USPTO Shoots Down DABUS’ Bid For Inventorship

“The USPTO reasoned that ‘conception—the touchstone of inventorship—must be performed by a natural person.’” week, the United States Patent and Trademark Office (USPTO) issued a decision refusing to vacate a Notice of Missing Parts in U.S. Patent Application No.: 16/524,350 (the ‘350 Application), titled “Devices and Methods for Attracting Enhanced Attention”, for failure to “identify each inventor by his or her legal name” on the Application Data Sheet (ADS). The ADS listed a single inventor with the given name DABUS and the family name as “Invention generated by artificial intelligence.” DABUS stands for “Device for the Autonomous Bootstrapping of Unified Sentience”. The Application listed Stephen L. Thaler as the Assignee, Applicant and the Legal Representative. The USPTO denied the petition to vacate the Notice of Missing Parts.

USPTO Cites U.S. Code, CAFC Precedent,  MPEP in Denying Petition

The Petitioner asserted that the claimed invention in the ‘350 Application was developed by a creativity machine named DABUS, which was trained with general information in the field of endeavor to independently create the invention.” The Petitioner further argued that DABUS should be named as the inventor and that inventorship in general should not be limited to “natural persons”.

The USPTO noted that Title 35 of the United States Code repeatedly refers to inventors as natural persons. For example, the USPTO explained, 35 U.S.C. § 101 states that “Whoever invents or discovers…”, wherein the term “whoever” suggests a natural person. The USPTO also referred to 35 U.S.C. § 115, which uses terms such as “himself”, “herself”, “individual”, and “person”. The USPTO reasoned that interpreting the term “inventor” to include machines “would contradict the plain reading of the patent statutes that refer to persons and individuals.”


The Office also referred to U.S. Court of Appeals for the Federal Circuit (Federal Circuit) precedent to support the conclusion that an inventor must be a natural person. With reference to Univ. of Utah v.Max-Planck-Gesellschafl zur Forderung der Wissenschaflen e. V, the USPTO explained that the Federal Circuit has ruled that a state could not be an inventor because inventors are individuals who conceive of an invention and conception is a “formation of the mind of the inventor” and “a mental act”. The USPTO reasoned that “conception—the touchstone of inventorship—must be performed by a natural person.” The agency added that the Manual of Patent Examining Procedure (MPEP) defines “conception” as “the mental part of the inventive act” and “the formation in the mind of the inventor”, thereby reinforcing the USPTO’s view that conception must be performed by a natural person.

The USPTO next considered several of the Petitioner’s additional arguments, but concluded that the arguments did not overcome the plain language of the patent statutes and Federal Circuit precedent. For example, the Petitioner argued that the USPTO has previously granted patents directed to the DABUS machine, and in doing so, “implicitly legalized the process by which DABUS arrives at an invention” for the ‘350 Application. The USPTO rejected Petitioner’s argument, stating that the “granting of a patent under 35 U.S.C. § 151 for an invention that covers a machine does not mean that the patent statutes provide for that machine to be listed as an inventor in another patent application any more than a patent for a camera allows the camera to hold a copyright.”

EPO and UKIPO Decisions

The USPTO declined to make any determinations regarding who actually “created” the invention, despite the Petitioner’s urging to consider the positions taken by the European Patent Office (EPO) and the UK Intellectual Property Office (UKIPO). In January 2020, the EPO and the UKIPO each rejected patent applications that designated DABUS as the inventor. The UKIPO published a decision setting out its reasons for refusing the DABUS. In its decision, the UKIPO explained that it accepted that DABUS created the invention, but noted that the application must be withdrawn because DABUS is not a person as envisioned by Patent Act and cannot be considered an inventor. The UKIPO also noted that AI Inventions are likely to become more prevalent and it is correct to debate the issues surrounding such inventions, since the current patent system does not provide a means for handling such inventions.

The EPO’s opinion explained that the name of a machine does not meet the requirements of Rule 19(1) EPC, which requires a family name, given names, and a full address of the inventor. The EPO further explained that legislation would be required to create a legal personality for AI systems or machines because “legislative history shows that the legislators of the EPC were in agreement that the term ‘inventor’ refers to a natural person only.” According to the EPO, legal personality is assigned to a natural person as a consequence of being human or to a legal fiction as a result of being created by legislation or jurisprudence. The EPO also pointed out that requiring an inventor to be a natural person appears to be an “internationally applicable standard” followed by the majority of the EPC contracting states, as well as China, Japan, Korea and the USA. Further, the EPO stated that “No national law has been determined which would [recognize] a thing, in particular an Al system or a machine, as an inventor.”

Image Source: Deposit Photos
Image ID: 212678242
Copyright: phonlamai


Warning & Disclaimer: The pages, articles and comments on do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author as of the time of publication and should not be attributed to the author’s employer, clients or the sponsors of

Join the Discussion

49 comments so far.

  • [Avatar for Pat]
    May 24, 2020 03:28 pm

    Anon, But it’s the AIPLA. I haven’t been spammed by them yet…

  • [Avatar for Anon]
    May 24, 2020 08:53 am


    Not liking the link, as it is behind a “clickwrap” wall.

  • [Avatar for Pat Eriksson]
    Pat Eriksson
    May 23, 2020 12:28 pm

    I found a wonderful clarification of DABUS through a recent AIPLA presentation by Thaler himself,

  • [Avatar for Anon]
    May 13, 2020 07:56 am

    I see that this thread is fading into the back pages of the blog without return engagement (on point) from MaxDrei and ted.

    Thanks to MaxDrei, we see that the story is not over, and we will likely have another chance to pick up the discussion left incomplete here.

  • [Avatar for Anon]
    May 10, 2020 08:09 pm

    Regardless whether Thaler’s Dabus machine is the “actual deviser”

    Full stop.

    This is simply not a ‘regardless’ situation.

    This is entirely outside of whatever proposition that Jehan may advance. See my posts herein.

    I think I see why the error of your attempt at inserting the ‘training of the patent attorney’ into the situation, and quite in fact, this attempt points out the error of your thinking.

    Further, it is entirely unclear just what point it is that you wish to make with introducing copyright into the discussion. As you (should) know, copyright and patent are neither the same nor mutually exclusive. These two vehicles of intellectual property protection are instead intended to protect two different ASPECTS of IP in two different ways.

    One is intended to protect expression, while the other is intended to protect utility.

    Thus your attempt to use “reduce to material form” is inapposite to a discussion entirely within the discussion of patent protection. Any comparing and contrasting then will be a non sequitur to the current conversation and the distinguishing between conceiving and inventing – particularly in the thought experiment provided that shows just how different it is (and must be) between a conception and a mere later reading of that conception.

    As I have noted more than once now, unless and until YOU (and the likes of ted) walk through this thought experiment presented to you and can account for the points presented therein, your assertions simply fall flat and cannot carry the legal point that you wish to assert.

  • [Avatar for MaxDrei]
    May 10, 2020 04:19 pm

    See link in #38. Thaler’s patent attorney, Jehan, imagines a situation in which an AI system i) generates a new concept ii) evaluates that concept, and iii) concludes whether the concept has technical value and utility. If the AI system has done all of that, Jehan suggests, then it is proper to name that AI system as an inventor of that concept.

    Question. Regardless whether Thaler’s Dabus machine is the “actual deviser” of any technical feature combination, do we find in Dabus an AI system that does all of what Jehan prescribes as those steps indicative of inventorship, namely ii) a step of evaluating, and iii) a step of concluding?

    Or was it something (or somebody) other than Dabus that performed the evaluation and the drawing of conclusions?

  • [Avatar for Anon]
    May 10, 2020 06:54 am

    I accept your concession that the one that who actually expresses the subject matter of the application in its independent claim (who is TYPICALLY a patent attorney) is NEITHER devisor, nor inventor.

    This of course weakens your position and strengthens mine.

  • [Avatar for MaxDrei]
    May 10, 2020 05:09 am

    As to your scrivening point, there is an difference between copyright and patent protection. Copyright protects the skill and labour that is expended in reducing the concept to material form. By contrast, the inventive concept, as such, is that which a patent protects. Skill and labour are still very much needed, many years of training and experience, accurately to reduce that ethereal concept to the material form of a patent claim that corresponds to that inventive concept. But the patent is not there to protect that skill and labour. The patent practitioner (agent or attorney) who drafts the claim, to present the invention in material form, is (except when is is the attorney who made the invention) as you say, a mere scrivener, and not the inventor or the owner of the invention.

  • [Avatar for Anon]
    May 9, 2020 07:12 pm

    I will be looking forwards to the appeal briefing. I earnestly hope that several tidbits from the musings put forward here (and elsewhere) are understood and put to the Court (including the notion that the musings below do provide for judicial recognition of juristic (non-real) persons).

    As you say, we shall see.

  • [Avatar for Anon]
    May 9, 2020 07:07 pm

    … the subject matter of the application, as expressed in its independent claim.”

    I am not sure if you mean to stress what you are stressing, but the scrivener who actually expresses the subject matter of the application in its independent claim is TYPICALLY a patent attorney — who is NEITHER devisor, nor inventor.

    I probably should have noted this the first time you indicated this path, but I was tracking the MISS from ted along a similar pathway.

  • [Avatar for Anon]
    May 9, 2020 06:07 pm

    the one who conceived the definition of the invention

    You are both not paying attention AND trying to make things up in order to fit.

    Read again my admonition at post 14. You are doing exactly the wrong thing.

  • [Avatar for MaxDrei]
    May 9, 2020 04:16 pm

    Anon, here is a Link to the most recent document on the UK IPO file:

    which confirms that Thaler has appealed the PTO Decision to the court. It remains to be seen what the court will decide, if the appeal gets that far.

    The court in England is stuck with the statutory definition in the UK Statute, that it is the “deviser” that is to be named as inventor. I don’t know of any other jurisdiction which is stuck with that definition. You say that the conceiver and the inventor need not be the same person. To the contrary, I think the person to be named as inventor is the one who conceived the definition of the invention, the subject matter of the application, as expressed in its independent claim.

    BTW Robert Jehan, who wrote the Linked article, is Thaler’s attorney, both at the UK IPO and at the EPO. He has an agenda, namely, his client’s.

    You point out that machines already write the specifications of patent applications. I can see that a machine might be able to write a detailed written description of a specific embodiment. Whether machines have advanced to the stage of accepting as input a statement of a technical problem to be solved, and then as output printing out claim 1, i doubt. But if ever they get to that stage it will be hard to say that anything other than the machine is the inventor.

    And, Observer, what you write makes to me a lot of sense.

  • [Avatar for Anon]
    May 9, 2020 02:02 pm

    Based on this precedent, I would expect Thaler to file patents on those inventions in his own name as inventor.

    I don’t think that you are paying attention to what Thaler is actually saying.

    Sometimes, angling for the “practical” overlooks the content.

    Inventor and conceiver are not the same legal concept. You too should take the ‘black box’ thought experiment. It is NOT a given in any future type of litigation that this denying of patents based on AI as not being the inventor must mean that there is a ‘ready’ human that legally could be deemed inventor.

    You see “inoculation’ where there is none.

  • [Avatar for Practical Observer]
    Practical Observer
    May 9, 2020 12:23 pm

    *EDIT to above at 35: “. . . it removes one obvious argument to their invalidity” (facepalm)

  • [Avatar for Practical Observer]
    Practical Observer
    May 9, 2020 12:19 pm

    This is a smart move by Thaler because it forces the various patent granting authorities to go on record with interpretations of their statutes/rules/regs to thereby inoculate *other* DABUS inventions against attack in the future. It would probably be worthwhile to try to get judicial precedent as well.

    Think about this from a practical standpoint. The two DABUS inventions currently at issue (a food container and a light blinking pattern) are interesting, but probably aren’t wildly commercially valuable. So if the governments say “no patent and you’re barred from re-filing in your own name” then it’s no big deal because they aren’t that valuable to begin with. But does anyone think these are the only two areas in which DABUS can generate inventions?

    I would bet DABUS has (or soon will) provide inventions that are much more commercially valuable. Based on this precedent, I would expect Thaler to file patents on those inventions in his own name as inventor. Think 10-15 years down the road to any potential eventual litigation involving those patents – what would the defense try to discover? Any facts that pertain to potential invalidity. What’s one example? That the listed inventor isn’t the true inventor. The defense would depose Thaler and the truth of DABUS would come out and they would try to argue the patent is invalid because Thaler wasn’t the true inventor. But now with this precedent in hand Thaler (or the patent owner at that time) can point and say that he is indeed the inventor because DABUS cannot be an inventor per the 2020 interpretation by the USPTO/EPO/UKIPO. This inoculation will make the patents covering these other DABUS inventions more valuable to potential purchasers/licensees because it removes one obvious defense to their validity.

    It will be interesting to follow up in 18-24 months (after the nonpublication period expires) to see what patent applications, if any, list Thaler as an inventor. If the above theory holds, I would expect to see numerous applications listing Thaler as an inventor and probably in a variety of different subject matter fields.

  • [Avatar for Anon]
    May 9, 2020 10:36 am

    MaxDrei @ 32 – thank you for the link.

    I suggest that you read the article again.

    Now read it again.

    Now read it yet again.

    In each of these reads, I suggest that you read with a mind willing to learn (as opposed to a mind stuck with your own ‘I’ve been around this block” mindset.

    Then read my post at 14 above, and notice how your post at 32 falls exactly to what should NOT be done.

  • [Avatar for Anon]
    May 9, 2020 09:38 am

    name the person who “conceived” the notion that the output of the AI machine is an indication of a potentially patentable “invention”.


    Walk through the thought experiment with the black boxes.

    You need to account for this in order to maintain your view.

    (this is one reason why I have been adamant on this point)

    Unless and until this is gone through, any (and I do mean any) insistence otherwise is a non-starter.

    and by the way, there ARE non-AI programs that can write applications already out there.

  • [Avatar for MaxDrei]
    May 9, 2020 08:22 am

    Have a read of the linked article. I say that the UK Patents Act definition of “inventor” (as the “actual deviser”) is beginning to look flawed. Rather, you have to name the person who “conceived” the notion that the output of the AI machine is an indication of a potentially patentable “invention”.

    Alternatively, devising an embodiment is not the same as devising the invention which was the subject of the patent application. Who conceived, visualised, reduced to words, drafted claim 1, the definition of “the invention”? Was it Dabus, I doubt it.

  • [Avatar for Anon]
    May 9, 2020 06:59 am

    … and maybe we can have ted re-engage .

    ted, what are your answers to your own two part question that I answered at post 15 above?

    Certainly, you did not mean those questions to be rhetorical, did you?

  • [Avatar for Anon]
    May 9, 2020 06:54 am

    I would also state that “first instance” and “first governing body” are different things.

    The first instance is the front line decision in the first governing body. Review of that first instance — even if that review is in the same body — is no longer a “first instance.”

    My error is thinking the review was in a second body is a small error in comparison to thinking that “first instance” applies to a reviewing of the actual first instance.

  • [Avatar for Anon]
    May 9, 2020 06:48 am

    I quite agree (and have all along) that much of what Thaler appears to be doing is as much marketing and theatrics as it is legal advancement.

    But there still be legal advancement.
    There still be the real notion that the devisor is NOT a real person.

    Ponder that single thought and do not be so fast to dismiss it.

    Hold that thought in mind and walk through the multi-black box example I have put forth that shows why the view that conception merely occurs to some other person who happens along to read what is in the black box cannot be correct.

  • [Avatar for MaxDrei]
    May 9, 2020 04:44 am

    Also, at the EPO, the decisions thus far are by the the first instance, the “Examining Division”.

    It remains to be seen what happens in any appeal to the EPO’s Technical Board of Appeal, that happens to be filed and prosecuted.

  • [Avatar for MaxDrei]
    May 9, 2020 04:36 am

    Thaler could appeal to the Court the decision in the Office. But if we suppose that in naming Dabus as inventor his motive all along is to gain attention for his company, is the expense of an appeal something he would even contemplate?

    The UL PTO Register access is down so I can’t check today, to see if the appeal period has already expired, without any appeal having been filed. That would be my guess though.

  • [Avatar for Anon]
    May 8, 2020 06:39 pm

    Ah, MaxDrei, I see that I had missed the distinction between a regular court and the reviewing body that backed up the initial decision.

    This has not made its way into the US Sovereign ‘equivalent’ (?) of an Article III court.

    Is that an option? Or is the reviewing body the last step in U.K. patent law for this matter?

  • [Avatar for MaxDrei]
    May 8, 2020 08:30 am

    This is the Link in the Blog post above, to the UK Patent Office Decision of December 2019. Its Para 15 includes the words:

    “To be clear, the Office accepts that DABUS created the inventions set out in the two applications”

    UK law defines the inventor as the devisor. The words “create” and “conceive” are not synonyms. The machine creates and the human who peruses the machine output perceives in it an “invention”. Harder though, to see a difference in meaning between “create” and “devise”.

    What UK court Decision are you referring to? I know of none.

  • [Avatar for Anon]
    May 8, 2020 07:08 am

    You remember incorrectly MaxDrie, as the U.K. decision is from the U.K. courts.

  • [Avatar for MaxDrei]
    May 8, 2020 04:45 am

    If I remember right, anon, the UK Decision is one made by a PTO Examiner. Is there anything yet from the UK courts?

  • [Avatar for Anon]
    May 7, 2020 03:49 pm

    U.K. has already agreed with me, MaxDrei, as I have repeatedly try to get you to notice.

    And let me repeat the warning from earlier: philosophy is not law, so you ARE banking on the wrong thought pattern to affirm the legal position here — it is NOT whether or not you agree with me. It is whether or not you agree with the law.

  • [Avatar for MaxDrei]
    May 7, 2020 01:18 pm

    Thanks, anon, for the link, which I enjoyed as a vindication of my suggestion that data processing rises to the level of “thinking” only when the “thinker” experiences doubt.

    Machines don’t do “doubt”. If they don’t doubt, well then they don’t “think”. And if they are incapable of thought, ergo, they cannot “conceive” an invention.

    You don’t agree. That’s OK. It remains to be seen, of course. whether the courts, in this developing area of jurisprudence, arrive at an agreement with me or with you.

  • [Avatar for Anon]
    May 7, 2020 09:28 am

    Some good points — and I will have to come back with more comments.

    In the meantime, ted, you are slipping back to colloquial, and the attempt at law (MPEP) is notoriously off.

    The bottom line is that the legal definition of ‘conceive’ does NOT require human mind. Note that I have already taken your first attempt and shown the ‘or’ alternative meaning may well apply.

    Second, you appear to err in thinking that the fact that some machines may well be merely “used by” someone else and that someone else is still the conceiver, as a universal “must,” and this is simply not true.

    MaxDrei, a small youtube philosophy video on Descartes that you might enjoy:

    Both you and ted may want to watch this — it places my obstinacy in a perspective that may help you both

  • [Avatar for ted]
    May 7, 2020 06:21 am

    @Anon, I simply do not understand why you are defining conception in a way that it has never actually been interpreted by a court. What source of actual law explicitly articulated your NOTION of what conception is?

    For instance, let’s look at 2138.04 of the MPEP: conception is “the complete performance of the mental part of the inventive act” and it is “the formation in the mind of the inventor of a definite and permanent idea of the complete and operative invention as it is thereafter to be applied in practice….”

    From that definition, if I complete the mental part of building the machine, and then interpret that the output is patentable for a real-world purpose, then under that definition of conception, it seems like I may be the inventor.

    Now, I agree with your point that this is a somewhat unsatisfying result based on our logical notion of conception. But it also seems silly to claim that everyone else is a moron about patent law, while requiring people to accept your conclusory argument, when even the USPTO left open the possibility that humans can be the inventor in that case.

  • [Avatar for MaxDrei]
    May 7, 2020 04:11 am

    Good point from anon about the consequences of naming as inventor a person who was not the inventor. But what’s that Latin expression to the effect “No harm, no remedy”? Under the 1973-minted European Patent Convention, for example, naming the wrong inventor can render the patent unenforceable, but only if the wronged true inventor initiates and successfully carries through the prescribed legal actions to enforce their rights. I await with interest the first time that happens, with a Dabus-type inventor.

    Third parties might fume, but if ownership of the invention is beyond dispute there is nothing they can do about it no matter how uncertain it is to ascribe inventorship.

    Where’s the harm, in that set-up? Rightful ownership is all that matters, no?

  • [Avatar for Anon]
    May 6, 2020 04:18 pm

    Mr. Lewis,

    I recommend that you read up on the Dabus case. Falsely naming a human who was not the actual conceiver is Fraud on the Office. Those figuring out what inputs, as well the authors of the program are NOT the conceivers.

    You err on both facts and law, and your suggestions are ill-advised.

    ANY attorney processing AI inventions best be careful. One may have not only furthering fraud issues, but malpractice issues as well.

  • [Avatar for David Lewis]
    David Lewis
    May 6, 2020 01:51 pm

    It seems to me that the owner of the invention could have figured out a human to name as the inventor – the one that figured out what input to give the AI program to arrive at the invention. Possible, the authors of the programs would also be inventors.

  • [Avatar for Anon]
    May 6, 2020 04:32 am

    So I ask, what good comes from computers receiving inventorship, and what damage does it do to progress if they can’t

    As I first noted, this question is not one of “what is the law?”.

    Second, I will note that this is a compound question — the answers to which may not converge nicely for any particular mindset.

    Third, the first part of my answer is that I have not come to a decision on whether or not it IS good or bad that AI machines currently cannot receive legal inventorship. If you have been around the patent blogosphere for a while, you may have picked up that I am NOT a fan of juristic persons. To me, juristic persons have gained a level of power that exceeds (far exceeds) any sense of balance of power and risk – or responsibility. To the degree that AI is not a real person, I place them in the (possible) category of juristic person. In law, you have real persons, juristic persons, or NOT a person. Law generally is aimed at the betterment of real persons. Law modified to account for juristic persons, and as noted, I feel that it goes off the tracks when the betterment of juristic persons is outsized in comparison to the betterment of real persons. Thus, I am internally conflated, as I am a true fan of innovation and a strong believer in a strong patent system to foster that innovation, and I do believe that the Supreme Court (perhaps surprisingly) ‘got it right’ in Roche v. Stanford.

    Fourth, the second part of my answer also flows from my personal beliefs in the value of innovation and strong beliefs in a strong patent system to foster innovation. To a certain degree, ANY instance of depriving the value and protection of innovation is anathema. As I indicated with a different matter above, if you have been around the patent blogosphere for awhile, my views are pretty well known. Another of my views has to do with the foundational aspect of the particular patent law system that the US Sovereign chose to create. From its onset, the US patent system was geared to be a protection system of fully alienable property. What works about this is that freedom of invention need not be locked down to ability to put invention into practice. As an avid student of innovation, this particular feature is amazingly powerful, and i jealously guard it. This is why I do NOT ascribe to making substantive patent power to be dependent on the ‘who’ of who owns the property (as many of the ‘small inventor’ readers of this blog are aware). I share this because this driver shapes my answer that ANY detriment to innovation should be taken seriously, and the fact that putative inventions may lack a legal mechanism of obtaining legal protection of the invention is deeply troubling to me. Of what harm? The direct answer is the harm that lacking protection has universally. We have a patent system for a reason (and, as those that created our legal system put it: the benefits of having a patent system practically go without saying), and our nation ascended into its position DUE in no small part TO having a patent system. Innovation will naturally flow to where it is best protected. We (the Royal We) need to make sure that our Sovereign IS that place.

  • [Avatar for Anon]
    May 6, 2020 04:03 am


    You still miss, even as you appear to have had a type of ‘A Ha’ moment.

    All along I have maintained that the state of US Sovereign law does not permit an AI machine to be deemed an inventor.

    That has never been ‘an error’ in my reasoning.

    I have repeatedly stressed that there must be understanding between the differences between the legal terms of ‘inventor’ and ‘conceiver.’

    Because of those differences, AI may be a conceiver, and NOT an inventor. In such cases, you run into the ‘problem’ that there may well be a putative ‘invention’ without there being a LEGAL inventor.

    Some have bristled at this and demanded that ‘law must be different.’

    And THAT would be a mistake, as the law is what the law is (the merely desired consistency may well NOT be there).

    But this is mistake NOT of my explaining of what the law IS. It is a mistake of those wanting the law to be something that it is not.

    And because of this, I simply need not ‘win over’ ANY crowds. I have already ‘won’ by simply being correct about the law. I did not write the law, but I am of critical enough mind to understand the law — as it is.

  • [Avatar for ted]
    May 5, 2020 08:10 pm


    I’ve genuinely appreciated the discussion. You make compelling arguments that a machine may stumble upon an “invention” without direct supervision of the machine’s inventor. However, as a matter of fact, constructing the machine is technically contribution to the conception of the invention. (MPEP 2137.01.II).

    However, even if we accept that the human did not invent, that still does not mean the machine can be an inventor. I have no doubts that a machine could meticulously determine the novelty of the invention it happened to amble upon, but I’m also certain that it could not appreciate the real-world utility of the invention. (35 U.S.C. 101).

    Like many times happens in patent law, it appears that we disagree upon what some words mean. The USPTO chose to interpret the law in a way that is not cogent with your point of view. However, it does not meant that the law cannot change when it hears a compelling argument.

    Best of luck on your journey to liberate the robots.

  • [Avatar for Anon]
    May 5, 2020 05:49 pm

    MaxDrei, aside from the fact that you are attempting to treat philosophy as law (always a dubious assertion, even when being direct about it), you have — again made an error in viewing whether the legal point of conception is necessarily one of human thought.

    Yet again — please get the facts right.

    It is unhelpful for you to go on these tangents of colloquialism.

  • [Avatar for Anon]
    May 5, 2020 02:40 pm


    If one goes by the legal definition, then YES, an AI machine can possibly provide a suggestion as to a possible course of action (the alternative “or” provides the condition is met when one aspect OR the other is met).

    I did warn you that you have entered a LEGAL discussion.

    If you want to argue, then you need to recognize the forum of the argument and what it takes to carry a legal point.

    You seem more intent on having a colloquial conversation — replete with the pre-determined notion of conception cannot be done by a machine.

    You lose this point (legally), and thus will not be able to get out of your own way.

    Further, note that I answered your question as to nuance. Instead of then reciprocating and following along with the thought experiment that I provided, you instead have charged ahead outside of the structures of this being a legal discussion.

    On top of that, you display an ignorance of just what an AI machine is about. Your “just run” may be applicable to a standard machine or a standard computer. An AI machine is not a standard machine or standard computer. That is a critical point for you to understand and accept for this legal discussion.

    That you view “reading” as “conceiving” tells me that you are not capable of understanding the nuance provided. Again — these are legal terms. Legally, merely reading is NOT conceiving. That you think otherwise just shows that you do not understand enough about patent law to even understand why that statement cannot be true. There is proof of something alright — just not the proof that you want to jump to.

    As to “they put it there in the first place” — putting the box has nothing to do with the legal point, and the thought experiment can be easily adapted to have a machine do that. You offer a non-sequitur that is simply easily dismissed.

    Finally, your comment on the Constitution is entirely unclear. I doubt that you understand how the Constitution impacts the point at hand. The legal point of denying a patent to an AI machine IS the current state, and it is that current state entirely outside of the discussion of whether or not an AI machine is the conceiver.

    As I have noted, this LEGAL state is because the legal term of “inventor” is not the same as the legal term of “conceiver.” The US Supreme Court has already deliberated on whether or not a real person must be the legal inventor. That question has been decided (see Roche v. Stanford). Your proclamation is thus already violated.

    You ask yet another question — even as you have NOT answered mine: “ So I ask, what good comes from computers receiving inventorship, and what damage does it do to progress if they can’t (I would appreciate a single contiguous answer,…

    If I give you an answer, will you then give me the courtesy that you have not given me yet?

    (by the way, that answer that you seek is NOT a question of what the law IS — you do understand the difference, right?)

  • [Avatar for maxdrei]
    May 5, 2020 02:31 pm

    Dubio ergo cogito. Cogito ergo sum. Everybody knows the second sentence. But the first is just as important. I doubt therefore I think. Computers don’t do doubt so they don’t think.

  • [Avatar for ted]
    May 5, 2020 02:01 pm

    Notice the word “idea” here. Is a computer capable of having an “idea”? An idea is “a thought or suggestion as to a possible course of action.” Two parts of the definition of “idea” seem to quash any argument that a computer can conceive.

    First of all a “thought,” is an idea produced by thinking. Computers fundamentally cannot think, because thinking requires a mind. A mind can only exist inside an individual. A mind contains “opinions” and “beliefs” that computers are incapable of possessing.

    Second, an “idea” is fundamentally about a “possible course of action.” Computers don’t have possible courses of action. They just run. They are inanimate objects that receive inputs and produce outputs.

    Not only does this point prove that computers cannot conceive, but it also proves that what you are calling “reading” is … conceiving. Humans are not just “receiving the box,” they put it there in the first place. The fact that a computer did work in the middle is no different than the way chemical compounds have worked for many, many, years.

    Finally, and this is an important one, the Constitution says that “[t]o promote the progress of science and useful arts.” This means that patent laws could be interpreted in a way that doesn’t promote progress, we can avoid that interpretation, or even explicitly change the law.

    So I ask, what good comes from computers receiving inventorship, and what damage does it do to progress if they can’t (I would appreciate a single contiguous answer, I could not find all of the links and thread that you sent me to last time). If the argument for DABUS were more clear and persuasive, maybe the PTO (and the rest of the world) would have taken your side.

  • [Avatar for Pro Say]
    Pro Say
    May 5, 2020 09:53 am

    No patent for DABUS?

    Robby the Robot weeped:

  • [Avatar for Anon]
    May 5, 2020 07:47 am


    The nuance is the difference between invention and a sub-part of invention: conception.

    And yes, the fact of the matter of “working outside the scope” is enough to break the legal thread between the two.

    You want to ignore the legal consequences here — and I am not sure that you even understand what legal points are in play.

    You do NOT get to jump to an “At the end of the day” statement, because in law (especially in law), the Ends do not justify the Means, and it is critically important just HOW the Ends are reached. The Means very much matter.

    A real person (or even team of real people) may have invented an AI machine that can conceive.

    But that is what they invented: the direct AI machine.

    They did not invent whatever disconnected and out of scope material that that AI machine itself conceives.

    More details on those points have already been shared on this blog and over at Patently-O.

    There is NO disrespect to software developers in my comments or in my position. I have full respect for such, and believe fully that such inventors should be rewarded strong patent protection for what they themselves invent (as that legal term encompasses).

    But that legal term requires — as a nuanced lower level portion of what it means to invent — that the conception itself be BY the software worker.

    AI has advanced past that point.

    In the US Sovereign, there are OTHER sticking points though, beyond the lower level aspect of conception, that prevent AI from being an inventor.

    What you are faced with, in the legal sense and under the US Sovereign legal setup, is that AI innovation cannot be patented because a mere reader of the result of AI conception did not themselves conceive (as that term is understood in its legal sense) of the innovation, they merely read the result of the conception.

    If you are up for it (and goodness knows, even attorneys who profess to have taught AI classes have shown an unwillingness to engage), walk through the following thought experiment (keeping in mind that the terms are legal terms, and need to be understood in that context):

    As found at:

    Specifically, for that type of result that is out of scope and as such, is not what is merely programmed or traceably directed by the programmer, what you end up with is ANYONE (and not just limited to a programmer, the AI machine owner, or anyone even remotely intersecting with the output) being ‘charged with’ being the inventor when that person happens to be the person reading the output.

    People have attempted to coopt the different word of ‘recognition’ as being the legal word of ‘conception.’

    So at the link (which unfortunately has a link of its own, as the person that that discussion was aimed at refused to engage and was being chased around), there is a relatively straight forward thought experiment set up to show the legal difference between one who merely reads an output and understands the output and one who conceives and creates the output that another reads.

    From the link:

    I have deconstructed your mere human reading the output of AI’s conceptions several times now – and like EVERY all element of this discussion, you have NOT engaged ANY of those counter points put to you (across SEVERAL THREADS).

    So let’s modify the point put to you a little bit more and see if you will engage ON POINT or merely run away (again) after claiming that you have engaged on point….

    In addition to post 1.3 [embedded link omitted] I had added at post an additional ‘mere reader’ conundrum for you:

    As to ‘first time’ – see above at 1.3 (this is now the second time that you are skipping over counter points provided to you). You may want to slow down and contemplate those points before you insist on maintaining your pre-conceived position.

    A further note on that point: mere recognition of someone else’s work is NOT viewed as invention OF that work. You have a serious logistical problem easily shown by a simple example.

    You would have ‘invention’ be by a person sitting in a room when that person merely opens a box shipped to him and reads something in the box.

    Note that the antecedent of what that item in the box is, and how that was developed is left out.

    To your view, THAT is not material.

    But it IS material. Visit again the posts and the items that I have presented to you that you have skipped over.

    You have moved from an assumption that the builder or owner is the first one (merely) reading the output of the AI machine and thus necessarily being the inventor, to eliminating the requirement of conception AT THE TIME that the idea is formed to some later time of merely reading that formed idea (formed prior to the box being shipped and opened by whomever happens to be receiving the box.

    Your assertion of “The key point here is that according to the legal definition a machine cannot invent but only a human brain.” is the conclusion that you are presuming. This entirely misses the underlying point that NO HUMAN has done the actual conception performed by the AI.

    What happens in our little thought experiment if we exchange the single person sitting in a room opening a box and the entire world sitting at a computer screen reading the contents of the box being presented en masse to everyone? Does your attempted logic make everyone the inventor?

    Now if you want to go down a different path of Discovery, well, we can do that. But YOU need to realize that such is indeed a different path.

    … [this is where I make the thought experiment a little more interesting]

    To this (unanswered) point, let’s add three additional rooms. In one of these additional rooms, there is a separate REAL person doing the conceiving, and writing that conception down, let’s say in a patent application format. That patent application is put into a black box. In the other additional room, there is an AI doing the conceiving, and with automatic transcription (of which, this capability exists today), that conception is ALSO transcribed into a patent application format. This other patent application is put into a separate black box. Each of these two black boxes are put into the third room. In that third room, there is a randomizer that takes all the black boxes given to it, and completely randomly selects a single black box for presentation to the aforementioned gentlemen – who for argument’s sake let say has nothing to do with building or owning the AI machine that was in the second room.

    Now let’s return to YOUR supposition that a real person opening that black box and merely reading what is in the box MUST be deemed the inventor.

    Do you see the logical problem?

    [chasing comment omitted]

    And then let’s return to my previous counter point which you have not answered, and replace that nondescript person merely reading what is in the black box with millions of people simultaneously. Are EACH of these millions of people who are merely recognizing what has been captured on paper (randomly and may or may not be the AI captured item) “inventors?”

    ted, do you understand the nuance now?

  • [Avatar for ted]
    May 5, 2020 12:37 am

    @Anon, what is this nuanced issue you are referring to… humans go to work and invent things. Computers are created by humans. They are machines that hold states in the form of 1s and 0s. Yes, in the era of cloud computing and distributed computing, maybe a piece of software can work outside the scope of a specific model that it is training against, but at the end of the day… it was programmed by a human.

    This argument seems disrespectful to software developers more than anything, and I just can’t fathom a world where we award a letters patent to a d*** computer instead of the human workers that toiled to fix the bugs and clear the critical design hurdles to facilitate whatever the computer ended up doing.

    So please, explain what you mean by nuance.

  • [Avatar for Anon]
    May 4, 2020 07:48 pm


    I deal with AI that is not constrained to mere trained models, so, no, your statement is not accurate.

    Further, the discussion is NOT directly about inventorship, as that legal term is to be used.

    The underlying issue here is far more nuanced.

  • [Avatar for Anon]
    May 4, 2020 07:45 pm


    You present an error of fact:

    Dabus, under the control and instruction of Thaler, performed a machine computation and outputted a statement.

    The explicit point here was that the factual act of conception was NOT at the instruction of Thaler (in a traceable manner).

    It is critical to the legal discussion that you do not get these facts wrong.

  • [Avatar for ted]
    May 4, 2020 05:20 pm

    Difficulties in determining who actually contributed the changes in the code base that led to conception of the invention is not an excuse to start lazily calling software/hardware (fundamentally a state machine) an inventor.

    C’mon people, AI is just a trained model, right? What am I missing here.

  • [Avatar for MaxDrei]
    May 4, 2020 04:33 pm

    In Europe, what counts is ownership of the subject matter for which patent protection is sought. Inventors have to be named so that the ownership issue can be nailed down.

    Here, Dabus was not able to execute an instrument of assignment in favour of Dr Thaler. Thaler signed the USPTO form on behalf of Dabus.

    Dabus, under the control and instruction of Thaler, performed a machine computation and outputted a statement. It was Thaler who conceived, from that statement, the invention which was then made the subject matter of a patent application. Dabus lacked any capability to appreciate that its output constituted a patentable invention but Thaler did. He was the first and only “inventor” of that subject matter, under the Act.

  • [Avatar for Anon]
    May 4, 2020 12:45 pm

    The weakest part of the EPO reasoning was: “legal personality is assigned to a natural person as a consequence of being human or to a legal fiction as a result of being created by legislation or jurisprudence.

    It should be evident that ‘created by jurisprudence’ COULD HAVE BEEN DONE in that immediate decision, and that ‘legal personality’ — as it exists OUTSIDE of real persons — may easily be seen as a path to recognizing AI (past the UK stage of devisor).

    I find it rather interesting that the usual patent blog pundits so keen on UK and EPO jurisprudence have – to date – steered clear of this ‘AI as devisor’ subject.

Varsity Sponsors

IPWatchdog Events

Patent Portfolio Management Masters™ 2024
June 24 @ 1:00 pm - June 26 @ 2:00 pm EDT
Webinar – Sponsored by LexisNexis
August 22 @ 12:00 pm - 1:00 pm EDT
Women’s IP Forum
August 26 @ 11:00 am - August 27 @ 5:00 pm EDT
IP Solutions, Services & Platforms Expo
September 9 @ 1:00 pm - September 10 @ 2:00 pm EDT

From IPWatchdog