out of the mind of an individual.

Accordingly, DABUS was not an inventor for the purposes of UK law.

Thaler v The Comptroller-General of Patents, Designs and Trade Marks [2020] EWHC 2412 (Pat)


  1. University of Western Australia v Gray [2009] FCAFC 116; 179 FCR 346 at [248] citing [Polwood Pty Ltd v Foxworth Pty Ltd][polworth] (2008) 165 FCR 527. For the counterparts to s 7(1) of the UK Act see ss 29 and 31.  ?

DABUS “over there”

Judge Brinkema, sitting as a District Court Judge in the Eastern District of Virginia, has upheld the USPTO’s rejection of Thaler’s DABUS applications on the basis that DABUS cannot be an inventor under the US Act.

In the United States, Dr Thaler has two patent applications – US Application Serial Nos 16/524,350 and 16/534,532. In both, DABUS was the nominated inventor and Dr Thaler claims entitlement on the basis of assignment.

As you will no doubt recall, DABUS is a “creativity machine” or artificial intelligence.

To highlight the ludicrousnessfictional nature of the universe we are operating in, Dr Thaler as the owner of DABUS executed the assignment to himself:

In view of the fact that the sole inventor is a Creativity Machine, with no legal personality or capability to execute said agreement, and in view of the fact that the assignee is the owner of said Creativity Machine, this Assignment is considered enforceable without an explicit execution by the inventor. Rather, the owner of DABUS, the Creativity Machine, is signing this Assignment on its behalf.

When the America Invents Act was passed, amongst other things it inserted a definition of “inventor” into the Act so that 35 USC §100(f) provides:

(f) The term “inventor” means the individual or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention.

Perhaps (with respect) unsurprisingly, Judge Brinkema ruled that “individual” meant a natural person.

In doing so, her Honour was fortified by the natural or ordinary meaning of the word. Contextually, there were also other references in the Act where Congress had used the term “individual” in reference to the inventor. (For example, §115(a)(1) and (b)(2).)

In addition, the Supreme Court had construed the term “individual” in the Torture Victim Protection Act as referring to a “natural person”. And several Federal Circuit decisions had declared that “inventors must be natural persons” albeit not in the context of the meaning of §100(f).

Judge Brinkema then explained that Dr Thaler “having neither facts nor law to support his argument” contends that policy considerations and the general purpose of the Constitution’s Patent Clause required the statute to be interpreted to permit AIs to be inventors:

Allowing patents for AI-Generated Inventions will result in more innovation. It will incentivize the development of AI capable of producing patentable output by making that output more valuable …. Patents also incentivize commercialization and disclosure of information, and this incentive applies with equal force to a human and an AI-Generated Invention. By contrast, denying patent protection for AI-Generated Inventions threatens to undermine the patent system by failing to encourage the production of socially valuable inventions.

Patent law also protects the moral rights of human inventors and listing an AI as an inventor where appropriate would protect these human rights …. [I]t will discourage individuals from listing themselves as inventors without having contributed to an invention’s conception merely because their name is needed to obtain a patent. Allowing a person to be listed as an inventor for an AI-Generated Invention would not be unfair to an AI, which has no interest in being acknowledged, but allowing people to take credit for work they have not done would devalue human inventorship.

Judge Brinkema considered that binding rulings of the Supreme Court and the Federal Circuit repeatedly held that policy arguments could not override a statute’s plain language. Her Honour also pointed out that, when Congress passed the America Invents Act, AIs were in existence and it was aware of them. Moreover, the USPTO’s own consultations had not exposed any strong support for AIs to be inventors.

Ruling against Thaler, Judge Brinkema concluded:

As technology evolves, there may come a time when artificial intelligence reaches a level of sophistication such that it might satisfy accepted meanings of inventorship. But that time has not yet arrived, and, if it does, it will be up to Congress to decide how, if at all, it wants to expand the scope of patent law.

What does this mean for Australia?

Plainly, the American context is not directly applicable to Australia since, as Beach J pointed out at [118], our Act does not have a definition of “inventor”. So, there is much greater scope for policy arguments to operate.

In that connection, the USPTO report cited by Judge Brinkema can be found here.

Ordinarily, I would be on the side of progress: the NRDC view of the world rather than D’Arcy v Myriad. Our courts, of course, must fit within the D’Arcy v Myriad world view unless Parliament were to bestir itself.

Apart from South Africa (which I understand does not undertake substantive examination of patent applications), Dr Thaler’s applications have been rejected on the ground that an AI is not an inventor by the UKIPO and EPO as well as in the USA. Government policy, which appears to have aligned with the Productivity Commission‘s argument that Australia as an intellectual property importing nation should not be out of step with the international environment, would suggest that an AI should not qualify as an inventor. Can we really afford to keep repeating the mistake made in the 3M case? However, an appeal is pending in the EPO. Maybe there will be an appeal in the USA too but the Federal Circuit’s prior indications do not augur well for the success of that.

It is also difficult to comprehend why, if as our Courts have ruled, that authors for copyright purposes must be humans, the same does not apply to inventors. Of course, our law now explicitly recognises moral rights as part of an author’s rights and there is no corresponding provision under Australian patent law. But both types of rights are justified by the same rationales – natural law or Lockean theory of property and, even, the so-called utilitarian theory.

I guess we shall see.

Thaler v Hirshfield ED VA, 2 September 2021 1:20-cv-903 (LMB/TCB)

Lid dip, Prof. Dennis Crouch at Patently-O.

DABUS Down Under take 3

Following last month’s ruling in Thaler that an AI could be an inventor for the purposes of Australian patent law, the Commissioner of Patents has announced her intention to appeal the decision to the Full Court.

Pursuant to s 158(2), the Commissioner requires leave to appeal. Bearing in mind that Beach J’s decision is the first substantive consideration anywhere in the world to accept that an AI could be an inventor for the purposes of the Patents Act, however, that should not prove too much of an obstacle in this case.

Thaler v Commissioner of Patents [2021] FCA 879

Thaler: the robots have arrived DownUnder

In what may well be a world first,[1] Beach J has upheld Thaler’s appeal from the Commissioner, ruling that an AI can be an inventor (or at least that someone who derives title to the invention from an AI can be an entitled person).

Stephen L. Thaler applied for a patent, AU 2019363177, entitled “Food container and devices and methods for attracting enhanced attention”.[2] The application named the inventor as:

DABUS, The invention was autonomously generated by an artificial intelligence

The application was made through the PCT so, as a result, reg. 3.2C(2)(aa) required the applicant to provide the name of the inventor. The Commissioner had used the identification provided to reject the application on the basis that an “inventor” must be a natural person, which an AI obviously was not.

Beach J rejected this approach. At [10], his Honour summarised his conclusions:

in my view an artificial intelligence system can be an inventor for the purposes of the Act. First, an inventor is an agent noun; an agent can be a person or thing that invents. Second, so to hold reflects the reality in terms of many otherwise patentable inventions where it cannot sensibly be said that a human is the inventor. Third, nothing in the Act dictates the contrary conclusion.

There are a number of strands to his Honour’s reasoning. Perhaps, the most striking feature of his Honour’s reasoning is his Honour’s emphasis the role of patents in encouraging technological development and the importance of not impeding progress by shutting out new developments.

DABUS

Dr Thaler was the owner of the copyright in DABUS’ source code. He was also responsible for and the operator of the computer on which DABUS operated. He did not, however, produce the claimed invention.

As the description of the “inventor” indicates, the claimed invention was an output from the operation of DABUS.

Beach J did not propose to offer a definition of “artificial intelligence”. His Honour considered that DABUS, at least as described in Dr Thaler’s evidence, was not just a “brute force computational tool”. Instead, it was appropriate to describe DABUS as semi-autonomous.[3]

DABUS itself consisted of multiple neural networks. At first, the connections between these networks was trained with human assistance by presentation of fundamental concepts. As the system accumulated knowledge, the networks connected themselves into increasingly longer chains. And then DABUS became capable of generating notions itself – the unsupervised (by humans) generative learning phase. Once in this phase, DABUS was capable of randomn generation of concepts and identifying those which were novel. In addition, it was trained or programmed to identify significant concepts which continued operation further reinforced. At [41], Beach J summarised:

The upshot of all of this, which I accept for present purposes, is that DABUS could be described as self-organising as a cumulative result of algorithms collaboratively generating complexity. DABUS generates novel patterns of information rather than simply associating patterns. Further, it is capable of adapting to new scenarios without additional human input. Further, the artificial intelligence’s software is self-assembling. So, it is not just a human generated software program that then generates a spectrum of possible solutions to a problem combined with a filtering algorithm to optimise the outcome.

Beach J accepted at [42] Dr Thaler’s evidence that:

DABUS, and its underlying neural paradigm, represents a paradigm shift in machine learning since it is based upon the transient chaining topologies formed among associative memories, rather than activation patterns of individual neurons appearing within static architectures ….

Beach J’s reasons

At [118], Beach J pointed out that no specific provision in the Act precluded an artificial intelligence from being an “inventor”.

Secondly, Beach J considered that patent law is different to copyright law which specifically requires human authors and recognition of moral rights.

Thirdly, as it was not defined in the Act, the term “inventor” should be given its ordinary meaning. Dictionary definitions did not help with this as more was required “than mere resort to old millennium usages of that world.”[4] Instead, at [120]:

the word “inventor” is an agent noun. In agent nouns, the suffix “or” or “er” indicates that the noun describes the agent that does the act referred to by the verb to which the suffix is attached. “Computer”, “controller”, “regulator”, “distributor”, “collector”, “lawnmower” and “dishwasher” are all agent nouns. As each example demonstrates, the agent can be a person or a thing. Accordingly, if an artificial intelligence system is the agent which invents, it can be described as an “inventor”.

Importantly, Beach J took into account the purpose of the patents system. At [121], his Honour explained:

in considering the scheme of the Act, it has been said that a widening conception of “manner of manufacture” is a necessary feature of the development of patent law in the twentieth and twenty-first centuries as scientific discoveries inspire new technologies”.[5]

Accordingly, it made little sense to apply a flexible conception of subject matter – “manner of (new) manufacture under s 18(1)(a) – but a restrictive interpretation of ”inventor“. This would mean an otherwise patentable invention could not be patented because there was no ”inventor”.

Beach J’s purposive approach was reinforced by reference to the newly enacted objects clause:

The object of this Act is to provide a patent system in Australia that promotes economic wellbeing through technological innovation and the transfer and dissemination of technology. In doing so, the patent system balances over time the interests of producers, owners and users of technology and the public.

It was not necessary to identify an ambiguity before resorting to the objects clause. The objects clause should always be considered when construing legislation.

It was consistent with the object of the Act to interpret “inventor” in a way which promoted technological innovation. Allowing “computer inventorship” would incentivise computer scientists to develop creative machines. Moreover, the object of the Act would not be advanced if the owners of creative computers resorted to trade secret protection instead of the patent system.

At least, with respect, his Honour’s approach shows that s 2A can operate in favour of technological development by encouraging patenting rather than, as many feared, a cat’s paw justifying resort to an ex post analysis[6] in favour of ‘user rights’.

In a further bold development, Beach J considered ([135] – [145]) that the Act is “really concerned” with inventive step. Under s 7(2), the issue was whether or not the patent claimed a sufficient technological advance over what had gone before to warrant the grant of a monopoly. How that advance was made was not relevant.

Descending to the detail of s 15, Beach J first pointed out at [157] that s 15 is directed to who may be granted a patent and does not define who is an inventor. Beach J accepted that DABUS was not a person and so under s15 could not be a person entitled to the grant of a patent. Even though s 15(1)(a) identified “a person who is the inventor” as the first person entitled to the patent, this was not determinative. It was directed to a different issue than definition of “an inventor”.

Although DABUS could not be an entitled person, Beach J found that Dr Thaler qualified as an entitled person at least on the basis of s 15(1)(b). At [167], his Honour explained:

Dr Thaler is the owner, programmer and operator of DABUS, the artificial intelligence system that made the invention; in that sense the invention was made for him. On established principles of property law, he is the owner of the invention. In that respect, the ownership of the work of the artificial intelligence system is analogous to ownership of the progeny of animals or the treatment of fruit or crops produced by the labour and expense of the occupier of the land (fructus industrialis), which are treated as chattels with separate existence to the land.

By this means, his Honour neatly side-stepped philosophical questions, some of which he had adverted to earlier. For example, at [131] Beach J asked:

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? ….

Beach J returned to this aspect of the problem at [194]:

more generally there are various possibilities for patent ownership of the output of an artificial intelligence system. First, one might have the software programmer or developer of the artificial intelligence system, who no doubt may directly or via an employer own copyright in the program in any event. Second, one might have the person who selected and provided the input data or training data for and trained the artificial intelligence system. Indeed, the person who provided the input data may be different from the trainer. Third, one might have the owner of the artificial intelligence system who invested, and potentially may have lost, their capital to produce the output. Fourth, one might have the operator of the artificial intelligence system. But in the present case it would seem that Dr Thaler is the owner.

As Dr Thaler combined in the one person the roles of owner, programmer and operator, he was entitled to the fruits of its operation. Would different problems arise if, instead of being embodied in the one person, each of the functions identified lay in a different person?

Returning to [131], Beach J continued immediately following the extract quoted above, saying:

…. In my view, in some cases it may be none of the above. In some cases, the better analysis, which is consistent with the s 2A object, is to say that the system itself is the inventor. That would reflect the reality. And you would avoid otherwise uncertainty. And indeed that may be the case if the unit embodying the artificial intelligence has its own autonomy. What if it is free to trawl the internet to obtain its own input or training data? What about a robot operating independently in a public space, having its own senses, learning from the environment, and making its own decisions? ….

If one can start with the AI as the inventor, that arguably simplifies the analysis in terms of entitlement as the person claiming to be the entitled person will need to show some claim over the results of the operation of the machine.

One final point. It is not clear from the reasons the extent to which Beach J was referred to the controversies around the world arising from Dr Thaler’s applications. It didn’t matter.

At [220], his Honour pointed out that they were irrelevant to the task before him: the interpretation of the words of the Australian Act.

I am not at all sure that “the world” has reached a settled position about the question whether AIs can be inventors. One would hope, however, Australian law does not head down yet another path where “we” are granting patents for things which are not patentable in their ‘home’ countries. It will, for example, be another 12 years or so before we are finally in something like parity on inventive step with the “rest” of the world and have escaped the constraints of the pre-Raising the Bar tests of inventive step.

Two questions

This short summary cannot do justice to the detailed arguments developed over 228 paragraphs.

As discussed above, Beach J accepted that DABUS could not be granted a patent as it was not a person and s 15 specifies that the grantee of a patent must be a person. The Commissioner had proceeded on the basis that the terms of s 15 were predicated on the “old millennium” understanding that title to an invention flowed from the person who was the inventor just as subsistence and ownership of copyright flows from the person who is the author of the work. Beach J has sidestepped that on the basis that s 15 is concerned only with entitlement, not definition of who is an inventor. Nonetheless, one might think s 15 was drafted in this way on the basis that entitledment flowed from the inventor. It does also seem somewhat curious that an AI can be an inventor but not entitled to a patent because it is not a person.

Secondly, much of the controversy overseas has been about whether Dr Thaler’s creation acts autonomously or semi-autonomously or is just an exercise in brute computing. See for example Rose Hughes’ report on IPKat. Beach J appears to have had some evidence from Dr Thaler about how DABUS was designed and worked. Thus at [43], his Honour accepted Dr Thaler’s “assertion” that:

DABUS, and its underlying neural paradigm, represents a paradigm shift in machine learning since it is based upon the transient chaining topologies formed among associative memories, rather than activation patterns of individual neurons appearing within static architectures. From an engineering perspective, the use of network resonances to drive the formation of chaining topologies, spares programmers the ordeal of matching the output nodes of one [artificial neural network] with the input nodes of others, as in deep learning schemes. In effect, complex neural architectures autonomously wire themselves together using only scalar resonances.

Reinforcement or weakening of such chains takes place when they appropriate special hot button nets containing memories of salient consequences. Therefore, instead of following error gradients, as in traditional artificial neural net training, conceptual chains are reinforced in proportion to the numbers and significances of advantages offered. Classification is not in terms of human defined categories, but via the consequence chains branching organically from any given concept, effectively providing functional definitions of it. Ideas form as islands of neural modules aggregate through simple learning rules, the semantic portions thereof, being human readable as pidgin language.

Later his Honour asked at [127] – [128]:

Who sets the goal for the system? The human programmer or operator? Or does the system set and define its own goal? Let the latter be assumed. Further, even if the human programmer or operator sets the goal, does the system have free choice in choosing between various options and pathways in order to achieve the goal? Let that freedom also be assumed. Further, who provides or selects the input data? Let it be assumed that the system can trawl for and select its own data. Further, the larger the choice for the system in terms of the algorithms and iterations developed for the artificial neural networks and their interaction, the more autonomous the system. Let it be assumed that one is dealing with a choice of the type that DABUS has in the sense that I have previously described.

Making all of these assumptions, can it seriously be said that the system is just a brute force computational tool? Can it seriously be said that the system just manifests automation rather than autonomy? ….

If by “assumptions” his Honour is referring to the evidence from Dr Thaler at [42] and [43] which his Honour accepted, that is one thing. It may be quite another thing if they were in fact assumptions.

Where to now?

At the time of writing, the Commissioner is understood to be considering whether or not to appeal.

Thaler v Commissioner of Patents [2021] FCA 879


  1. The EPO refused the corresponding patent application with the oral hearing of the appeal to be heard on 21 December 2021. Apparently, the corresponding patent has been granted in South Africa which, I am given to understand, effectively does not operate a substantive examination system.  ?
  2. For an interesting discussion, see “The first AI inventor – IPKat searches for the facts behind the hype” and the later report on the USPTO’s consultations “Is it time to move on from the AI inventor debate? ?
  3. At [19] – [29], Beach J provides an overview of how his Honour understands artificial neural networks work.  ?
  4. At [15]. For the lawyers, [148] – [154] set out the technical arguments for the limitations of dictionaries in some detail.  ?
  5. Diplomatically (and consistently with precedent) citing D’Arcy v Myriad Genetics Inc (2015) 258 CLR 334 at [18] perhaps the single biggest retreat from the teleological approach declared in NRDC.  ?
  6. For example, Mark A Lemley, ‘Ex Ante versus Ex Post Justifications for Intellectual Property ?

Artificial intelligences and inventions Down Under

The Commissioner of Patents has rejected the DABUS application Down Under.

Stephen L. Thaler applied for a patent, AU 2019363177, entitled “Food container and devices and methods for attracting enhanced attention”. The application named the inventor as:

DABUS, The invention was autonomously generated by an artificial intelligence

The application being made under the PCT, there was a formalities check, which, in reg. 3.2C(2)(aa), requires the Commissioner to check whether the named inventor has been identified.[1]

When the Delegate objected that an inventor had not been identified. Thaler explained why he considered DABUS was the inventor (in part):

The sole contributor to the invention is DABUS, an artificial intelligence machine that includes artificial intelligence programs written by the applicant. DABUS is capable of devising inventions without the involvement of a natural person who traditionally qualifies as an inventor. For the present invention, the machine only received training in general knowledge and proceeded to independently conceive of the invention and to identify it as novel and salient. How DABUS functions is described in detail in US Patent 10,423,875 and other patent specifications.

The Delegate understood from this response that DABUS is not a person as understood in law – an individual, a corporation or a body politic.[2]

There is no definition of “inventor” in the Act. Thus, Wilcox and Lindgren JJ had declared that the word bears its ordinary English meaning.[3]

At [12], the Delegate said:

…. Any standard dictionary shows that the traditional meaning of inventor is a person who invents. At the time that the Act came into operation (in 1991) there would have been no doubt that inventors were natural persons, and machines were tools that could be used by inventors. However, it is now well known that machines can do far more than this, and it is reasonable to argue that artificial intelligence machines might be capable of being inventors. I have no evidence whether the ordinary meaning of “inventor”, assessed at the present day, can include a machine. But if this were the ordinary meaning, would this be consistent with the other provisions of the Act?

So far as the other provisions and context provided any (limited) assistance, the Delegate considered at [20] that it was clear a patentee must be a person. This implied that an inventor also needed to be a person and, in any event, an inventor who was not a person could not be a patentee.

Although it was not part of the decision, it may also be noted that an author for the purposes of copyright must be a natural person. A computer-generated work is not an original work for the purposes of copyright as there is no author.[4] Of course, a patent can protect ideas or function while copyright protects the “expression” of ideas. At least to the extent that the rationale for granting protection in either system is the natural rights of a person to the fruits of their mental labour,[4] one would think the same considerations should apply.

Thaler has enlisted the services of Allen’s pro bono and appealed, No. VID 108/2021.

Stephen L. Thaler [2021] APO 5]


  1. Correct identification of the inventor(s) is important as a patent can be revoked if it is not granted to an “entitled person” (or all the “entitled persons” (see s 138(3)(a)) and a person can be an “entitled person” only if they can trace their chain of title back to the inventor(s): s 15(1)(a).  ?
  2. Citing Acts Interpretation Act 1901 (Cth) s 2C.  ?
  3. Atlantis Corporation Pty Ltd v Schindler [1997] FCA 1105; 39 IPR 29 at 54.  ?
  4. Telstra Corporation Limited v Phone Directories Company Pty Ltd [2010] FCAFC 149 at [90] (Keane CJ) and [117] – [119] (Perram J).  ?

IP Australia and Indigenous Knowledge – consultations

In September 2020, IP Australia published its work plan to make provision for the better protection of Indigenous Knowledge in Australia’s Intellectual Property System.

Now it has released a consultation paper on four topics from that work plan it wishes to advance.

As summarised in the Introduction, the consultation paper raises four topics:

  1. Establishing an Indigenous Advisory Panel – providing a formalised Indigenous voice to IP Australia.
  2. Measures for trade mark or designs using Indigenous Knowledge – changes to processes to ensure IK owners benefit from, or have consented to, the use of their IK as the basis for rights.
  3. New requirements to declare the source of Indigenous Knowledge used in new innovations – make it easier to determine if IK has been used in a patent or plant breeder’s right, and encourage conversations about access and benefit sharing.
  4. Labelling to promote authentic Indigenous Products – exploring interest in labelling schemes that distinguish authentic Aboriginal or Torres Strait Islander goods.

Topic 2, relating to trade marks and designs, notes that IP Australia may currently reject an application if the application uses “IK” which is secret or sacred; the name of a group or a nation where there is no connection to that group or nation or uses a an Aboriginal or Torres Strait Islander word which should be available for other business to use.

Noting the limitations in that range, the consultation paper seeks input on a range of issues including:

(a) whether people would have concerns providing a statutory declaration etc. as evidence to support an objection to an application;

(b) whether IP Australia should ask applicants whether they have consent to use the “IK”;

(c) the introduction of a check to assess whether an application would cause cultural offence to a community or communities;

(d) whether IP Australia should assess whether the the application involves a use of “IK” in a way which falsely suggests a connection to an Indigenous person, community or nation.

Other questions relate to tools for better identifying applications which involve the use of “IK”.

Topic 3 raises questions about whether patent applicants should be required to declare the source of (1) genetic resources and/or (2) traditional knowledge included in their applications or on which an application is based.

Additional questions relate to how such requirements would be implemented, enforced and, if not complied with, penalised.

The Overview page states that consultations close on 24 May 2021.