AI-Generated Invention at the Federal Circuit

On Monday, June 6, the Federal Circuit will hear oral arguments in the pending  AI-inventorship case of Thaler v. Vidal, Appeal No 21-2347. The court does not release the identity of the individual judges until just before oral arguments, so all we (and the parties) know is that it is Panel A in Courtroom 201.

This case raises the novel legal issue of whether a patent can be obtained for an invention created by an artificial intelligence (AI) in the absence of a traditional human inventor (“AI-Generated Invention”). The United States Patent and Trademark Office (“USPTO”) and the District Court for the Eastern District of Virginia have barred such inventions from being patented. This ignores fundamental statutory and constitutional principles and also stymies innovation. While other countries are promoting the progress of science, the USPTO is belatedly adopting luddism.

Appellant’s brief.  In their responsive brief, the US narrowed the issue somewhat:

Whether the District Court correctly concluded that an artificial intelligence device comprised solely of source code cannot qualify as an “inventor” under the Patent Act, where the plain statutory language specifically defines “inventor” to be an “individual” and refers to an “inventor” using personal pronouns.

Appellee Brief.

That same panel is hearing two other IP cases:

  1. Samsung v. Dynamics, 21-2352.  PTAB sided with the patentee Dynamics.  On appeal, Samsung argues obviousness — that the Board failed to consider some aspects of the prior art and required motivation-to-combine be found in the primary reference.  Dynamics patents cover electronic emulation magnetic stripes (on, for instance credit cards).  The infringement argument is that Samsung’s digital wallet infringes. John O’Quinn (Kirkland Ellis) will argue for Samsung and Robert Morris (Eckert Seamans) for Dynamics.
  2.  In re A. Zeta S.R.L., 22-1178. Zeta is an Italian company seeking to register the mark PARMA COFFEE to be used for coffee, chocolate, tea, and other food/beverages.  Parma is a mid-sized city in Italy (200k people) and the goods in question apparently originate from Parma. The PTAB refused to register the mark — finding the COFFEE portion of the mark generic, and PARMA to be “Primarily Geographically Descriptive” under Section 2(e)(2) of the Lanham Act.  One issue here is about “origination” — the coffee is not grown in Parma but the PTO still argues that there is sufficient connection with the city (and also that Zeta waived that argument before the TTAB).  The parties waived oral arguments in the case and so it will be submitted on the briefs .

 

 

 

 

89 thoughts on “AI-Generated Invention at the Federal Circuit

  1. 11

    I missed the first half of Thaler’s opening arguments, but I doubt I missed much. However, as the USPTO accurately pointed out, this is a straight-forward statutory construction issue about the meaning of the word “individual.” Put simply, an AI is not an individual. If Thaler wants to change that, he needs to go to Congress and have them change the language.

    I would suggest, to Thaler and any like-minded individuals, that before they go to Congress, they should come up with some better so-called AI-generated inventions. A flickering light and fractal-surfaced shipping container are hardly going to pull on the heartstrings of those in Congress and convince them that the world would be a better place if AIs were allowed to be inventors.

    Thaler’s attorney talked briefly (in his rebuttal) about Congress setting up the patent act to ‘encourage innovation’ and that the patent act should be interpreted in that vein. Perhaps this attorney should have contemplated how the Federal Circuit has interpreted 35 USC 101 before making that argument.

    1. 11.1

      Perhaps this attorney should have contemplated how the Federal Circuit has interpreted 35 USC 101 before making that argument.

      Ha! Point well taken. For whatever little it is worth, I agree with all of the other substantive points in #11 as well. Title 35 is not on DABUS’ side as it stands.

      If Thaler wants to convince Congress to amend Title 35 to get him where he wants to go, I think that he would be ill-advised to point to any particular AI-generated inventions. Instead, he should try to organize large corporations to lobby for Title 35 to be amended to allow corporate entities to be listed as “inventors” and not merely as “applicants,” much in the same way that corporations may already be listed as “authors” in the copyright context. That would solve the problem without having to spend any time convincing anyone of the value of any given AI-generated inventions.

      1. 11.1.1

        Hmm,

        Sounds like somebody needs BOTH that Lockean understanding AND to read the Stanford v Roche case…

  2. 10

    Does DABUS have an employment agreement with Thaler’s company, or does it still own the rights to its invention? Can DABUS walk away and sell or license its rights in the invention to another company? If it is a work for hire situation, has DABUS ever been paid for its inventing labor, or is this a case of involuntary servitude?

  3. 9

    Above is why the doctrine of “simultaneous” conception and reduction to practice exists. It could easily apply to AI inventions. (And again, otherwise the AI invention rightfully will get zero patent coverage in the USA).

    Back to my grown up real job.

  4. 7

    This reminds me of the Star Trek: Voyager episode, entitled “Author, Author,” which deals with whether an artificial hologram’s creations are copyrighted.

      1. 7.1.1

        OTOH, the Trek shows suggest that, even in the year 2364, virtually everyone is skeptical an AI could invent something.

        1. 7.1.1.1

          Yes, but the Roddenberry science fiction world had some serious other “reality” items that one had to overlook in order to enjoy the shows.

      2. 7.1.2

        As I recall, Commander Data’s case was cited as precedent in the “Author, Author” episode.

  5. 6

    “This case raises the novel legal issue of whether a patent can be obtained for an invention created by an artificial intelligence (AI) in the absence of a traditional human inventor (“AI-Generated Invention”).”

    Here. Fixed:

    “This case raises the ludicrous idea that any entity other than a human can obtain a patent.”

    In some wacky alternative universe where AI can obtain patents (didn’t we see this in a Three Stooges skit circa 1955?) . . . a SCOTUS – CAFC tag-team would invent ways to crush innovation . . . by making up their own, unconstitutional rules for what constitutes patentable subject matter . . .

    Like such a thing could ever happen in America.

    1. 6.1

      You have the timing of those things mixed up — and thus your attempted “causal logic” does not reach.

    2. 6.2

      I don’t think the case is about whether an AI can ever own property. Rather, I think it is about whether an AI can ever be named, truthfully, as the inventor of the subject matter of a duly issued claim, duly issued to a legal person, namely the owner of the output of the AI.

      Is devising and then outputting a statement purporting to be a solution to a technical problem (as DABUS did) sufficient to qualify DABUS as the true “inventor” of the claimed subject matter. I suspect not.

      The way I see it is that we do not yet know how to write a program to create a computer that makes conjectures. Knowledge is created when a conjecture is conceived, formulated, examined, tested and found to be well-founded.That is something which, in the entire known universe, only the human brain can do. And the human brain has not yet figured out how to build an artificial conjecture generator. We don’t yet know how long it will take, to invent true AI.

      1. 6.2.1

        I don’t think the case is about whether an AI can ever own property.

        Actually – leastwise in the US – this very much IS what the case becomes about.

        1. 6.2.1.1

          IDK. This seems more like a narrow statutory interpretation problem. The broader, law-review fodder questions (e.g., could an AI ever be an author, inventor, property owner, etc) don’t need to be answered quite yet.

          1. 6.2.1.1.1

            OC,

            While I do see your point that the doctrine of “judicial economy” (in the sense of avoiding Constitutional issues when a simpler issue may resolve the matter directly in front of the court), my comment is meant more to pierce to the heart of the matter, as such may inform multiple resolutions.

            1. 6.2.1.1.1.1

              Fair enough (though my complaint is more a matter of the proper/legitimate role of the judiciary than of judicial economy)

              1. 6.2.1.1.1.1.1

                I hear you – in this case though, pure statutory reading is likely not enough (I have noted the weakness in “but pronouns are used,” given that pronouns are also used for the non-real person Legal Fiction of Person Having Ordinary Skill In The Art.

                Not certain why, but previously Greg has attempted to dismiss the Supreme Court case of Stanford v. Roche, even though that case contains the most on-point reasoning pertaining to “inventor” through the “initial inventor” reasoning (which directly implicates the basis of that “initial” through the Lockean nature that informs the presence of the Constitutional grant of authority to the Legislative Branch to write patent laws.

                It is important to note (and distinguish) the difference between to whom a patent may be granted and the satisfying of the legal definition of inventor.

                In the DABUS case, it is asserted that there is NO human inventor that meets the legal definition of inventor.

                It has also been asserted that DABUS meets that legal definition.

                As — it is a given — that an invention IS present, this necessarily leads to an easy conclusion.

                DABUS meets the legal definition of inventor.

                That being said, that is not enough under US law for DABUS to receive a patent grant.

                Under US law, the analogy of the Simian Selfie is clearly on point.

                There, a creative expression was made and captured in a fixed media — meaning that the analogy-to-invention was present.

                There too, the non-human “person” was determined to be the analogy-to-inventor.

                But just because BOTH of those conditions were met, did NOT mean that the analogy-to-patent-grant was made.

      2. 6.2.2

        Knowledge is created when a conjecture is conceived, formulated, examined, tested and found to be well-founded.

        Knowledge and inventorship are very different things (as you no doubt know).

        You seem to fall into the same general category of people wanting the AI to be something more – akin to The Singularity.

        First things first – that (later) thing is NOT needed for the present legal situation.

        1. 6.2.2.1

          You seem to fall into the same general category of people wanting the AI to be something more – akin to The Singularity.
          No. The bar isn’t that high. However, it has to be more than a glorified microscope — a tool that enables someone to perceive something that they wouldn’t otherwise be able to perceive.

          First things first
          Funny coming from you. First thing first is to see if US law permits AI to be an inventor. The district court made the legal finding, based upon its interpretation of the statute, that it cannot.

          If the Federal Circuit doesn’t set aside that statutory construction (which I highly doubt that it will), then whether DABUS is, in fact, capable of invention is not relevant. Again, putting first things first.

          To sum up, the order of things necessary for DABUS to be a named inventor (i.e., first things first, second things second, etc.) is the following:
          (1) A statutory construction that an AI can be an inventor under US law (if not, “do not pass go, do not collect $200”)
          (2) Assuming that AI is capable of being an inventor under US law, then there needs to be factual findings that DABUS (i.e., as an AI) contributed something sufficient enough to warrant being named as an inventor.
          (3) The patent application has to issue as a patent.

          (1) isn’t going to happen. As such, points (2) and (3) are moot. Even if (1) happened, then (2) isn’t likely to happen based upon my understanding of the current capabilities of AI and the law of inventorship.

          Really, the only real suspense lies in whether the Federal Circuit chooses to issue an opinion (I suspect they will) or whether the Federal Circuit disposes this under Rule 36. I doubt SCOTUS wants to touch this — the only reason they would is if the Federal Circuit, in a grand surprise, decided for Thaler/DABUS.

          Once disposed of by the Federal Circuit (and cert defined by SCOTUS), Thaler’s only recourse is to lobby Congress to allow for AIs as inventors. Good luck with that.

          1. 6.2.2.1.1

            I’m interested in who has standing in the US to challenge point 2 (is DABUS the real inventor). I know in UK only a person claiming to be the true inventor could make the challenge, what is the position in the US?

            1. 6.2.2.1.1.1

              The position in the UK is not quite how you express it. The right to challenge inventorship is enjoyed only by the person claiming to be the rightful owner of the patent application or patent. See Section 37 of the UK Statute of 1977. Human knowledge is not yet at the point when an AI can claim to be the rightful owner and assert its rights in a court of law.

              It’s different in the USA, as you surmise.

            2. 6.2.2.1.1.2

              I’m interested in who has standing in the US to challenge point 2 (is DABUS the real inventor).
              The USPTO could. See In re Verhoef, 888 F.3d 1362 (Fed. Cir. 2018) (affirming rejection under 102(f) that application did not accurately name the correct inventors).

          2. 6.2.2.1.2

            Actually Wt – the “first things first” has been abundantly addressed (as you may seek to redefine it) by me: As I have stated, the Lockean nature of US patent law precludes the granting of a patent to an AI.

            This though is different — in fact — as to whether or not an AI meets the legal definition of inventor.

            As I have also pointed out, the analogous case of the Simian Selfie informs what may happen to iinventorship.

            But you would rather avoid OTHER issues that do arise – such as the impact on joint-inventorship, and how is the other non-real-person legal fiction of Person Having Ordinary Skill In The Art informed of State of the Art considerations.

            You would rather insist on clenching tight your eyes or demanding some type of The Singularity event that need not be present FOR the current legal considerations.

            1. 6.2.2.1.2.1

              As I have stated, the Lockean nature of US patent law precludes the granting of a patent to an AI
              The so-called “Lockean nature of US patent law” may be worth an article in a law journal but it has no apparent relevance to the statutory construction on who may or may not be an inventor — putting first things first.

              As I have also pointed out, the analogous case of the Simian Selfie informs what may happen to iinventorship.
              Which means to say that an AI cannot be an inventor. Or do you read Naruto v. Slater any differently?

              But you would rather avoid OTHER issues that do arise – such as the impact on joint-inventorship, and how is the other non-real-person legal fiction of Person Having Ordinary Skill In The Art informed of State of the Art considerations.
              First things first. You cannot qualify as a joint inventor if you cannot qualify as an inventor. As best as I can tell, the “non-real-person legal fiction of Person Having Ordinary Skill In The Art” also cannot be an inventor.

              You would rather insist on clenching tight your eyes or demanding some type of The Singularity event that need not be present FOR the current legal considerations.
              This case will soon be dead. IMHO, it is DOA as it is. If I want to explore the implications of an AI (or microscope) capable of invention, I’ll read some good science fiction.

              Oh, and as I mentioned before, I don’t need a singularity event from AI. However, I need more than a glorified Eliza (or microscope).

              1. 6.2.2.1.2.1.1

                The so-called “Lockean nature of US patent law” may be worth an article in a law journal but it has no apparent relevance to the statutory construction on who may or may not be an inventor…

                Just so.

                AI cannot be an inventor. Or do you read Naruto v. Slater any differently?

                I think that we can all agree that: (1) the analogy of Naruto v. Slater to the patent context is a good one; and (2) extending the analogy, we should conclude that Title 35 does not presently permit a non-human to be named as an inventor. I gather that there is live disagreement about what the law should be, but I think that there is actually overwhelming consensus about these parts as to what it actually is.

                1. I gather that there is live disagreement about what the law should be
                  I would like to hear the pros for allowing AI to be a named inventor. Thaler has outlined some. Does any of them resonate here? Does anyone have any better reasons?

                2. Just to be precise, the statute cannot really control whether an AI can be an inventor. If the AI invents something, then it is an inventor, whether the statute acknowledges as much or not. I think that all participants around these parts, however, are in agreement that Title 35 (as properly interpreted) does not permit an AI to be acknowledged as an inventor, or to be listed as an inventor.

                3. This one needs a more subtle touch:

                  Title 35 (as properly interpreted) does not permit an AI to be acknowledged as an inventor, or to be listed as an inventor.

                  Title 35 may well acknowledge — but that is distinctly different than be listed.

                  Acknowledging requires no more than what the Simian Selfie case did: the simian was the one that captured the expression in fixed media.

                  Being listed – however – requires the (much more than a mere academic paper) understanding of the Lockean nature, as that would occur only with the grant of the patent. As I have (correctly) mapped out, the Simian Selfie analogy yields a “no copyright == no patent,” which is indeed distinctly different than attempting to say “no invention.”

              2. 6.2.2.1.2.1.2

                Which means to say that an AI cannot be an inventor.

                No – you miss the analogy.

                Being the inventor is analogous to the Simian taking the snapshot.

                Obtaining a patent is analogous to obtaining the copyright.

                No one is arguing that the Simian did not take the snapshot.

                1. No one is arguing that the Simian did not take the snapshot.
                  You are both missing the analogy and making a nonanalogous point.

                  The ruling was that without express statement, by an Act of Congress, that permits animals to have statutory standing as it pertains to the Copyright Act, they do not. The analogy is that for an AI to have standing (as an inventor), then this needs to be expressly found in an Act of Congress regarding patents. Since an AI being an inventor is not expressly found in the patent statutes, these statutes are interpreted so as to not permit an AI to be an inventor.

                  Also, your point that “No one is arguing that the Simian did not take the snapshot” is not analogous here. There are many with severe doubts that DABUS invented what was purportedly invented. One of the reasons is that facts behind this so-called invention have been shrouded in mystery. Another reason is that the facts are inconsistent with the known capabilities of AI. While the Court may presume that the pled facts are true for purposes of summary judgement, this does not mean that these facts are uncontested.

                2. You are both missing the analogy and making a nonanalogous point.

                  No – not at all.

                  I am the one (the first one) to provide the analogy.

                  You may not like how it maps — but that is a “YOU” problem.

                3. If you are still having trouble mapping the analogy, I walk through it (slowly) here:
                  Why bury your analogy under a link? If you think your analogy is apt, then repeat it here.

                  Also, I have little interest trying to reconstruct your thought process that may span multiple different comments. However, you did something elsewhere here that I’ll address.

                  In the DABUS case, it is asserted that there is NO human inventor that meets the legal definition of inventor.

                  It has also been asserted that DABUS meets that legal definition.

                  As — it is a given — that an invention IS present, this necessarily leads to an easy conclusion.

                  DABUS meets the legal definition of inventor.

                  What kind of twisted logic is this? There are enough holes in these few statements to make swiss cheese blush with envy.

                  First, don’t confuse assertions with facts. Also, the person (Thaler) doing the asserting is not legally trained. As such, he is poorly-position to assert whether or not “NO human inventor that meets the legal definition of inventor” and “DABUS meets that legal definition.”

                  As for an invention being present — again, that is an allegation, not a fact. Your conclusion is essentially that since Thaler alleges that DABUS is an inventor therefor DABUS meets the legal definition of inventor.

                  First things first. It is for the Court to decide whether the legal definition of inventor encompasses an AI.

                  There too, the non-human “person” was determined to be the analogy-to-inventor.
                  That is incorrect. I read both the District Court and Appellate Court decisions. At no point did they concede that Naruto (the name of the monkey) was an “author.” In fact, this is what they stated:
                  Naruto is not an “author” within the meaning of the Copyright Act.

                  Naruto was not an author, and DABUS is not an inventor. That is the analogy you fail to see.

                4. “What kind of twisted logic is this?”

                  The kind learned by spending years “winning” arguments by showing examiners to be “wrong” based on the amendments.

                5. I did not find this:

                  fact, this is what they stated:
                  Naruto is not an “author” within the meaning of the Copyright Act.

                  What is there is different: that Naruto does not have standing to sue under the copyright act.

                  The majority found Article III standing. To do so — de facto — the Simian is recognized as the author. It is simply — as I mapped out — that author or inventor Being present And expression in fixed media OR invention being present, does not mean being able to sue under copyright law OR award a patent grant.

                6. Further Wt, the Wiki supports my views , as does link to copyright.gov

                  See pages 54 and 68 (at a minimum) – CLEARLY a non-human MAY be deemed an author (the one that captures an expression in a fixed media). And just as clearly, just because and author exists, and an item of expression has been captured in a fixed media, that does not mean that SUIT under copyright law is permitted.

                  This is exactly as I have mapped to the invent/grant scenario.

                  A non-human MAY WELL BE an inventor.
                  An invention may be made without a human inventor.

                  And just as clearly, just because and inventor (analogy to author) exists, and an invention (analogy to item of expression has been captured in a fixed media), that does not mean that patent grant (analogy to SUIT under copyright law) is permitted.

                  Your assertions are simply false.

                7. … and (of course), Ben misfires as usual.

                  Time again for you to change your shoes, Ben – you really stepped into it.

                  Not only did you also miss the point, your counter of “with argument” assumes facts not present – and which quite frankly do not apply to me.

        2. 6.2.2.2

          Agreed, that “knowledge” and “inventorship” are not the same thing. But what about the three verbs conceive, devise and invent. What important differences do you see, between these three acts? In particular (and given the premise adopted by the UK and DE courts in the DABUS case) if the AI devised the claimed subject matter, does that make DABUS, alone, the true and first “inventor” of the claimed subject matter?

          It might not matter under US law but it sure does under the UK patent statute which defines the inventor as the devisor of the claimed subject matter. Will the Fed Ct tell us that devising is not enough, for an AI to acquire the status of “inventor”?

          1. 6.2.2.2.1

            Before we jump to the “wahtaboutISM,” let’s fully appreciate the admission that you make here.

      3. 6.2.3

        “ We don’t yet know how long it will take, to invent true AI.”

        Probably in a few years!

        Hahahahahaha. Invest in ProphetCoin now, folks.

      4. 6.2.4

        Max,

        …devising and then outputting a statement purporting to be a solution to a technical problem (as DABUS did)…

        Can you direct me to any evidence that this is what happened? Or even to a direct assertion attributable to Thaler that what DABUS does involves “outputting a statement purporting to be a solution to a technical problem”? Because I have not seen such a thing. And while courts around the world have been compelled to proceed on the basis that DABUS is the inventor, because Thaler has insisted upon it, to my knowledge none has actually been presented with any evidence, or been asked to adjudicate the veracity of Thaler’s assertions.

        What are the consequences if all of this effort is being expended based upon a lie?

        1. 6.2.4.1

          Mark, as you know, I’m a bit of a dilettante on these boards, throwing stuff against the wall (as anon constantly observes) to provoke a response. So, no, I can’t cite any specific ebvidence.

          What I had in mind is that it is implicit in naming DABUS as inventor under the EPC that DABUS has come up with a technical solution to a technical problem, for that is the only thing that counts under the EPC as a patentable invention.

          My own view is that for the purposes of patent law the inventor is likely to be the human that “conceived” of a patentable invention upon evaluating the output of the DABUS machine.

          Greg deLassus on these boards has good knowledge of the use of AI in the pharma field, to output molecular structures as candidates for research programs. He (I understand) is open to the idea that the “inventor” in such cases is the machine that outputs the structure rather than the human that evaluates that output.

          Me, I know next to nothing about the current capabilities of AI. So I just throw up stuff here and see what comes back at me. Thanks for the comment

          1. 6.2.4.1.2

            Does Greg also think that a machine which screens a giant library of compounds for a function is also an “inventor”?

            If so, there’s even more bad patents out there than we thought.

            1. 6.2.4.1.2.1

              a machine which screens a giant library of compounds for a function is also an “inventor”?

              Screening FOR a function is such a pretty strawman – as the words imply that the object of the screening is not the choice of the machine – in other words, the machine is ONLY a tool — this is just NOT what AI is about (and a purposeful obfuscation).

          2. 6.2.4.1.3

            “ Greg deLassus on these boards has good knowledge of the use of AI in the pharma field”

            LOL We all have “good” knowledge of the use of computers in all kinds of disciplines. That kind of “expertise” has no bearing on the issue here which is simply whether a machine can legally be named an “inventor” in the US patent system.

            Here’s a hint: no.

            1. 6.2.4.1.3.1

              What do you think it takes to be legally named as an inventor?

              (You do realize that this is different than actually being awarded a grant of a patent – much like a Simian can act and capture an expression in a fixed media, but cannot be awarded a copyright on that fixed creative expression, eh?)

        2. 6.2.4.2

          Mark,

          As you are the one making the assertion (“What are the consequences if all of this effort is being expended based upon a
          L
          I
          E?
          “), you would have the burden to show that such IS a
          L
          I
          E.

          Charges of fraud do carry a higher burden on the accuser.

          If shown to be true, Fraud on the Office carries known consequences.

          But what I find “amusing” is the trenchant Zeal to even avoid the legal issues.

          If this case turns to Fraud, would the next? Or the next after that?

          We are ALL better off facing these issues now.

          1. 6.2.4.2.1

            …and Mark disappears again…

            Yet another who for whatever reason just does not want to address a very specific legal issue.

        3. 6.2.4.3

          Mark, Thaler’s assertions raise credibility issues, to say the least.

          Patent lawyers should rely on AI experts for advice on that, as Rose Hughes rightly pointed out in her post “An AI inventor or the Emperor’s New Clothes” published on IPKat on Sept 15, 2021.

          Another telling bit to the credibility issue is the specification of Thaler’s application 16/524,350. It was not published but you can look at the equivalent EP 3563896, para 19-21.

  6. 5

    Whichever actual human(s) wrote the source code of the AI should be named as the inventor of whatever the AI’s code “created.” Simple.

      1. 5.1.1

        Simply wrong — those writing the source code
        F
        A
        I
        L
        to satisfy the legal requirement for being named inventors.

        This is not something that can — or should — be made “merely convenient” to be some none-inventive (but tangentially related) person “AS Inventor.”

    1. 5.2

      Does this remain true forever? If the AI keeps running after the code-author’s death, should we continue to list the code-author as inventor even then? Mr. K—the code author—should be said to have continued conceiving of new inventions long after he has died?

      1. 5.2.1

        Death of an inventor is irrelevant after the patent application is filed.

        Moreover, in the genetic engineering space, we don’t know the new organism as the inventor of itself. Take for example putting in bacterial genes into plant DNA to make herbicide resistant plants. At least in the old days, the people doing it used quasi-random procedures that put in the foreign DNA by chance (most of their experiments failed). Precisely WHERE the foreign DNA attaches into the plant’s genome was not know or intended by the humans. And that location (called a genetic “event”) has a profound impact on whether and how much herbicide resistance the chimeric plant cell would possess–it is up to the organisms and not the humans. But the humans are the inventors. Not a perfect analogy but has similarities to the AI context where the people that programmed the AI could be inventors. If they are NOT inventors, then nobody is.

        1. 5.2.1.1

          Mycogen Plant Science, Inc. v. Monsanto Co., 61 F. Supp. 2d 199, 227 (D. Del. 1999):

          According to Mackey, the gene gun functions like a shotgun in that a small gunpowder charge is used to shoot tiny tungsten pellets into corn cells. Prior to shooting, these pellets are coated with genes. Mackey testified that using the gene gun is a “pretty crude” process because “[y]ou are literally just shooting genes into these cells and hoping something happens.” Most cells and genes do not survive the shooting.

          According to Mackey, after shooting genes into the cell clumps, she and her colleagues found survivors in the form of successful gene transplants in the cell clumps. They worked on getting the gene transplant into a corn seed, as opposed to a cell clump, because corn seed is the commercial end product; it is what farmers want to purchase for planting. Mackey testified that, in order to do this, DeKalb scientists began to grow a corn plant from the cell clump. This corn plant initially grew but began to die before it reached maturity. DeKalb scientists performed a process “like a C section” on the plant to save the little seed that was in the dying corn plant’s ear. They then used this seed to grow a plant which they nicknamed “Henry.”

          Mackey testified that “Henry” was the first plant into which she had successfully transferred genes. On December 22, 1989, she tested “Henry” and found the gene present in the plant. On this date, she knew she had made a fertile transgenic corn plant. Mackey testified that DeKalb is “selling descendants of Henry today.”

          According to Mackey, after shooting genes into the cell clumps, she and her colleagues found survivors in the form of successful gene transplants in the cell clumps. They worked on getting the gene transplant into a corn seed, as opposed to a cell clump, because corn seed is the commercial end product; it is what farmers want to purchase for planting. Mackey testified that, in order to do this, DeKalb scientists began to grow a corn plant from the cell clump. This corn plant initially grew but began to die before it reached maturity. DeKalb scientists performed a process “like a C section” on the plant to save the little seed that was in the dying corn plant’s ear. They then used this seed to grow a plant which they nicknamed “Henry.”

          Mackey testified that “Henry” was the first plant into which she had successfully transferred genes. On December 22, 1989, she tested “Henry” and found the gene present in the plant. On this date, she knew she had made a fertile transgenic corn plant. Mackey testified that DeKalb is “selling descendants of Henry today.””

          1. 5.2.1.1.1

            Your (extensive) citation does NOT support your assertion, by the way.

            What your assertion stands for is that the “process” can well be a haphazard one. The inventors though did NOT stop at a haphazard step. They continued their inventive efforts UNTIL they had a legal invention, and thus were named legal inventors.

        2. 5.2.1.2

          Death of an inventor is irrelevant after the patent application is filed.

          I suppose. Is the death of the inventor before the filing of the application also irrelevant? Is it irrelevant that the inventor has died before the claimed invention was even conceived?

        3. 5.2.1.3

          “Precisely WHERE the foreign DNA attaches into the plant’s genome was not known or intended by the humans.”

          The more important point for this conversation is that the DNA being inserted has known structure and if you know that the DNA is inserted into the organism then you can structurally distinguish that modified organism from the prior art organism (and that distinguishing structure needs to be recited in the claim). This sometimes adds hundreds of “pages” to the specification! And patent offices around the world created a entire separate set of rules and requirements so that the structural info could be tracked, searched, examined and managed. Ask a grown up about it.

          1. 5.2.1.3.1

            Yes the DNA of the combined organism can be sequenced (after it exists) and the humans can figure out where the exogenous DNA has inserted into the host genome. But they could not have (and did not) conceived of that before the event was created. The humans had no idea where (or if) the foreign DNA would insert. But the humans are still named inventors.

            Here with AI, the human guy that coded the AI didn’t know in advance what the AI would do. But once the AI “invents” something, it can be written down and described in the patent application–just like in genetic engineering. The human is thus the inventor. If not, NOBODY is the inventor under the USA patent laws.

            1. 5.2.1.3.1.1

              The alternative of “then nobody” is simply the better choice – especially given an admission that no human rises to the actual legal requirement for being an inventor.

              And of course, inventorship is only one of the issues that would flow from recognizing non-human innovation.

              Another would be the direct impact on another non-human legal fiction: that of what a Person Having Ordinary Skill In The Art would know (need I remind you that this legal fiction is meant to reflect State of Art, rather than real person inventorship, and thus is impacted regardless of the “no one gets a patent” outcome)….

            2. 5.2.1.3.1.2

              “ The humans had no idea where (or if) the foreign DNA would insert. But the humans are still named inventors.”

              Wrong. If they never determine that the foreign DNA was inserted then they are not yet inventors of anything. A research plan is not an invention.

          2. 5.2.1.3.2

            “Prophet” – you seem to think that just because something can be described that this alone determines inventorship. Apples and oranges my grown up friend. We should stay away from irrelevant ad hominem attack and just make a point or stay quiet.

            1. 5.2.1.3.2.1

              “ you seem to think that just because something can be described that this alone determines inventorship.”

              Wrong again. The issue is that if you can’t describe what you’ve allegedly invented in a non-functional manner that distinguishes your article of manufacture from the art, then you are not an “inventor” of anything patentable.

              Again: ask a grown up. And this isnt about “picture claims”. This is about describing a structure in a way that that distinguishes the structure from the art in a physical, non-functional manner.

              The CAFC dealt with this head on in the context of product-by-process claiming. You can’t do an end-around the requirements for claiming a new article or machine by relying on function or method of manufacture. You have to know what you allegedly invented before you can claim it. You at least have to know the structural differences between what you invented and the art.

              1. 5.2.1.3.2.1.1

                LOL – you are not correct in the law on that Windmill chase of yours of “describe in non-functional

                And we both know that, eh?

                And your head has been handed to you by David Stein (among many others) across MANY examples of even non-computing arts.

                But please, continue to “be Malcolm.”

          3. 5.2.1.3.3

            Lol – and yet, the so-called “grown-up” arts are not predictable and MAY thus need exacting picture claims.

            You may have to discuss this with a grown-up with at least a modicum of innovation understanding (and can recognize why detailed “you may patent this” ‘thinking’ (using that term VERY loosely)) — that understanding including the Kondratiev waves of innovation.

            Unless you like being stuck with marginal-advanced exacting picture claims that generate less than marginal patent protection and don’t really care about the direction of Real innovation.

            1. 5.2.1.3.3.1

              The issue is not claim scope. The issue is that in the genetic engineering examples I listed, like the AI context at issue in the original post, the human inventor did not conceive of the invention before it existed. In the former case, the biological cells (coupled with the random nature of the gene gun) determined WHERE and IF the foreign DNA attaches. That, in turn, dictated whether the chimeric plant has robust herbicide resistance. In the latter case, the “AI invented” something based on the programming from the human being. Thus, the human is the inventor or there is not one (and hence there can be no USA patent protection on the alleged invention).

              1. 5.2.1.3.3.1.1

                Above is why the doctrine of “simultaneous” conception and reduction to practice exists. It could easily apply to AI inventions. (And again, otherwise the AI invention rightfully will get zero patent coverage in the USA).

                Back to my grown up real job.

                1. Arguably, here it is reduction to practice, then conception – by a human who sees the output of the AI as an invention, and is hence the inventor. AI isn’t the one telling the patent attorneys what the invention may be.

                2. But (though – now having gone through your above comment) – I note that your above comments does NOT reach or support your analogy to AI.

                  See above.

                3. Hitting the Count Filter…

                  Your comment is awaiting moderation.

                  June 1, 2022 at 9:05 am

                  agent,

                  Your message was being processed during my replies above.

                  As you may (merely) want – and you are not alone — the mere act of reading someone else’s invention does not make that person a legal inventor.

                  I have covered this many times now in the details of the multiple rooms and single black box (with optional world wide streaming capability).

                  Further, it is in fact asserted that DABUS did in fact “tell” the patent attorneys what the invention was.

              2. 5.2.1.3.3.1.2

                That some bio (or Pharma, or combination) patents are granted without possession (or inventorship) at time of filing is a bug — not a feature.

    2. 5.3

      Does anyone know if this is how it works in the copyright space? Do we list as “author” the person who writes the code that creates a design? If so, do we continue to list that person as author even after they die?

      1. 5.3.1

        Greg – as you also (on a same thread as me) provided the Simian Selfie case, maybe someone can remind you of that post so that your question can be asked again (as needed)…

    3. 5.4

      “ Whichever actual human(s) wrote the source code of the AI should be named as the inventor of whatever the AI’s code “created.” Simple.”

      Wrong is more like it. Absurdly wrong. The inventor needs to have a fixed and permanent idea of what the invention is in order to have conceived of it. After that, the invention needs to be described in a non-functional matter that distinguishes it from the prior art. If the invention is an article of manufacture, that means the new physical structure of the machine needs to be described in a manner that distinguishes it from the art.

      Good luck with that.

      Oh and by the way: innovation in computing will continue just fine without patent protection. Bank on it. But the worst attorneys in the world will cry and whine as they always do.

      1. 5.4.1

        needs to be described in a non-functional matter

        Collides with:

        Wrong is more like it. Absurdly wrong.

        Unintended irony at its finest.

  7. 4

    As to… :

    and refers to an “inventor” using personal pronouns

    I do not believe that this “buys you anything” — the legal fiction (non-real person) of the Person Having Ordinary Skill In The Art is also referred to using personal pronouns.

    That being said (at least in the US), the issue (one issue – to be more precise, as other issues DO abound) is resolved by looking at the Lockean nature of patent law.

    See also (again) the Simian Selfie case.

  8. 3

    The grifters and con artists promoting “AI” also tried to get everyone to believe that YEARS AGO we’d all be buying cars that could drive themselves across the country. And truckers would be out of jobs! How’d that turn out?

    The only “innovation” that makes a genuine difference to scammers like Thaler is radical, short-sighted judge-made changes in the law that provide grifters with new soft targets to exploit at every one else’s expense.

    The entire US system is still struggling to coherently deal with logic claiming … so let’s let computers file their own patents? LOL What a c e s s p o o l.

  9. 2

    The most just result of all this is that Thaler and his attorneys end up sanctioned and preferably disbarred for a while.

    1. 2.1

      Prophet, I assume that your comment was intended to be frivolous. In the UK, it was accepted that the AI devised the claimed subject matter. In Germany, the law requires a human to be named as inventor. That problem was overcome when the patent court in Munich accepted a Declaration of Inventorship in the form (thanks to James Nurton on another patent law blog):

      “Stephen L. Thaler, PhD who prompted the artificial intelligence DABUS to create the invention”.

      1. 2.1.1

        News flash: the US has different patent laws than the UK and Germany.

        Also, those UK and German patents are junk.

        1. 2.1.1.1

          Junk patents? Not sure about that, Prophet, because, as far as I understand it, the DABUS patent applications in UK and Germany have yet to be examined on the substantive merits, if any, of their patentability. The court proceedings are to resolve the question whether Applicant has timeously filed an acceptable form of declaration of inventorship (a filing formality).

  10. 1

    Computers can’t apply for patents and can’t be named as inventors because they aren’t people.

    But sure let’s waste thousands of dollars and Federal Court resources “debating” an inanity.

Comments are closed.