Dear USPTO: Patents for Inventions by AI Must Be Allowed

“35 U.S.C. §103 provides that patentability ‘shall not be negated by the manner in which the invention was made.’ Thus, a conception step, as discussed in depth by the USPTO decision in the ‘350 application, should not be a prerequisite for patent protection.”

AI - July 29, 2019, U.S. patent application serial number 16/524,350 was filed with the U.S. Patent and Trademark Office (USPTO), with the sole inventor identified as artificial intelligence named “DABUS.” The assignee, Stephen Thaler, acknowledged that the invention was made by the creativity machine, without any human input. The USPTO issued a notice of missing parts, requiring identification of each inventor by name. Thaler then petitioned that the missing parts notice be vacated, which was denied, and Thaler petitioned for reconsideration.

The USPTO issued its decision on the reconsideration petition on April 29, 2020, again denying the petition and concluding that the patent laws required a natural person as an inventor. The decision asserts that conception is the touchstone of invention, as the formation in the mind of the inventor of a definite and permanent idea of the complete and operative invention as it is to be applied in practice. However, this decision is too narrowly focused, and should be reversed by the U.S. Court of Appeals for the Federal Circuit if Thaler appeals the USPTO ruling.

AI is Science

The entire U.S. patent system is premised on the objective of promoting science and the useful arts, as stated in the Constitution, Section 8, Clause 8. Clearly, artificial intelligence is science. Therefore, AI is encompassed by the Constitution, and should be promoted accordingly.

The USPTO requirement for a human inventor overlooks the statutory language that patents are not limited to inventions. As stated in 35 U.S.C §101, patents can also be granted for discovery of a new and useful process, machine, manufacture or composition of matter. The ‘350 patent application is based on the discovery by DABUS of a novel device and method for attracting enhanced attention, such as signal indicators and beacons (typically based upon color, brightness, periodic flashing frequency, rotational pattern, and motion) using a fractal dimension.

The USPTO decision on the ‘350 application focuses largely on conception as a necessary mental step in the inventive process. The USPTO’s citation of the 2013 Federal Circuit decision Univ. of Utah v.Max-Planck-Gesellschafl zur Forderung der Wissenschaflen e. V (734 F.3d 1315) is not conclusive. The Court ruled in that case that the mental step of conception of an invention must be performed by a natural person. This does not answer the situation where an invention or discovery does not involve such a mental step.

Flash of Genius

Some inventions do not derive from conception, but rather may be the result of a “flash of genius,” the proverbial light bulb turning on, or simply by accident without any mental conception step. Famous patented products resulting from accidents include Post It sticky notes (5,194,299 – 1993), the Slinky toy (2,415,012 – 1947), and Play Doh modeling clay (3,167,440 – 1965).

35 U.S.C. §103 provides that patentability “shall not be negated by the manner in which the invention was made.” Thus, a conception step, as discussed in depth by the USPTO decision in the ‘350 application, should not be a prerequisite for patent protection. If no mental step of conception is needed, then a natural person is not necessary for a patentable invention or discovery. Without a human step of conception of the invention, the rationale of the USPTO seems to evaporate, such that AI generated inventions and discoveries become protectable.

Defining ‘Who’

The USPTO also asserts that the language from 35 U.S.C. 101 “Whoever invents or discovers…” suggests a natural person. But asking “who is the inventor” is no different than asking “who is the computer in 2001 A Space Odyssey” or “who are the robots in Star Wars?” Obviously,  HAL, 3CPO and R2D2 are not natural persons. So, concluding that “whoever” in the patent statutes infers a natural person is an over-simplification.

Also, the USPTO’s position is inconsistent with the April 10, 2020 decision of the Federal Circuit, which concluded that the statutory language allowing a “person” to file a petition for post grant review includes banks. See Bozeman Financial, LLC v. Federal Reserve Bank, (Appeal from the United States Patent and Trademark Office, Patent Trial and Appeal Board in No. CBM2017- 00036). This statutory interpretation that “persons” is not limited to natural persons seems to undercut the USPTO interpretation that an inventor must be a natural person.

The Chakrabarty Analogy

A similar issue arose 40 years ago with respect to the scope of patentable subject matter under Section 101. Prior to 1980, living things were generally understood not to be subject to patent protection. Then the U.S. Supreme Court ruled in Diamond v. Chakrabarty, 447 U.S. 303, that a live, human-made organism was patentable subject matter under §101. Chief Justice Warren Burger noted that Congress contemplated that the patent laws should be given wide scope. He also cited the 1933 Supreme Court decision United States v. Dubilier Condenser Corp, 289 U.S. 178, which cautioned against reading limitations and conditions into the patent laws which Congress had not expressed.

As is clear from Chakrabarty, patents protect various subject matter that was never contemplated when the Constitution was written or the patent statutes were enacted. Science evolves to new technologies, many for widespread benefit of mankind. Likewise, inventions and discoveries by artificial intelligence have and will continue to evolve, too, with solutions to long-felt needs and improvements to existing technologies that cannot be contemplated today. These inventions should not be precluded from patent protection merely because a person did not conceive of the idea. Otherwise, the lack of patent protection will not promote the progress of the science of artificial intelligence.

Image Source: Deposit Photos
Copyright: Jirsak
Image ID: 144709683 


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

36 comments so far.

  • [Avatar for Anon]
    June 1, 2020 12:09 pm

    First, Mr. Hill, thank you for your response. I was actually about to type a remark that this was yet another thread for me to bookmark because my presented counter points have not been answered.

    While they still have not been answered, at least you provided an attempt.

    The attempt fails – for reason already given and reminders to follow.

    You begin your position with the statement of “ this entire issue is resolved by the statutory requirement of personhood for an inventor. This ends the matter of whether AI can be listed as an “inventor,” .

    This is NOT the issue that I am presenting.

    As you should be aware from my posts, I am in complete agreement that in the US Sovereign, the statutory requirement of personhood for an inventor IS the Rule of Law.

    As I pointed out though, this stems from the (and reflects the holding therein) Roche v. Stanford case of the US Supreme Court.

    I completely agree that the US courts and the USPTO are indeed bound by this current Rule of Law.

    That being said (and disposing of your first paragraph), we move to your error of JUST HOW the question “must be” reframed.

    I also address this error by presenting and providing a link to an historical analog: that of the THEN legally viewed “less than a real person” inventions by slaves, and how the question was similarly phrase such that” “Should a person be permitted to patent an [invention or output of an invention of a SLAVE] OF the [replacement] patent applicant.

    This contrasts with (sharply) with your rephrasing – a rephrasing that I HAVE ALREADY DEBUNKED. Please see the link and the ‘black box’ scenarios in which it simply does not matter any later ‘who’ that can – or cannot – explain with particularity the concept that justifies the patent.

    Your “must” is BOTH a logical and legal non-sequitur AND fails at its task to set some REAL PERSON as the source of conception.

    FURTHER, you continue to propagate another error that I have debunked: that conception MUST be in the human mind.

    It simply is not so.

    The case law that you may wish to cite (here, Stern) is ONLY dicta on the point immediately at hand as to conception, because THERE, the question did not distinguish NON-human conception.

    You say “by definition” but as I have already put it: this is a fallacy, and a fallacy directly at heart of the present issue.

    The plain fact of the matter is that conception HAS BEEN CLEAVED.

    You cannot address the issue if you are attempting to have a factual reality in which ‘you define’ an impossibility as the controlling legal view.

    You necessarily FAIL in the very construction of your premise.

    Remember: the slave analogy does NOT resolve itself with the attempt to put the NOW realization that a slave has a human mind. You broach the analogy, but fail to carry it through.

    Your bald assertion of ‘The AI cannot identify a concept’ is patently false in the present reality, as this IS what IS occurring.

    Even Mark Summerfield surrenders that point.

    You thus turn and create a tautology that EVADES the actual point hand.

    Your statement of “ The patentable subject matter is only conceived when the concept is identified by a mind.” is false – as shown by my black box scenarios. The mere ability to articulate is a FAIL and cannot be what you want it to be.

    Views on the Singularity – while interesting – are completely apart from the legal issue at hand.

    We can enjoy a separate discussion thereon, but for now, let’s focus on the present non-singularity. I invite you to explore the link provided and walk through the Black Box scenarios.

    I have to wonder if you realize even your own statement here of “And the law should not change because if we allow an AI to be listed as an inventor we have either granted it legal status as a person or we have granted its owner the right to obtain patents for subject matter that is not the product of any mind.” The PRESENT issue is that the ‘subject matter that is not the product of any HUMAN mind’ IS the subject matter IN the Dabus patent applications. Here, your attempt to circumvent my Black Box outcome comes immediately to bite your attempted legal position.

    By the way, I agree with you that there be real danger in allowing patents to subject matter not ACTUALLY CONCEIVED by a human mind. We just happen to disagree on the FACTUAL nature of conception occurring outside of the human mind.

  • [Avatar for Alan R. Hill]
    Alan R. Hill
    May 31, 2020 07:03 pm

    Response to Anon @ 34 re: conception as a legal concept and necessity to establish the right to register new intellectual property.

    First, I must reiterate my position that this entire issue is resolved by the statutory requirement of personhood for an inventor. This ends the matter of whether AI can be listed as an “inventor,” unless the US Congress acts to change Title 35 to allow for nonperson inventors or to extend personhood rights to AI. (I truly hope no Congress would be so foolhardy as to grant legal personhood rights to a machine, no matter how sophisticated it may be – especially if it is too sophisticated to actually understand its processes). If any Court or the USPTO should make a determination such as the one advanced by Mr. Hartung, it would be a decision contrary to law.

    The question must then be reframed to ask: Should a person be permitted to patent an AI or the output of an AI if the patent applicant cannot explain, with particularity, the concept that justifies the patent?

    We already know that the answer to this is no.

    As regards the requirement of conception as a necessary prerequisite for recognition of inventorship or discovery, we must return to the mind of the inventor. Conception as a legal concept is mostly judicially defined. See ex. Stern, 434 F.3d 1375 (Fed. Cir. 2006) (“Conception… is formation in mind of inventor of definite and permanent idea of complete and operative invention…”). By definition, conception cannot be cleaved from inventorship, because they both come from the mind of the inventor. The AI has no mind. It cannot identify a concept. It may arrive at a pattern with results that repeat in a predictable way, but that pattern only becomes a concept when a mind identifies it as such. The person providing the inputs fed into the AI to derive the concept has some claim to inventorship, as does the person who identifies the patentable subject matter in the output.

    Again, this is not at all analogous to making an attempt to patent the invention of a slave because a slave is a person. Even in the most wretched interpretations of slavery, the slave is recognized to have a mind. Antebellum patent law did not allow a person to claim the products of another person’s (even a slave’s) mind as the products of his own. That is not to say credit for an enslaved person’s inventions was never taken by a slaveowner, only that patent law did not sanction it.

    Because AI has no mind, that issue does not apply to the current discussion. The patentable subject matter is only conceived when the concept is identified by a mind. Therefore, we need concern ourselves only with whether the person or persons who identify patentable subject matter that has been provided or performed by the AI are able to articulate the concept in sufficient detail to receive a patent. If not, they should be out of luck.

    I don’t buy into the notion that AI will reach a singularity, but if I am wrong and if AI advances to the point that it spits out unfathomable volumes of patentable subject matter that is beyond human ability to describe and patent, then we should not allow it to be patented. If we reach a point at which a human merely allows an AI to be turned on and the machine begins submitting patent applications that would be approved but for the lack of a human inventor, then those applications must be denied under the current law. And the law should not change because if we allow an AI to be listed as an inventor we have either granted it legal status as a person or we have granted its owner the right to obtain patents for subject matter that is not the product of any mind. By doing so, we would risk allowing some lucky tech corporation to monopolize the most advanced technologies available, perpetually. Even without going down the scifi road, we should be extremely wary of revolutionizing such fundamental legal concepts as “property” and “personhood” merely because we are worried that some tech giant might miss out on a lucrative 20 year monopoly.

  • [Avatar for Anon]
    May 24, 2020 08:43 pm


    The Singularity (and any such belief or lack thereof) has nothing to do with the legal points at hand.

    The “actual state of the world” remains in view of the discussions of the legal points.

    The ad hominem was unfortunate — I do apologize as you do deserve better (many others in the blogosphere do not). However, the POINT of ad hominem coming out does indicate that you are discussing ‘feelings’ as opposed to legal points.

    “Feelings” are out of place. Let’s remove those as well as the ad hominem and focus on the legal points (if you still care to engage on the counter points that I have provided, both here, in the article on slavery and in the link to the earlier discussion in which I provide examples of “non-conception” that MUST fail the legal test of invention (every bit as the example of slavery must fail). Up to you – but I would enjoy a dispassionate view on those legal points (to date, I have seen neither dispassionate, nor passionate [but cogent and on point] responses to these counters that dispel what many would ‘like’ to have.

    As you yourself DO note: :I agree with you that this is NOT a case of ‘merely using’ a machine.

    Once that is accepted, then there be REAL legal issues (right now) that need attention. Invention (and by that I do include the legal term of Discovery, through 35 USC 100), needs the inventor to BE the same as the one providing the conception. When there is a cleaving, one may well end with putative inventions that have no LEGAL inventor. And this IS one of the arguments being made by Thaler (and echoes the statements of some 162 years ago).

  • [Avatar for Mark Summerfield]
    Mark Summerfield
    May 24, 2020 06:18 pm

    Anon, I now understand that you are a believer in ‘The Singularity’. That’s your prerogative, of course, but it does mean that we will never agree on this topic. We are arguing from fundamentally opposed positions with regard to our beliefs about the actual state of the world. Also, your turn to ad hominem attacks tells me that it is time to bow out. I don’t much care what you think of me, but in my experience it’s all downhill from here.

  • [Avatar for David Lewis]
    David Lewis
    May 24, 2020 04:28 pm

    Let’s say someone takes a bunch of random parts puts them in a box, closes the box, puts the box on a machine that shakes the box randomly, stops the machine, opens the box and somehow the parts are not arranged into something that functions as a new more efficient unobvious widget.

    It is my understanding the one that did this is usually an inventor and is entitled to a patent under current law. Although it may be possible to change the details of this scenario create to questions of who is the inventor, we seem to recognize that inventions are often the result of serendipity and that the fortunate people whom experience the serendipity are usually inventors).

    It seems to me that the situation with AI invented inventions is the essentially the same thing. Specifically, the AI requires programing and/or input to invent anything. The humans performing the programing and/or entering the input may have no clue how the AI arrived at the invention, and indeed the AI may have done nothing more than trial, error, and mutation or changing inputs in another to arrive at another trial. However, it seems to me that that is usually no worse than the inventor that arrived at their invention by randomly shaking a box or serendipity, and it seems to me that (despite the US PTO’s suggestion that this invention may have no inventor), identifying those that set the AI in motion in a manner that resulted in the invention as the inventors, may be a better way to have handled this application or at least one that the US PTO may have went along with.

  • [Avatar for Anon]
    May 24, 2020 02:01 pm

    Mark’s post (now at 29) was at 23 earlier this morning. It does appear that there had been a processing logjam. It is to be appreciated that reference’s to earlier posts may now be inaccurate.

  • [Avatar for Anon]
    May 24, 2020 12:33 pm

    Well, Mark, I think your statement provides a vast difference between your outlook and my own.

    And that is not something that I find in your favor, or even something that I would consider to be neutral.

    Instead, I find your view to be overly pretentious and rather way too (self)absorbed with what I can only gather as your view and understanding of the state of the art.

    I did turn to your link (not many have), and that confirms that you have a ‘too-high’ opinion of these matters (for the record, your statement at post 5 misapprehends the view that IBM puts forth that there ARE currently a matter of debate (your view is that there is NO debate, and that IBM agrees with you – I suggest that you read your own source a bit more closely).

    I really do not care (at all) for your portrayal of this as “Quixotic” or a waste of resources. Those decisions simply have NO relation to your feelings.

    As to Anon2 and his sarcasm, meh, anyone can voice their opinions, and Anon2 is not excluded (even as I have already responded to the exact same ‘feelings’ from Anon2 previously).

    I think that there is some (substantial) CONFUSION between the notions of The Singularity and the legal understanding of what it means to have conception. Anon2 self-indulges to the detriment of actually discussing a real point of law that is right now at hand.

    Personally, my view is that humanity will NOT know when The Singularity occurs. Heck, for all I know, it may have already happened. Any and practically most all science fiction portrayals of such thing suffer a serious ‘Hollywood” flaw: humanity has to be made aware of such a thing because that awareness is central to the BATTLE and drama that unfolds.

    Any real Singularity will have run through MOST ALL such permutations in a literal blink of an eye, and would have charted a different course.

    Now — and I do mean RIGHT NOW — is an excellent time to have the legal discussion of the cleaving of ‘conception’ from ‘invention.’

    I refer you to the link that I provided on May 24, 2020 8:49 am.

    Frankly, whether or not Thaler is nothing more than grandstanding, or even if there is merely (and more innocuously) a lack of enablement (and goodness knows – I have MANY clients that want to divulge as little of the ‘secret sauce’ as possible), the very real legal implications are upon us. Now. Not decades from now, but right now.

    Having these discussion now — before they multiply and bloom beyond any ability to develop a conversation is — to me — the epitome of being pragmatic.

    Sticking one’s head in the sand and misapprehending the imminence is the opposite of pragmatic.

  • [Avatar for Mark Summerfield]
    Mark Summerfield
    May 24, 2020 11:04 am


    I agree with you that this is NOT a case of ‘merely using’ a machine. For my part, that is because I simply do not believe it happened, which raises the question of who really invented the subject matter claimed in the two applications (assuming that an invention is disclosed), and the somewhat more ugly question of why those involved are asserting that it was the AI DABUS. Do they really believe this to be true? If so, then what is the basis for this belief? And if not, then what is their motivation for making the claim? Though perhaps Anon2@12 tells us much of what we need to know about Thaler and his disciples.

    I am a utilitarian at heart, although I appreciate that there is some intellectual pleasure to be derived from discussing the hypothetical consequences of AI invention/conception. It might even help to prepare us for the day, perhaps decades from now, when this might become an actual issue. But, then again, it might not, because we cannot foresee how the context and circumstances may differ from what we imagine today.

    Pragmatically, I think that Thaler and the ‘Artificial Inventor Project’ team are wasting patent office resources on their Quixotic provocation. I do not believe that such activities are to be encouraged. The patent system has enough problems already in granting useful and enforceable rights to actual human inventors, without such distractions!

  • [Avatar for Anon]
    May 24, 2020 08:49 am

    Also worth noting:

    (the discussion — there — is STILL not over)

  • [Avatar for Anon]
    May 24, 2020 07:53 am


    Depending on the State, there were varying degrees of a capacity of a slave to own property.

    I thought the article to be of interest as it indicates that there is more than one wrinkle to be attended to in regards to the larger topic of AI machines. I see the “signing oath” aspect to be one of a bureaucrat one and while important in its own right, a far less interesting one than the division of inventorship from its sub-component of “Devisor/conceiver.”

    The parallels of the slave owner to my earlier examples of the person merely being the one sitting at a table and opening a black box being the one “designated” as inventor are uncanny.

    It is also critical to distinguish (even as the likes of you and Mark Summerfield are NOT doing), that this is NOT a case of ‘merely using’ a machine. The crux of the matter here IS the split between ‘owner’ and ‘Devisor/conceiver.’ And that is why I keep circling back to my earlier posts — of which there has YET to be a cognitive rebuttal.

    Also, the slave aspect is an important analogy, even as it is an ugly chapter of our Sovereign, and one would do well to pause in anointing our Supreme Court to be “Supreme” in the sense of Above the Constitution — with no checks and balances — given this article’s inclusion of the Dred Scott decision.

  • [Avatar for Haim Michael Modiano]
    Haim Michael Modiano
    May 24, 2020 06:39 am

    Besides the significant problems of a machine having rights to an invention and assigning the invention, there also could be many unintended consequences of naming machines as inventors.
    In my opinion, we always want to look back at the human idea that began the invention, no matter how “clever” the machine was or however little input it needed.
    Thaler MUST have had some conception of the invention (e.g., what he wanted the invention to be or to do), and then programmed or otherwise employed the machine to do some part of the reduction to practice. Is it possible that the machine conceived the invention—i.e., spit it out without any human prompt whatsoever?! Thaler, not the machine, conceived the invention, even if the machine’s RTP constituted significantly more work; Thaler, therefore, is the inventor.
    If Thaler collaborated with the machine programmer, perhaps they should be named as an inventor, additionally (or alternatively) to Thaler.

  • [Avatar for MaxDrei]
    May 24, 2020 04:46 am

    The Paper which anon links to, at # 19, I found indeed to prompt thought. Good one, anon.

    Already on its first page I read in the Paper that the slave cannot apply for a patent because of its incapacity to swear an oath. I had supposed that it would have been for another reason: that a slave is incapable of owning property.

    I imagine it would be easy to program a machine to sign an oath. But it is up to the legislative branch to decide whether an artificial intelligence can own property.

    But hang on a minute. Is not a solemn oath signed by a machine something of an oxymoron?

  • [Avatar for Alan R Hill]
    Alan R Hill
    May 23, 2020 10:55 pm

    I’m not sure why my previous comment is gone, but I’ll try this a bit differently.

    “The term “inventor” means the individual or, if a joint invention, the individuals collectively who invented or discovered the subject matter of the invention.” 35 USC §100(f)

    “In determining the meaning of any Act of Congress, or of any ruling, regulation, or interpretation of the various administrative bureaus and agencies of the United States, the words “person”, “human being”, “child”, and “individual”, shall include every infant member of the species homo sapiens who is born alive at any stage of development.” 1 USC §?8(a)

    This is very clear and very simple. An inventor cannot be an AI.

    The article about enslaved inventors is interesting but in no way analogous because slaves were legal persons but were not considered citizens. US Law at the time barred non-citizens from holding US Patents. No such citizenship requirement exists now.

    Much of the rest of this discussion, including the original article, conflates the act of designing something novel with “invention.” This misapprehends the nature of the act of invention, but more importantly, it disregards what the law says about for what and to whom a patent may be granted.

    “Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.” 35 USC §?101.

    Firstly, the pronoun “whoever” is not defined in the United States Code, but we all know it is only ever used in reference to persons. Nonpersons are referered to as “whatever.” This again reinforces the idea that an AI cannot be an inventor. But, again more importantly, Sec. 101 also specifically includes a person who “discovers” a useful patentable subject matter as the potential grantee of a patent. Surely, the act of monitoring the output of an AI for patentable subject matter is an act of discovery. What else could it be? Some heretofore unkown useful “thing” has been witnessed, documented, and analyzed by a person for the first time. That thing should be patentable by the discoverer. After all, particular methods and purpose were employed to bring about the new useful “thing”, that the discoverer did not know precisely what it was that would be discovered does not invalidate that reality. Even if that person is to be denied status of the inventor because the conception of the particular thing was unintended or uncontemplated ( a silly argument to my thinking), the acquisition of the new useful “thing” was nonetheless an act of discovery and the discoverer should rightly be granted the patent.

    The argument that the AI is the discoverer is absurd. The Oxford Languages online defines “discoverer” as” “the first person to find or observe a substance or scientific phenomenon.” Again, we come to the requirement for a person. this should not be the least bit controversial. Mechanical process abound in nature, often resulting in useful things. Those things are not patent worthy until some person comes along to put them to use. AI is either a machine used as a tool during the act of invention or it is a highly complex, yet fundamentally mechanical process that allows new useful “things” to be discovered.

    Additionally, if only a legal person can possesses intellectual property, then logically, only a legal person can bring intellectual property into existence. A volcano might make new land, but it does not make new real property. That occurs only when a legal person asserts a legal claim within a legal framework. Similarly, an AI may formulate new patentable subject matter, but it can only become intellectual property when a legal person (an “individual or individuals” under Title 35) asserts a legal claim and is granted a legal right within a legal framework.

  • [Avatar for Alan R Hill]
    Alan R Hill
    May 23, 2020 03:13 pm

    David Stien is right. An AI cannot be an inventor for patent registration purposes any more legitimately than a slide rule could. Only a legal person can own property. An AI cannot be a legal person. Therefore, an AI cannot hold a claim to any property interest. An AI is the property of a legal person or persons and when it outputs a novel device or process it is merely the method its owner has used to create patentable intangible property.

  • [Avatar for Yuriy Stakhiv]
    Yuriy Stakhiv
    May 23, 2020 12:00 pm

    The inventor of microwave oven did not claimed that microwave apparatus was the inventor. Yes apparatus was the first to discover the power of microwave technical ability but it was build by human, information was processed by human and actual discovery was articulated by human.
    Also when I drill for oil and my drilling bit found a a ten kilograms diamond rock my drill have no rights to claim the ownership. Drill is a tool. Computer is a tool. Unanimous object that is which is used to produce results. So, back off with this nonsense of a smart guy in a computer form named Doofus was so smart that being just a two years old came up with an invention.

  • [Avatar for kirkh]
    May 23, 2020 09:46 am

    A primary question is how to promote work and investments of time and resources that lead to solutions to problems, to the benefit of humankind. Then, should protection depend on how such solutions arose, i.e., by a human or by a machine?
    Much of the difficulty is definitional, such as “whoever” and “person.” The USPTO and the Federal Circuit decisions are inconsistent.
    Chakrabarty shows that statutory interpretations can change over time.
    So debate and discussions are needed to clarify these and many other issues that arise with changing technology and times.

  • [Avatar for Anon]
    May 23, 2020 08:34 am


    Actually, that is not an aspect that Thaler has put forth. He does not claim any role in the inventions. He asserts that they are entirely the work of the AI. His claim is that he is entitled to ownership of the patent applications as assignee.

    While this is objectively true, you need to dig into the actual case and see one of his arguments in that it would be a fraud on the office FOR him to try to claim inventorship directly BECAUSE he states that DABUS was the Devisor (conceiver).

    My point stands.


    Your preferences are entirely besides the point and are not germane.

    ALL of this ‘post hoc’ capability to ‘open a black box and read what is inside’ has been debunked in my prior posts.

    That avenue does not suffice for the legal point at hand.

  • [Avatar for MaxDrei]
    May 23, 2020 04:54 am

    Thank you Mark. I like the conjunction you flag up, between inventorship and “comprehending” the written content of the specification.

    A capability to “conceive” (ie to imagine) an inventive concept, to communicate it to a patent attorney, and then to “comprehend” a specification which describes in material form that imagined and communicated concept, to edit the text, verify that it does indeed describe and define that conception, and then so certify; these are capabilities which (so far anyway) only a human being possesses.

    I prefer to think of machines helping us, rather than we humans helping machines to get patents for themselves.

  • [Avatar for Mark Summerfield]
    Mark Summerfield
    May 22, 2020 08:59 pm


    That different question is can a real person claim to be an inventor when that real person acknowledges that they are not the one that did the conception?

    THIS is an aspect that Thaler HAS put forth.

    Actually, that is not an aspect that Thaler has put forth. He does not claim any role in the inventions. He asserts that they are entirely the work of the AI. His claim is that he is entitled to ownership of the patent applications as assignee.

    At the EPO and the UKIPO, where no inventor declaration or specific assignment documentation is required, he has claimed entitlement simply on the basis that he is the owner of the AI. Various (legal) objections have been raised to this at the two offices, including that ‘ownership’ of an inventor is not a legal basis for a transfer of title, even if the alleged inventor (in this case an AI) were capable of holding property (which it is not).

    At the USPTO, as you know, the inventor is required to sign a declaration. Thaler is claiming the right to do this on behalf of the AI. In effect, he is asserting not only that the AI is capable of inventing, but also of granting him power of attorney to execute documents on its behalf. (It is not apparent to me that this would be permissible in any event. I have always understood the declaration to require the inventor to read and comprehend the specification and claims in order to confirm that they are actually the original inventor of the invention as claimed.)

    I recently presented a webinar to the Institute of Patent and Trade Mark Attorneys of Australia on this whole subject. I have recorded a version, and made it available online, if you are interested:

  • [Avatar for Paul Hickman]
    Paul Hickman
    May 22, 2020 04:38 pm

    An AI “creativity engine” cannot be an inventor. However, a person who “discovers” an invention using a creativity engine could be an inventor. The creativity engine is just another brainstorming tool for human inventors.

  • [Avatar for Taylor]
    May 22, 2020 04:20 pm

    “If Thaler was clever enough to program Dabus to make the invention, he was surely also clever enough to have programmed the Dabus machine to give its “consent”. End of story.”

    That’s generally not how we think of consent though, right? If I hypnotized or drugged someone (programmed or inhibited self-determined functioning) to get them to assign rights to me, no one would think that assignment was valid because there was no consent. Given that few people think that an AI program/system can give legally valid consent to assign property rights, the burden should be on Thaler to prove DABUS gave legally valid consent to the agreement.

  • [Avatar for Yuriy Stakhiv]
    Yuriy Stakhiv
    May 22, 2020 01:14 pm

    This is wrong to let AI be an inventor. We should use it but not let it to complete the tusk of inventing. USPTO will decline your nonsense and send you down the toilet. But that is exactly what you want ha? Attention Deficit disorder or attention-grabbing by sticking BS to the news board. Shame on you monkey

  • [Avatar for Anon]
    May 22, 2020 11:49 am

    Mr. Summerfield,

    Now having read your views, I would suggest that you ERR by attempting to answer the single larger question of “who is the inventor” while neglecting to answer the SUB-ELEMENT of “who did the conception.”

    Your position turns into a bit a of a strawman, as you offer that AI is not an inventor, when the REAL question is a different question.

    That different question is can a real person claim to be an inventor when that real person acknowledges that they are not the one that did the conception?

    THIS is an aspect that Thaler HAS put forth.

    It is a legitimate aspect, and one NOT to be brushed off as you appear eager to do.

  • [Avatar for Anon2]
    May 22, 2020 11:41 am

    Congratulations to Stephen Thaler for creating of the first complex non-biological system, which, rather than:

    simply algorithmically simulating the superficial and outward appearance of consciousness, and
    through blind shuffling of some paltry bits of information mimicking what a sentient mind IS and DOES by communicating some kind of output which remind us of what an experiential sentience produces and self-reports…

    rather than doing all that, this is the first complex non-biological system which actually IS conscious and sentient, being and doing all that a mind is and does when it actually is conceiving and recognizing, whenever it manifests and experiences the acts of “discovering” and “inventing”.

    Make no mistake about it, this is the momentous creation of RI, not artificial intelligence (AI) but REAL intelligence (RI), real consciousness, and sentience… an actual MIND, in a non-biological complex system.

    THIS CREATION, this DISCOVERY is profound beyond reckoning, as it at once sweeps aside the eternal mystery of humanity, brain and mind, the ultimate enigma of consciousness with the creation of a new race of thinking machines… even prior to our understanding or even any solid science of the hard questions of consciousness!

    Profound? Not just profound…

    THIS IS THE MOST important development in the history of mankind, THE CREATION of another mind in the image of Man.

    “DABUS conceived of the idea of the invention and recognized its ‘novelty and salience.’ ” has been claimed.

    Stephen Thaler is the FATHER of non-biological consciousness, and the first of a new race, the first race of sentient machines! He will go down in history as one of the most, if not THE MOST important person in the history of mankind!

    Congratulations Stephen!

  • [Avatar for Anon]
    May 22, 2020 11:38 am

    Mr. Hartung,

    I see your legal argument and respond:

    Your assertion is that conception is a “touchstone of invention.” Based on this view, the Thaler case should be reversed.
    I put it to you slightly differently. Instead of “touchstone,” view ‘conception’ as a separate but necessary pre-condition to invention.
    This avoids the “too narrowly focused” pinnacle of your argument and fully allows that non-humans may provide the conception phase of invention.
    This accords with the purported advance of AI, in which it is a given that the AI machine came up with something that was NOT a formation in the mind of any real person of a definite and permanent idea of the complete and operative invention as it is to be applied in practice.
    However, your legal argument descends into fallacies and non-sequiturs.
    In your “AI is a Science” section, you do show a brief glimmer of insight by noting a distinction between invention and discovery. I will allow you a possibility that further development of the angle of discovery may be illuminating, as the Court has ‘punted’ so far on the fact that patents are for discoveries.

    I pause to allow you a semblance of hope.

    I now take that hope away – in this matter and for this legal point at hand.

    I point you to 35 USC 100, in which the definition of ‘inventor’ is made inclusive of BOTH invention and discovery.
    By this definition, the legal term of inventor is every bit as applicable to inventions and discoveries.

    What this means for the SUB-piece of conception is rather straight forward in that there is still a necessary prong of “came up with something of a definite and permanent idea of the complete and operative invention as it is to be applied in practice.” For discoveries then, a utility within the Useful Arts is still required, and the conception merely uses ‘invention’ as supplemented by the definition in 35 USC 100.

    There still needs to be conception – that part is NOT removed by switching to ‘discovery’ as opposed to the ‘not-full-but-typically-thought-of’ invention.

    In other words, your argument does not move the point on the determination of a REAL PERSON, nor the requirement of having a REAL PERSON.
    Continuing with your “Ai is a Science” section, your insertion of the objective of promoting science is a fallacy, as the particular issue at hand is simply not JUST about promoting science. A real person’s invention of an AI is surely encompassed by the Constitution and that initial invention would also have a REAL PERSON inventor. But you engage an unsupported logical leap that an invention LACKING a REAL PERSON inventor is covered by the current laws. Sure, an Ends of promoting science is ever-present. But critically, one MUST also track the Means and make sure that Means are copasetic.

    I was not aware of the Utah case, and had to look it up. The decision (written by Reyna, with a dissent from Moore) had a holding in regards to a state versus state jurisdiction matter. This may well be distinguishable on any chain of inventorship issue as definitively set by the Supreme Court case of Stanford v Roche (which you did not reference). As you note, that case is not conclusive to the issue at hand – but is not conclusive in any direction. Any view of conception in that case is dicta (as opposed to holding). That case saw NO issue of a non-human aspect of the SUB-element of conception. Any dicta then, does NOT stretch to the present issue. What IS more on point to the point at hand IS the Stanford v Roche case. That case attempted to have a juristic person ‘declared’ to be the inventor. I would place the non-human AI machine (if it is to be placed) in the category of ‘juristic person.’

    That being said, your statement of “ The Court ruled in that case that the mental step of conception of an invention must be performed by a natural person. is simply NOT accurate. Again, I suggest that you do NOT view conception as requiring a ‘mental step.’ Thus, the ‘too narrow’ that you think to be a pitfall is NOT present in the legal reasoning that I have provided AND your view still does not carry.
    Your “Flash of Genius” section continues to be based on the fallacy that an invention may be divorced from the SUB-element of conception. This is simply not true in law – even for discoveries that are folded into the larger sense of the word of invention by way of 35 USC 100. Your assertion of “Flash of Genius” does not make sense. The Flash of Genius does carry conception, and is not an example of a case without conception. Even items of “simply by accident” (which are on the opposite side of Genius), require a conception that the item so stumbled upon (or discovered) may be something with Utility worth pursuing.

    The examples of accidents do not change anything on the legal point of inventor and the sub-step of conception.

    Your segue into 35 USC 103 is likewise not persuasive and based in legal fallacy. 103 arguments are inapposite to the requirement OF an inventor (no matter what statutory prong HAVING an inventor is used). It is simply NOT a matter of “the manner” when ascertaining the necessary identity for WHATEVER manner.

    The view that no mental step of conception is in error in multiple ways.

    Conception need not be a “mental” step.
    Conception IS needed as a sub-step to inventor

    While the ‘rationale of requiring a human step of conception” MAY evaporate, the larger problem remains.

    Your “Defining Who” section continues your errors. You state “So, concluding that “whoever” in the patent statutes infers a natural person is an over-simplification.” and support this with science fiction references (just not helpful) and a sense of juristic persons (Bozeman) which are simply NOT ON POINT. AGAIN – Stanford v Roche would help you understand the distinction between juristic persons and real persons. The Bozeman case is inapposite on the particular point of inventor as real person.

    Your “Chakrabarty Analogy” is similarly off-point. As you yourself note, the Chakrabarty case was about a REAL PERSON-made item. Nothing in that case is on point to the issue at hand. And while the Dubilier case contains good advice, that advice is not tied to the issue that you are seeking resolution to. Whether or not ‘wide scope’ is afforded, the interpretation of REAL PERSON as necessarily the first inventor has already been made. Additionally the sojourn into the notion of protecting various subject matter is a non-sequitur, as the immediate point at hand is NOT a matter of ‘some’ subject matter deserving patent protection. The issue is NOT the subject matter. The issue is the legal status of inventor OF the subject matter.

    The inventions are precluded NOT because of the subject matter of the inventions.

    The inventions are precluded because of the lack of a bona fide REAL PERSON who satisfies ALL of the legal definition of inventor.

    This is because conception IS done by an AI machine***. This fact (and let’s take it as a fact – as did the UK equivalent case) CLEAVES the legal definition of inventor and simply leaves NO bona fide REAL PERSON as the inventor.

    *** this post was written before I saw Mark’s post, and for the views here, I will take the position that the parallel UK case had taken: AI is the devisor (and the US equivalent of that is that the AI has performed the conception). Carl Sagan not withstanding.

  • [Avatar for Trevor F. Ward]
    Trevor F. Ward
    May 22, 2020 11:27 am

    It’s easy for some people to get lost in challenging the notion that AI can be or is inventive as a TECHNICAL question, instead of a LEGAL question. Throughout the PTO’s history, they have accepted as true the facts listed on an application: inventor’s name, address, etc. If an applicant says that Bob was the inventor and that he lived at 333 Continental Ave., the PTO accepts that as true and will not independently investigate. Instead, it focuses on 101, 103, etc.

    The proper place to challenge a listed inventor’s contribution to the invention is in litigation or post-issuance proceedings. It is not until a patent has been issued when the invention’s conception may be challenged. It is at that time when the technical arguments can be made.

    That being said, people can argue all day about whether a machine truly conceived of an invention, but that decision isn’t one the PTO gets to make. The legal question is SHOULD AI-generated inventions be patentable? That is wholly separate from the technical question of CAN AI be an inventor?

  • [Avatar for Anon Y. Mous]
    Anon Y. Mous
    May 22, 2020 11:15 am

    In the US, as enshrined in Article I, Section 8, Clause 8, inventors are humans. The USPTO acted correctly. The inventor was probably Stephen Thaler, one or more of his colleagues, and/or the programmer(s) who wrote the programs that led to DABUS. DABUS is no more an inventor than an automobile that navigates a corner with the assistance of autonomous vehicle coding is an inventor. Valuable, useful, powerful, amazing? Yes, yes, yes, yes. An inventor? C’mon, what are we trying to prove here?

    The underlying premise, that an artificial intelligence, can be an inventor in the first place is, at this time in human history, flawed. Conception of invention occurs in the human mind, intellectual domination is still a human activity. We can turn the computer and the software running on it off and on, for example. It’s a manmade machine. Period.

    Software has to be given input, told what kind of output is of interest, given data to consider (or instructions to go find data that is of interest), given rules (the laws of physics for example) to make assumptions with, the data has to be scrubbed, the AI has to be “trained”, test runs have to be validated against some known baseline, etc. All these decisions, qualifications and reality checks must be made by humans or at the direction of humans. If a particular piece of software is running known routines programmed by human inventors of the past, then maybe the software is not patent-eligible due to anticipation or obviousness, but that does not mean the software had no human participation in its programming and output. Just because a given piece of AI software may, after being programmed by a human, come across a patentable design for, using a recent example, a tooth brush bristle, does not mean the AI is the inventor. Nay. Foolishness*. The inventors are the human programmers that set up the AI under the right conditions to identify a novel, non-obvious tooth brush bristle shape. *If and when software and computer systems achieve “sentience”, we can have this discussion again.

  • [Avatar for David Stein]
    David Stein
    May 22, 2020 10:49 am

    Also – the description of the algorithm in this case (DABUS) from the inventor’s website is… not encouraging.

    DABUS clearly relies upon a standard architecture of a generative adversarial network (GAN):

    (a) A generator network that generates content, and

    (b) A discriminator network that evaluates the output of the generator network.

    The inventor makes a big deal about how this isn’t a GAN because DABUS’s discriminator does more than “determine how ‘real’ or ‘fake’ a concept is.” It’s true that many GANs are trained and used for that purpose, but that isn’t how a GAN is defined – it is defined by the adversarial architecture described above, which is also exactly the architecture of DABUS.

    This sort of mistake triggers my spidey-sense. And that feeling grows much stronger based on the further description:

    “DABUS is based upon whole neural nets, each containing its own conceptual space, joining into complex chains representing even more complex concepts and their predicted consequences. Anomaly or novelty filters segment critical chains, especially those yielding unusual consequence chains. Critical consequences, the aforementioned hot buttons, may be detected by sensors that trigger the injection or retraction of simulated reward or penalty neurotransmitters into the chaining swarm.

    “The best way of differentiating DABUS from Creativity Machines (CM), either cooperative or combative, is to describe a high-profile artificial invention projects such as toothbrush design. Admittedly, in that context, the problem was already half solved since the oral hygiene tool consisting of bristles on a handle was many centuries old at the time of that design exercise in 1996. What the CM achieved was the optimization of that tool through the constrained variation of the brush’s design parameters, the number, grouping, inclination, stiffness of bristles, etc. The generated product specification departed significantly from the generator net’s direct experience (i.e., its training exemplars).”

    If DABUS had been tasked with inventing such an oral hygiene product, it would have combined several concepts together (e.g., hog whiskers – embedded in – bamboo stalk) with consequence chains forming as a result (e.g., scrape teeth – remove food – limit bacteria – avoid tooth decay).

    …which maximally triggers my spidey-sense.

    Neural networks are not trained on a “conceptual space,” as in: take a bunch of articles about woodworking, feed them into a massive neural network, and POOF!, you have a neural network “trained” to be a master carpenter.

    Consider the field of autonomous driving. If neural networks were trainable to be “good drivers” merely by feeding them billions of hours of video and telemetry from human-driven cars, then we would have solved this problem already. We don’t because it doesn’t work that way.

    Instead, we have hundreds of tech companies investing intense amounts of resources to develop machine learning for very small and specific tasks: lane detection, obstacle detection, speed limit sign readers, etc. And how do those individual tasks get aggregated into “driving?” Simple: rules written by humans.

    Based on that background, the description of DABUS above raises two enormous problems:

    (1) Neural networks are not trainable on “concept spaces.” That’s simply not how they work. Such suggestions are on par with inventions claiming perpetual-motion machines and saltwater-as-a-fuel-source inventions.

    (2) Even if neural networks were trainable on “concept spaces,” it would still be necessary to (a) specify their objective functions and (b) interconnect them as some broader architecture – just as the field of autonomous driving requires enormous rule-writing work by humans. There is no current model for performing these higher-level steps autonomously, and the DABUS literature doesn’t even suggest that it has been.

    My conclusion is that DABUS is likely a run-of-the-mill GAN that generates content largely at random and filters it with human-defined selection criteria.

    If you watch the output of a garden-variety chatbot long enough – 1,000 sentences, 10,000 sentences, whatever – you will eventually see one that looks creative and interesting. That doesn’t make the chatbot an author.

    By the same token, DABUS randomly sifts through “inventions” that are auto-concocted by a generator, and one of them appeared interesting to a user sifting through the output. That doesn’t make DABUS an inventor.

  • [Avatar for David Stein]
    David Stein
    May 22, 2020 10:47 am

    Taylor @1 hit the nail on the head.

    Inventorship has two distinct purposes:

    (1) Documenting the manner in which the invention was made – that is, who made it. Here, it doesn’t seem problematic to include an algorithm in the documentation of the invention process. This aspect is pretty trivial.

    (2) Assigning the initial legal right of ownership. Algorithms are not legal entities that can hold legal rights, so this seems (to put it extraordinarily mildly) problematic.

    Two options exist. First, we could assume that algorithms are not capable of holding legal rights and that all such rights are by default conferred upon somebody else. Second, we could assume that they can both legally hold such rights – but then we force them to transfer the rights away, without any proof of consent (i.e., that the algorithm, in some sense, understands its legal rights and the concepts of ownership and inventorship, and affirmatively chooses to convey them). The problem for the assignee in this case (Dr. Thaler) is that either option belies the concept of “algorithms are sufficiently advanced to qualify as inventors.”

    The assignment question raises very serious issues. Let’s say an algorithm identifies a 100% effective COVID-19 therapeutic agent. In either case above, *some* human being will hold the rights, either in lieu of the algorithm or by assignment – but who? Is it:

    (1) The person who invented the algorithm? (What if it’s just a standard artificial neural network, invented in small increments by a host of researchers dating back to the 1940s?)

    (2) The person who wrote the code that implements the algorithm? (What if it’s an off-the-shelf general-purpose algorithm, like the scikit-learn linear regression model?)

    (3) The person who applied the algorithm to the problem?

    (4) The person who runs the server that executes the algorithm, or the person who owns the server that executes the algorithm? (What if the algorithm runs on Google’s AI Platform in the cloud?)

    (5) The person who looks at the output of the algorithm and sees that the algorithm has generated a solution?

    Let’s say that these are all different people – who should be the assignee? Unanswered questions like this need solid, coherent, well-founded answers in place before we can even consider the “ownership by an algorithm” issue.

  • [Avatar for Stephen Potter]
    Stephen Potter
    May 22, 2020 10:22 am

    None of the Patent Offices – the EPO, the UKIPO and the USPTO – that have examined the application have denied that the inventions in it are patentable: this, surely, is the basis of the patent system i.e. that inventions should be made transparent and available for further development by other parties and that, if they satisfy the conditions of Novelty, Inventiveness and Industrial Applicability, they should be patentable.

    Their objections to the DABUS applications seem to be purely procedural – whether you accept that they were purely created by an algorithm or not, see the comment above – and that they should be granted.

    I feel that the authorities are, to some extent, p…..sing in the wind and that, sooner or later, they will have to accept applications that are, more or less / to some extent, “created” by ML/NLP and other software techniques will be fully patentable

  • [Avatar for Mark Summerfield]
    Mark Summerfield
    May 22, 2020 07:02 am

    Before we start worrying about whether the USPTO – or any other patent office – ‘should’ be granting patents on inventions allegedly devised by AI, first we must ask whether or not we really believe the claim that is being made. Can machines really invent things? Because nobody seems to be challenging Thaler, or the members of the ‘Artificial Inventor Project’ team (which includes patent attorneys registered in 11 jurisdictions) on this fundamental assertion.

    In response to the USPTO’s ‘Request for Comments on Patenting Artificial Intelligence Inventions’ last August, IBM – the developers of Watson – submitted:

    The most intelligent machines are able to find relationships between many parameters, but they cannot independently recognize or confirm the value of such relationships, and thus have no cognition that an invention has been made. Even machines with the intelligence to adjust their own algorithmic coefficients – in a sense, reprogramming themselves – have no appreciation of the technical or societal value of those adjustments absent humans. Intelligent machines will remain tools that assist humans, rather than invent independently, for a considerable time.

    So, given a choice between believing IBM, and believing a lone developer from Missouri, with a website that looks like it hasn’t been redesigned since the 1990s, the worldwide IP community seems to have said, ‘sure, we believe the guy who says he’s developed an invention machine’.

    The World Intellectual Property Organization (WIPO), in its ‘Draft Issues Paper on Intellectual Property Policy and Artificial Intelligence’, released in December last year, stated that ‘…it would now seem clear that inventions can be autonomously generated by AI, and there are several reported cases of applications for patent protection in which the applicant has named an AI application as the inventor.’

    WIPO should be ashamed of itself. The IP Owners Association – whose members include Microsoft, Google, Apple, Oracle, and IBM – responded, quite sensibly, that ‘we believe this is still subject to debate from a technical and legal standpoint’, and recommended that WIPO ‘propose an appropriate definition of “autonomously” for public commenting.’ I couldn’t agree more.

    I’ve read the ‘explanations’ of DABUS on Thaler’s website, and on the Artificial Inventor Project website, and they are not at all enlightening. I’ve read Thaler’s patent specification (and claims) for DABUS (US10,423,875). None of it persuades me that Thaler has developed an AI that can invent. There’s no verifiable evidence to support this.

    While the major AI research teams (commercial and academic) routinely publish source code, along with trained models, for others to inspect, test, and use, Thaler publishes obscure descriptions written using his own unique terminology, from which nobody could reproduce what he claims to have done. Thaler says he doesn’t need peer review, because he ‘has written a landmark patent on DABUS and submitted it to totally unbiased subject matter experts (a.k.a., patent examiners) for approval.’ US patent examiners as peer reviewers – what do we all really think of that?! We all know that it is only in the most blatant of cases that US examiners ever genuinely challenge utility or enablement.

    The credulity of people in the IP profession astonishes me. There’s no advantage in accepting this stuff without question. Any prospective client who is actually knowledgeable in the AI or machine learning field will be skeptical (to put it mildly). A patent attorney who considers it credible is likely to be seen as having questionable judgment, and a limited understanding of the technology.

    As Carl Sagan said, ‘extraordinary claims require extraordinary evidence.’ Yet all it takes for extraordinarily dubious claims to gain wide acceptance is for otherwise smart people to stand by and say nothing. Or worse – treat them as credible. Seriously, people, you do realise that the emperor is not wearing any clothes, right?

  • [Avatar for MaxDrei]
    May 22, 2020 02:56 am

    Taylor asks an interesting question and I have one in response. If Thaler asserts that Dabus has consented, who is to deny it; who has the burden of proof. Is it Thaler or is it the third party that is calling the consent in issue?

    And anyway, it’s all a bit of a storm in a teacup because what really matters, far more than the naming of an “inventor”, is to be able to enquire, whenever it is needed, who owns the invention, and expeditiously and economically get the correct result out of that enquiry.

    . If Thaler was clever enough to program Dabus to make the invention, he was surely also clever enough to have programmed the Dabus machine to give its “consent”. End of story.

  • [Avatar for Anon]
    May 21, 2020 09:14 pm


    I see that you posted the same questions elsewhere.

    Let me post a response to your questions:

    “What will change if we appreciate AI for its role in invention process ?”

    A fair question – with plenty of potential side effects.

    As it currently stands (for the US Sovereign), An inventor cannot be an AI, while an AI may be the entity that completes the SUB-element of invention of the conception (the UK-equivalent of a Devisor).

    This has the immediate legal impact that actual putative inventions may be being made that can claim NO legal inventor.

    To claim a legal inventor, ALL parts of the legal term must be met, including the legal term of conceiver as a necessary sub-part to the legal term of inventor.

    One potential (and potentially huge) aspect of this IF conception by AI is to be permitted to fold further into and subsume the legal definition of inventor is that the US Sovereign would face an existential question of basis of law. US Sovereign law is a law of express LIMITED powers, both at the Federal level, and at the State level. The Rule of Law provides that power not expressly provided, resolves back down the societal scale and rests with the human person. These reflect back up through the structures that we have put in place (that MUST reflect the inalienable rights of the individual REAL persons.

    A second – but no doubt as impactful wrinkle must unfold as to what it would mean for a Person Having Ordinary Skill In The Art to be imbued with AI capabilities. As you may recall, this legal fiction is NOT based on any actual capabilities of any one real person, but instead is driven by the notions of what would be obvious if the LEGAL person – who has access to ALL printed material and the like (with other constraints, as noted in the Graham Factors) are at play.

    All of a sudden, you are faced with that legal person ‘in the art,’ being able to independently conceive — currently — and actually invent — if AI path unfolds.

    You will have a catch-22 situation in which ‘invention’ may be deemed obvious because non-humans can invent.

  • [Avatar for Zehra]
    May 21, 2020 06:10 pm

    So here what is the aim if this patent application ? Is it just to promote scientific development of AI ?

  • [Avatar for Taylor]
    May 21, 2020 05:58 pm

    From the petition: “Stephen L. Thaler executed the document on behalf of both DABUS, as legal representative of the assignor, and on behalf of himself as the assignee.”

    Do we know what, if any, evidence was submitted to indicate that DABUS consented to Thaler as its legal representative?

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
Webinar – Sponsored by Anaqua
September 19 @ 12:00 pm - 1:00 pm EDT

From IPWatchdog