A Failure of Fact: What Has Been Missing from the Arguments in Thaler v. Vidal

“The discussions [thus far] leave a reader in the dark as to what exactly DABUS is and how DABUS went about conceiving and reducing to practice the particular container shape or light stick that are the subject of its patent applications.”

Whether or not the law recognizes a machine as the inventor-at-law, the facts are indispensable to determination of the true inventor-in-fact. In the case of Stephen Thaler’s attempt to obtain patent protection for a food container and light stick he says were independently invented by his AI machine, DABUS, the inventor-in-fact will be either Thaler or his machine. The procedural posture of Thaler v. Vidal caused the discourse to jump over the facts of how the food container and the light stick were invented by DABUS. These overlooked facts may reveal the true inventor, regardless of whether or not the type of inventor is recognized by the current law. 

Thaler v. Vidal is an appeal to the U.S. Court of Appeals for the Federal Circuit from a summary judgment by the United States District Court for the Eastern District of Virginia in favor of a decision by the U.S. Patent and Trademark Office (USPTO) that had denied Stephen Thaler’s patent applications because they failed to list any human as an inventor.

Give Us the Facts

Thaler’s brief relies on the following premise: “As noted by the district court, and never challenged on appeal by the USPTO, all factual statements in the patent application including the capability of the AI to create a patentable invention must be taken as true.” The USPTO did not evaluate the factual issues regarding inventorship beyond the proclaimed non-human nature of the inventor. At the district court, the standard for grant of summary judgment is to resolve all disputes of material fact in favor of the non-moving party. As a result, when a case is decided on summary judgment, the truth or falsity of the material facts is not evaluated. The district court is merely saying: even if we assume everything you say is true, you still lose, so we don’t need to get to the facts to rule against you. It does not mean that the district court actually determined that Thaler’s factual assertions were true.

Thaler has argued that the opinion of the Federal Circuit “conflicts with 35 U.S.C. 103, which prohibits denying patent application [sic] based on how inventions are made, and contravenes Diamond v. Chakrabarty, 447 U.S. 303, 315-317 (1980) in which the Supreme Court held that a categorical rule denying patent prosecution for ‘inventions in areas not contemplated by Congress … would frustrate the purpose of the patent law.’”  Neither argument is relevant to the issue of inventorship. 35 U.S.C. 103 pertains to obviousness and Chakrabarty to statutory categories of patentable subject matter, namely 35 U.S.C. 101. Whether the claims of Thaler are novel and nonobvious has not been evaluated by the USPTO and the current claims appear to fall within the statutory categories.

While a 35 U.S.C. 103 standard is inapposite to the inventorship issue, demonstrating “how the invention was made,” in this particular situation, may go a long way towards resolving the question of machine or man. Taking readers by the hand and guiding them through the process of how the machine conceived and reduced its purported “inventions” to practice could illuminate the concept of inventorship by a machine.

Thaler is now contemplating a petition for certiorari to the U.S. Supreme Court.) Facts, showing how a machine can invent, that were neither provided by the applicant nor evaluated by the USPTO or the other courts that ruled on the case, are precisely what is needed for establishing whether a machine can invent. When readers are convinced that the machine actually and truly did invent, it is only a short skip and hop to persuading them that the law must be interpreted or changed to identify the true inventor. In the next brief, perhaps, such facts will be given a role.

Thaler filed two patent applications at the USPTO, as well as a number of other patent offices. The version filed under the Patent Cooperation Treaty (PCT) combines the two applications and is published as WO 2020/079499 corresponding to PCT/IB2019057809.. One of the applications is directed to a “Food Container” or “Fractal Container,” such as a water bottle or milk carton, with an external wall pattern that helps two adjacent containers interlock and the other to a “Neural Flame” which is a beacon that emits light in a fractal pattern.

The independent claims provide:

Thaler asserted his “AI machine,” which he calls “Device for Autonomous Bootstrapping of Unified Sentience” abbreviated to DABUS, as the sole inventor of the inventions that are the subject of his patent applications.

This assertion of inventorship caused the parallel proceedings in the different countries to focus only on one question: whether AI can be an inventor under the laws of these countries? The arguments and opinions revolve around the history of the law and the meaning of certain terms such as “person” and “individual” in the statutes. Accordingly, the discussions leave a reader in the dark as to what exactly DABUS is and how DABUS went about conceiving and reducing to practice the particular container shape or light stick that are the subject of its patent applications.

We Need More

In the appeal to the Federal Circuit the question presented was “to decide if an artificial intelligence (AI) software system can be listed as the inventor on a patent application.” The Federal Circuit stated: “In July 2019, Thaler sought patent protection for two of DABUS’ putative inventions by filing two patent applications with the PTO: U.S. Application Nos. 16/524,350 (teaching a “Neural Flame”) and 16/524,532 (teaching a “Fractal Container”). He listed DABUS as the sole inventor on both applications. Thaler maintains that he did not contribute to the conception of these inventions and that any person having skill in the art could have taken DABUS’ output and reduced the ideas in the applications to practice.” (footnotes omitted.)

Footnote 2 of the above passage reasons: “While inventorship involves underlying questions of fact, see Dana-Farber Cancer Inst., Inc. v. Ono Pharm. Co., 964 F.3d 1365, 1370 (Fed. Cir. 2020), cert. denied, 141 S. Ct. 2691 (2021), for purposes of this litigation the PTO has not challenged Thaler’s representations, see D. Ct. Dkt. No. 25, at 11. Accordingly, our analysis must be consistent with the undisputed facts in the administrative record, drawing inferences in favor of the non-moving party. See Safeguard Base Operations, LLC v. United States, 989 F.3d 1326, 1349 (Fed. Cir. 2021) (discussing when it is appropriate to supplement administrative record and noting “[t]he focal point for judicial review should be the administrative record already in existence”) (quoting Camp v. Pitts, 411 U.S. 138, 142 (1973)).” Footnote 2, therefore, dismisses delving into any questions of fact regarding the purported inventions.

Thaler has made a conclusory assertion that “he did not contribute to the conception of these” but, because his inventor can be turned off by a switch, a bit more is required to convince one of DABUS’ autonomy.

Under one common scenario, a human engineer works for a technology company; he talks to his coworkers and attends conferences; he may arrive at an idea about improving a process or a machine that is used in his area. If he “conceives” an “invention” and “reduces the invention to practice,” either actually or constructively, then he writes an Invention Disclosure and submits it to the intellectual property department of his company where the Invention Disclosure is converted to a Patent Application and submitted to the USPTO for examination. The engineer uses computers and may use off-the-shelf software or write his own specific computer programs; he chooses which type of data to collect or use for his particular purpose. Further, the inventor reviews the patent application before he signs the oath.

How does the above chain of events map to the process of invention by DABUS and preparation of the patent application by Thaler? Did DABUS have any input to the claiming of its invention? What transpired between Thaler and DABUS that prompted DABUS to “invent” in the first place? Which part of the claims does DABUS consider as its invention?

Along the same lines: WHAT is DABUS? What is an “AI MACHINE”? A “Device for Autonomous Bootstrapping of Unified Sentience” tells very little. Is it a general-purpose computer running software programs that Thaler wrote?

A partial response to the question that is posed above is provided in the USPTO petition decision: “Petitioner asserts the invention of the above-identified application was generated by a machine named ‘DABUS.’ According to petitioner, this ‘creativity machine’ is programmed as a series of neural networks that have been trained with general information in the field of endeavor to independently create the invention. Petitioner asserts that DABUS was not created to solve any particular problem, and it was not trained on any special data relevant to the instant invention. Instead, it was the machine, not a person, which recognized the novelty and salience of the instant invention. Petitioner contends that inventorship should not be limited to natural persons and, therefore, the naming of DABUS as the inventor in the above-identified application is proper.”

The above characterization of DABUS serves as a mere introduction. What does “general information in the field of endeavor” mean? What was the “field of endeavor” and what is considered “general information” on which DABUS was trained? Can DABUS invent a baby formula for the milk in the container as well? Presuming that DABUS is a computer program, who wrote the program? Who selected the data provided to the program? What questions did the human operator, i.e. Thaler, ask of the computer program? Both purported “inventions” pertain to “fractal patterns.” One application asks that the outer body of a container have a fractal pattern shape and the other asks that a light beam emit light in a fractal pattern. Did Thaler use “fractal patterns” as an input and write a software to come up with articles of manufacture that could include “fractal patterns” as an element? Is this the basis of Thaler’s assertion that DABUS is the sole inventor and stating that he, Thaler, had “no” contribution to the invention? These are examples of facts that could be added.

Distinguishing DABUS

The input to the program and the output generated by the program that prompted Thaler to file a patent application demonstrate whether Thaler did or did not invent the subject of the applications. For example, if Thaler selected a specific machine trained on a specific set of data relevant to the invention and the machine generated choices, based on the input parameters, then the machine is a decision support system. See U.S. 5740328 which was arrived at by the help of a decision support process and publications that describe the process . Is DABUS a decision support program? If not, how does it differ from one?

Also note that the claims were not examined substantively and there was no determination of “novelty and salience of the instant invention.” To date, several X references have been cited by the various patent offices. An X reference is a “document indicating lack of novelty or inventive step,” and is the equivalent of a 35 U.S.C. 102 reference under U.S. law such that the claim is not considered novel and nonobvious in view of an X reference. See for example, the cited U.S. 5803301 titled “Seamless Can and Process for Making the Same” () and “Snowflake Tealight Holder” ().

In mathematics and engineering classes, the students learn by solving example problems. Thaler could use DABUS’ invention as an example of “AI Inventorship” and teach by leading the reader through the facts of how DABUS came to invent its “Fractal Container” or “Neural Flame.” Being led step by step through the facts of these “inventions” may prove more convincing than discussing the dictionary meaning of the English words or policy considerations favoring AI inventorship.

This article initially omitted the co-authors in error.

Image Source: Deposit Photos
Image ID: 408527856
Author: DmitriySk


Warning & Disclaimer: The pages, articles and comments on IPWatchdog.com 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 IPWatchdog.com.

Join the Discussion

12 comments so far.

  • [Avatar for Anon]
    March 23, 2023 12:17 pm


    That is not how the law works.

    Stop following those jumping into the weeds.

  • [Avatar for Addy]
    March 23, 2023 09:27 am

    Let’s have Thaler show facts supporting DABUS’ unaided conception of the invention and the subsequent reduction to practice.

    Then we can have a philosophical discussion of whether or not an AI can be an inventor.

  • [Avatar for Anon]
    March 23, 2023 07:06 am


    I have to disagree.

    There is a clear position here that the delegation of authority to Congress is not in fact unlimited.

    That delegation (at our Nation’s founding) comes with the Lockean nature and understanding of the words chosen for that delegation.

    The foundational aspect of Quid Pro Quo is to turn a real person’s inchoate right (a real person’s possession) into a basket of real personal property legal rights.

    This foundation is at a Constitutional level.

    Could Congress ‘just change the words of law?

    Perhaps — but THAT action would NOT be merely your standard legislative action, but instead would need be a Constitutional amendment going to the delegation of authority.

  • [Avatar for PeteMoss]
    March 22, 2023 04:43 pm

    This entire debate was settled in the Dean Koontz DEMON SEED movie. Proteus IV, an AI program, creates a cure for leukemia. Others ask the creator, “You’re going to patent [the cure], aren’t you?” Not, “Proteus IV is going to patent the cure, isn’t it?”
    I disagree with Koontz, however. This is my test: who transformed a blank piece of paper into a useful, novel and non-obvious invention? Proteus IV & DABUS, that’s who. The law on inventorship is just words written on a page. It is not an unchangable law, like the speed of light. Just re-write the statute and adding DABUS complies with “the law.”

  • [Avatar for Addy]
    March 20, 2023 09:05 am

    Thanks for raising the questions that the courts should have asked regarding the role of DABUS in the inventing process. Is an AI/ML (Artificial Intelligence/Machine Learning) model that has been specifically trained on a set of parameterized data to find a local optimal solution an “inventor” of the output, or a programmed tool (or decision support system)?

    DABUS didn’t select a problem to be solved, the final solution chosen (it likely came up with many), the data it was trained on, the parameters applied to any element of the training data, the model trained with the data, adjusting model parameters, etc. It’s unclear how that system could be an inventor.

    I do think that someday AI/ML can be an inventor, but we’re not quite there yet.

  • [Avatar for Curious]
    March 18, 2023 04:24 pm

    The fact as [alleged] is that Thaler IS NOT the inventor-in-fact.
    Fixed it for you.

    Facts are first alleged. Facts can then be shown to be (legally) established by certain burdens of proof.

    The law is clear the inventor is required to be a human being in order to obtain a US patent. Could other entities be inventors? Tools have been used by other types of animals, and consequently, humans may not be the only inventors. Regardless, if a dolphin invents something new, for example, they cannot obtain a patent therefor in the US — no matter if the invention is both novel and non-obvious.

    That, however, is a different question than determining whether DABUS, in FACT, was the inventor, or whether DABUS, in FACT, was a joint inventor, or whether Thaler, in FACT, was the sole inventor.

    There are some of us who question whether DABUS, in fact, could have been a sole inventor or joint inventor. In other words, there are some that question the credibility of any allegation that DABUS was the inventor.

  • [Avatar for Mark Summerfield]
    Mark Summerfield
    March 18, 2023 02:55 am

    When a court is presented with a case involving stipulated facts that have been agreed by all parties, it has no choice but to render a decision based on a presumption that the stipulated facts are true. This is clearly not the same thing as actually being presented with evidence.

    In Australia, the High Court denied Thaler’s application for special leave to appeal (analogous to a petition to the SCOTUS for cert) essentially because the Court did not see sufficient value in deciding a case in which it could not interrogate the factual basis for inventorship. It was right to do so. It is not the place of an applicant to predetermine inventorship and press the nation’s top Court for a binding decision based on what might well be a hypothetical proposition. And with the Commissioner of Patents accepting the agreed facts, there would have been no contradictor. Inventorship is a question of law, and the novel proposition that a machine should be named as an inventor requires the Court to be in a position to evaluate the factual circumstances in order to determine how the law should properly be applied.

    Unlike the Federal Circuit, SCOTUS has a choice as to whether or not it takes on a case that may or may not be grounded in reality. What are the odds it follows the same course as the Australian High Court?

  • [Avatar for Anon]
    March 17, 2023 10:53 am

    but, because his inventor can be turned off by a switch, a bit more is required to convince one of DABUS’ autonomy.

    This statement is a legal non-sequitur.

  • [Avatar for Anon]
    March 17, 2023 10:51 am

    the inventor-in-fact will be either Thaler or his machine.

    My immediate reply is that this cannot be correct.

    The fact as presented is that Thaler IS NOT the inventor-in-fact.

  • [Avatar for Eileen McDermott]
    Eileen McDermott
    March 17, 2023 09:32 am

    That was missed on my end, thanks for noting – it’s been updated.

  • [Avatar for Pat]
    March 16, 2023 05:13 pm

    The author has mangled the name of the technology, as well as other facts. It’s “Device for the Autonomous Bootstrapping of Unified Sentience,” not Science. See https://thenewstack.io/stephen-thaler-claims-hes-built-a-sentient-ai/ for a much more accurate description of it.