Comment USPTO RFC Regarding "Impact of the Proliferation of Artificial Intelligence on Prior Art, the Knowledge of a Person"

I intend to submit the following comments to this RFC. Let me know your feedback and if you would like to sign on.

?

Comments in Response to USPTO Request for Comments

Title

Impact of the Proliferation of Artificial Intelligence on Prior Art, the Knowledge of a Person Having Ordinary Skill in the Art, and Determinations of Patentability Made in View of the Foregoing

Document ID

PTO-P-2023-0044-0001

Comment Due Date

Jul 29, 2024

Submission Date

?

FR Citation

2024-08969

?

?

The Honorable Kathi Vidal?

Under Secretary of Commerce for Intellectual Property and Director?

U.S. Patent and Trademark Office?

600 Dulany St.?

Alexandria, VA 22314?

?

Via www.regulations.gov

Dear Director Vidal:

??????????? I am submitting the following comments and suggestions in response to the USPTO’s “Request for Comments on Impact of the Proliferation of Artificial Intelligence on Prior Art, the Knowledge of a Person Having Ordinary Skill in the Art, and Determinations of Patentability Made in View of the Foregoing” (RFC).? My comments will focus on the impact of AI on Design Patents.? These comments may also apply to Utility and Plant Patents, but they were prepared specifically for Design Patents.?

??????????? As background, I am a Registered Patent Attorney.? I personally filed and prosecuted over 1,200 US design patents, virtually all in the automotive space.? I have worked with automotive designers for over 20 years.? I litigated design patents before the International Trade Commission and the Federal Circuit.? I testified before Congress regarding the proposed Parts Act that would have eviscerate design protection for replacement components.? When I joined USPTO as Regional Director of the Elijah J. McCoy US Patent and Trademark Office, I met with designers around the country to explain design patents and understand their concerns.?

??????????? I expect the Office will receive hundreds of comments on this RFC.? Few, if any, will specifically address design patents.? Generative Artificial Intelligence (AI) will impact design patents in three ways.?

??????????? AI allows computers to self-generate designs as directed by a human.? Designers have been using AI to speed the inventive process for many years.? Automotive designers of parts with thousands of small features would create the basic pattern, say a hexagonal opening in a speaker grille or graining on vinyl film, and AI tools would fill the opening or sheet with repeating evenly spaced elements terminating around the perimeter with each element sized proportionally to provide a balanced visual appearance.?

??????????? AI allows users to easily generate billions of design variations based on an existing design.? This could be the originator that wants to maximize their freedom to operate, or a third party that wants to block others from patenting future iterations.

??????????? AI will be used by parties seeking to invalidate a design patent by finding each claimed design element in the near limitless prior art created by humans and AI tools.? AI will find every curve, angle, line, depression, radius, texture, and other feature in prior databases and construct a prima facia case on obviousness that prospective patentees must overcome. ?

??????????? Before I address the specific questions in the RFC, I would like to comment on the challenges design patent owners now face due to the Federal Circuit imposed KSR/Graham analysis to determine the nonobviousness of a design (LQK v. GM, 2021-2348, May 21, 2024).? The Federal Circuit significantly altered the landscape of design patent law by rejecting the longstanding Rosen-Durling test for determining obviousness. This test, which had been in place for decades, required a two-pronged approach: first identifying a primary reference that is “basically the same” as the claimed design, and then allowing modifications by secondary references if they were sufficiently related.

??????????? The Federal Circuit found this approach overly rigid and inconsistent with the Supreme Court’s decision in KSR International Co. v. Teleflex Inc. (2007). Instead, the Federal Circuit decided that the obviousness of design patents should be evaluated using the Graham factors used for utility patents, which include:

1.???????? The scope and content of the prior art,

2.???????? The differences between the prior art and the claims at issue,

3.???????? The level of ordinary skill in the pertinent art,

4.???????? Secondary considerations such as commercial success and long-felt but unsolved needs.

This decision means that design patent owners will face greater difficulty in enforcing a design patent.? Computers have constructed vast databases of prior art and made them free to everyone.? An automotive designer has access to every fender design (the patent at issue in the LKQ case) since antiquities.? AI tools can identify every difference between the claimed design and the prior art and look to find these differences in other references.? These computers will then formulate a “ransom note” argument that all of the claimed features are found in the various references by piecing together bits of fenders from the prior art.? Yes, this is the definition of hindsight reconstruction, but it now becomes a burden on the applicant to overcome this constructed invalidity argument.? To overcome hindsight, potential infringers will proactively generate billions of designs before the application to seed their later arguments.

??????????? The most thorough review of obviousness in design patents is found in “Overcoming the ‘Impossible Issue’ of Nonobviousness in Design Patent” Janice Mueller and Daniel Brean, Kentucky Law Journal, Vol. 99, No. 3, p. 419, 2011, U. of Pittsburgh Legal Studies Research Paper No. 2009-30.? This article reviews the history of US design patents and the mandate that the design be “nonobvious” in view of the prior art.? This is exactly the issue in the LKQ v. GM and one that designers, attorneys, examiners, courts, and juries struggle to understand.? After more than 40 years of practicing in this area, I believe that nonobviousness has no application to design.? This is especially true now that AI can generate billions of designs at the click of a button.? USPTO should assist Congress in updating 35 USC 171 to remove the nonobvious element.?

??????????? I have attached an article I wrote on this topic called “A New Era for Design Patents: Replacing Nonobviousness with Commercial Availability.”? I recommend that USPTO study the costs and benefits if the US were to eliminate the requirement that design patents are nonobvious and instead include a trademark-like requirement that the design be made commercially available.? Design patents without underlying commercial products provide the public with no benefit other than being a free source of drawings for cheap posters used in restaurant bathrooms.? Designers do not study issued design patents to garner inspiration as if they were the Torah. Jason Du Mont and Mark Janis studied the origins of US design patents in their 2013 paper “The Origins of American Design Patent Protection.”? Their paper documented that technological advancements in iron casting allowed for rapid copying of US products by foreign producers and spurred Congress to enact design protection.? The US-manufactured articles were the intended articles of protection, not the artwork of the drawings.?

??????????? The other change that the USPTO should consider is defining the word “original” in 35 USC 1.171 to be ascribed to the assignee and not to the inventor.? This has two benefits; it allows the manufacturer to retain a staff of designers that are tasked with extending and refreshing a product’s brand identity.? The manufacturer will likely produce many iterations of a design and extend the brand through continuously refreshed articles.? Imagine the 2107 iPhone 100.? “Original” should not preclude teams of designers working to extend a single corporate design theme.? “Original” should also discourage competitors from protecting designs too similar to those of their competitors.? The Federal Circuit achieved the correct result in International Seaway Trading v. Walgreens, 589 F.3d 1233, by ignoring basic anticipation law.? Perry Saidman describes in detail the Federal Circuit’s flawed analysis in his paper “Design Patents Are Sinking in International Seaway: Rethinking Design Patent Anticipation”.? The decision precluded International Seaway from patenting a Crocs-like design for anticipation would also preclude Crocs from protecting new designs.? Rather than using the International Seaway anticipation test, “original” should be defined to provide greater leeway for the creative enterprise to make incremental changes to a design that retains the consumer-recognized brand theme.? Interlopers that did not create this brand theme should be challenged as not “original”.?

??????????? The remainder of my comments will address your questions.?

?

A. The Impact of AI on Prior Art

1. In what manner, if any, does 35 U.S.C. 102 presume or require that a prior art disclosure be authored and/or published by humans? In what manner, if any, does non-human authorship of a disclosure affect its availability as prior art under 35 U.S.C. 102?

??????????? Ben Hattenbach & Joshua Glucoft directed address this question in 2015 but for utility patents in their article “Patents in an Era of Infinite Monkeys and Artificial Intelligence,” 19 Stan. Tech. L. Rev. 32, 42 (2015).? “On one end of the spectrum, if a computer generated a focused set of high-quality variations on claim language, then it would be easier to justify folding such knowledge into the scope of the prior art. At the other end of the spectrum, if a computer published millions of variations of claims such that all but a few were useless from a technical or grammatical perspective, then it would be easier to justify not requiring inventors to account for that sea of information.”? The authors suggest that focused, findable, and usable computer-generated art should be in the scope of prior art.? If USPTO and Congress accept this reasoning for utility prior art, they should exclude AI-generated art for design patents.? As described in my earlier publications, I found no case, article, testimony, or other discussion of design inventors using non-produced prior publications in their creative process.? There is no commercial insight into the consumer appeal of a non-produced design.? Designers see what sells.? Designers use the success of a manufactured article as a measure of public taste.? They combine this success with anything and everything they see around them to envision a new design for a future customer whose tastes will have inventively evolved.?

??????????? Lucas R. Yordy, The Library of Babel for Prior Art: Using Artificial Intelligence to Mass Produce Prior Art in Patent Law, 74 Vanderbilt Law Review 521 (2021) considers AI generated disclosures as possible prior art in utility patents.? “This Note argues that courts should adapt to AI-generated prior art by eliminating the existing presumption that prior art is enabling and by implementing a conception requirement for prior art.”? Whether or not that framework works for utility patents is unknown, but it is completely unworkable for design patents.? Computer-generated images are as detailed or more detailed than USPTO-compliant design patent drawings.? Unless the USPTO updated its drawing requirements to include 3d data, virtually all AI-generated images would be considered enabling.?

??????????? AI-generated art should not remove otherwise protectable designs from the universe of protectable works. That would discourage creators and encourage nefarious actors to preemptively clutter the public prior art with designs that had no chance of becoming available for consumers and needlessly adding cost and risk to genuine producers.? Moreover, allowing AI-generated designs to dominate the prior art landscape could stifle innovation and creativity. Designers might be deterred from pursuing their ideas, fearing that their unique creations could be rendered unprotectable due to the proliferation of AI-generated works. This could lead to a decline in the quality and diversity of new designs entering the market.

??????????? Furthermore, the potential for misuse of AI in generating prior art poses a significant threat to the integrity of the design and patent system. Unscrupulous entities could flood the system with low-quality, AI-generated designs to create a defensive shield against legitimate innovation. This tactic would not only burden the patent examination process but also unfairly disadvantage genuine creators who invest time and resources into developing original works.

??????????? The concern that computers could preemptively generate designs to preclude later protection by creators is not an abstract fear removed from reality.? Consider the case of Damien Riehl’s and Noah Rubin’s "All the Music0" project launched in January 2020. They publicly announced their initiative to generate and place every possible melody into the public domain to challenge the concept of melody copyright. This project aimed to democratize music creation by making all potential melodies freely accessible and ineligible for copyright protection.

??????????? “All the Music” project employed a sophisticated algorithm to preemptively place every conceivable melody into the public domain, thereby rendering them ineligible for copyright protection. The core objective of the project was to generate an exhaustive database of all potential melodies within specific musical parameters and release them into the public domain. This initiative sought to address and potentially invalidate the current framework of melody copyright by ensuring that every possible combination of notes is freely accessible and cannot be subject to exclusive ownership.

??????????? The methodology involves algorithmic generation of melodies. Riehl and Rubin’s program systematically produces all possible combinations of notes within defined constraints, such as a particular scale and melody length. For example, the program might utilize the 12-note chromatic scale to generate trillions of unique melodies, each stored as a MIDI file—a digital format that encodes the sequence and timing of musical notes.

??????????? Crucially, all generated melodies are dedicated to the public domain. This legal status allows anyone to use these melodies without any restrictions, including for commercial purposes, and precludes future copyright claims. By making these melodies publicly available, the project aims to prevent allegations of copyright infringement based on these pre-existing combinations. The implications of Riehl’s project are profound and multifaceted. By challenging the very foundation of musical copyright, the “All the Music” project raises significant questions about the protection of musical works. If every possible melody is already in the public domain, the scope for claiming exclusive rights to new compositions becomes severely limited. This project could fundamentally alter the landscape of music copyright by ensuring that basic musical building blocks are available to all.

? ?????????? While the project has yet to face significant legal challenges, its existence introduces a radical argument against the traditional understanding of music copyright law. The principle is straightforward: if a melody is preemptively placed in the public domain, it cannot be subsequently copyrighted. This approach could render countless potential copyright infringement lawsuits moot, as the melodies in question would already be part of the public domain corpus.?

??????????? This effort is not limited to music.? In fact, this team has created a proof of concept for a similar project called “All the Patents” where AI generates inventions automatically or modified by user input.? Listen to an interview with Damian Riehl where this project could collapse patents at minute 34 of this interview:? https://changelog.com/practicalai/232#t=2045.?

???????????? Let’s apply all this to designs.? Back in 2017, The Nutella Unica project was an innovative marketing initiative by Ferrero, the makers of Nutella, aimed at creating seven million unique jar designs. The project utilized an algorithmic approach to generate distinct designs for each jar, ensuring that no two jars were alike. This was achieved through a combination of colors, patterns, and shapes that the algorithm could randomly mix and match, creating a virtually infinite array of possibilities.

???????????? The process began with Ferrero collaborating with the creative agency Ogilvy & Mather Italia. They developed an algorithm that could generate unique designs by manipulating a predefined set of graphic elements. These elements included different color palettes, abstract patterns, and various shapes, which were digitally combined in myriad ways to produce the final designs. The designs were then printed onto Nutella jars, making each one a collectible piece of art.

???????????? Resources for the project included advanced software development, digital printing technology, and a robust supply chain to handle the production and distribution of millions of uniquely designed jars. The project was a significant logistical feat, requiring precise coordination between designers, programmers, printers, and distribution channels. The implications for designers were profound, as the project showcased the potential of combining creativity with technology to produce mass-customized products. It highlighted a shift towards personalized consumer experiences and demonstrated how algorithms could be leveraged to enhance the creative process without diminishing the role of human designers.

??????????? Following the success of the Nutella Unica project, other brands and industries began exploring similar initiatives. The concept of algorithmically generated, personalized products gained traction, influencing sectors such as fashion, packaging, and consumer goods. This trend towards customization continues to grow, driven by advancements in digital technology and an increasing consumer demand for unique, individualized products.

??????????? The use of AI doesn’t stop with graphic designs.? DALL-E, Midjourney, Stable Diffusion, and Vizcom are all capable of creating millions, billions, and eventually trillions of product designs based on user prompts.? Under our current legal framework, this will be done both defensively and offensively.? Defensively, those wanting to protect their freedom to operate will proactively iterate billions of design variations on their current products.? This allows a company having a design theme to extend it to other products for very little cost.? If they wish to preclude others from protecting these designs, they can publish the AI-generated designs on an indexed, publicly accessible website similar to the “All the Music” project.? This poses a herculean task for subsequent innovators to clear their human-produced designs from the trillions of AI-generated variations.?

?

2. What types of AI-generated disclosures, if any, would be pertinent to patentability determinations made by the USPTO? How are such disclosures currently being made available to the public? In what other ways, if any, should such disclosures be made available to the public?

???????????? In the near term, the USPTO must determine if a design is nonobvious. This involves evaluating whether AI-generated disclosures can be considered prior art. If these AI-generated disclosures are widely read by designers of similar products, or result in products that are in public use, on sale, or otherwise available, then they should indeed be considered prior art. Such disclosures would provide a relevant baseline for the nonobviousness analysis, ensuring that only truly original designs receive patent protection.

??????????? However, any reference that relies on a previous design by the same inventor or assignee should not qualify as prior art. This is because it lacks originality, a fundamental criterion for prior art. Prior art must be independently conceived and not derived from the inventor’s earlier works. Thus, ensuring that prior art adheres to the same standards of originality required for design patents is essential for maintaining the integrity of the patent system.

??????????? By adhering to these principles, the USPTO can effectively balance the integration of AI in design with the protection of genuine innovation. This approach will help prevent the misuse of AI to flood the prior art landscape with non-original designs, thereby preserving the incentive for designers to create and protect truly novel works.

?

3. If a party submits to the Office a printed publication or other evidence that the party knows was AI-generated, should that party notify the USPTO of this fact, and if so, how? What duty, if any, should the party have to determine whether a disclosure was AI-generated?

???????????? The USPTO should provide a method where the inventor of a design can create a portion of the design and then task AI tools to complete the routine, non-inventive duplication or extension of the human created portion. This approach recognizes the value of human creativity while leveraging AI’s efficiency in executing repetitive tasks. Below are several examples illustrating how this method could work:

?Example 1: Speaker Grill Design

???????????? A designer of a speaker grill for an automobile might design the perimeter and the hexagonal openings that characterize the grill. The designer’s creative input ensures that the overall shape and the arrangement of the hexagons meet specific aesthetic and functional requirements. The AI software can then take over to evenly fill the space within the perimeter with the hexagonal pattern, maintaining consistency and precision across the entire grill.? As the hexagonal pattern intersects the perimeter, the AI software will ensure that no partial hexagon is too large or too small to be perceived as unbalanced or unproducible.?

?Example 2: Vinyl Fabric Graining

???????????? A designer creates a small section of graining on a vinyl fabric to simulate a leather texture. This initial design reflects the desired look and feel of the texture. The AI software can then randomly extend this pattern to cover an entire sheet of vinyl fabric. By doing so, the software replicates the grain’s randomness and variability, which are characteristic of natural leather, across a larger area without the need for manual repetition by the designer.

?Example 3: Complex Geometric Patterns

???????????? Consider a designer working on a complex geometric pattern for a tile. The designer creates a unique section of the pattern, incorporating specific shapes and alignments. The AI software can then replicate and extend this section to form a seamless, continuous pattern across multiple tiles. This ensures uniformity and reduces the time required for manual pattern replication.

??????????? To implement this method effectively, the USPTO should provide clear guidelines enabling inventors to identify individually created sections and the AI replication and extension of these sections. These guidelines could include:

1.???????? Documentation Requirements: Inventors should document the portions of the design they created manually, specifying the details and creative decisions involved.

2.???????? AI Contribution Explanation: Inventors should explain how AI tools were used to complete the design, including the parameters and algorithms employed.

3.???????? Distinction in Filing: Patent applications should clearly distinguish between human-created and AI-generated elements, ensuring transparency and proper attribution of creativity.

4.???????? Verification Procedures: The USPTO could establish procedures to verify the originality of the manually created sections and the routine nature of the AI-extended portions.

??????????? By adopting these methods and guidelines, the USPTO can foster a collaborative environment where human creativity and AI efficiency complement each other. This approach not only streamlines the design process but also ensures that inventors receive appropriate recognition and protection for their creative contributions.

????????????

4. Should an AI-generated disclosure be treated differently than a non-AI-generated disclosure for prior art purposes? For example:

a. Should the treatment of an AI-generated disclosure as prior art depend on the extent of human contribution to the AI-generated disclosure?

??????????? References that contain a combination of human and AI-generated content should be treated similarly to patent applications that include both human and AI-generated content.? See the example in Question 3 above.? If an invention would be deemed patent-eligible under 101, then art of the same nature should be considered prior art under sections 102 and 103.?

b. How should the fact that an AI-generated disclosure could include incorrect information (e.g., hallucinations) affect its consideration as a prior art disclosure?

???????????? Errors in AI-generated art should not be considered any differently than human-generated art having errors.

c. How does the fact that a disclosure is AI-generated impact other prior art considerations, such as operability, enablement, and public accessibility?

??????????? AI-generated prior art are not impacted by these considerations in design patents.?

?

5. At what point, if ever, could the volume of AI-generated prior art be sufficient to create an undue barrier to the patentability of inventions? At what point, if ever, could the volume of AI-generated prior art be sufficient to detract from the public accessibility of prior art (i.e., if a PHOSITA exercising reasonable diligence may not be able to locate relevant disclosures)?

??????????? Only art that would be patent-eligible should be considered in design patent examination or validity determinations.?

?B. The Impact of AI on a PHOSITA

?

6. Does the term “person” in the PHOSITA assessment presume or require that the “person” is a natural person, i.e., a human? How, if at all, does the availability of AI as a tool affect the level of skill of a PHOSITA as AI becomes more prevalent? For example, how does the availability of AI affect the analysis of the PHOSITA factors, such as the rapidity with which innovations are made and the sophistication of the technology?

?

Only art that would be patent-eligible should be considered in design patent examination or validity determinations.?

?

7. How, if at all, should the USPTO determine which AI tools are in common use and whether these tools are presumed to be known and used by a PHOSITA in a particular art?

???????????? AI tools that generate designs are not relevant to the examination process at the USPTO. The primary factor that should be considered during the examination of a design patent is the amount and nature of human creativity involved in the creation of the design. Human ingenuity is the cornerstone of innovation, and it is essential that the patent system recognizes and rewards the creative contributions made by inventors, rather than the computational capabilities of AI systems.

The most relevant aspect in evaluating a design’s patentability is the originality and inventiveness demonstrated by the human inventor. AI tools can assist in the design process by performing routine, non-inventive tasks, but the core creative elements must originate from the inventor. Therefore, the USPTO should focus on the unique contributions made by human inventors, ensuring that the patent examination process remains true to its purpose of promoting genuine innovation.

To facilitate this focus on human creativity, the USPTO should provide inventors with a means of documenting their contributions to the design. This documentation could include detailed records of the initial creative steps taken by the inventor, descriptions of the design elements created manually, and explanations of how AI tools were used to assist with non-inventive tasks. By implementing such measures, the USPTO can ensure that the true value of human creativity is recognized and protected, while also acknowledging the supportive role of AI in the design process.

?8. How, if at all, does the availability to a PHOSITA of AI as a tool impact:

a. Whether something is well-known or common knowledge in the art?

?

See the response to question 7 above.?

?

b. How a PHOSITA would understand the meaning of claim terms?

?

See the response to question 7 above.?

?

9. In view of the availability to a PHOSITA of AI as a tool, how, if at all, is an obviousness determination affected, including when:

a. Determining whether art is analogous to the claimed invention, given AI’s ability to search across art fields? Does the “analogous” art standard still make sense in view of AI’s capabilities?

?AI tools that generate designs should not be relevant to the examination process at the USPTO. The primary focus of the examination process should be the amount and nature of human creativity involved in creating the design. Human ingenuity is the cornerstone of innovation, and it is crucial that the patent system recognizes and rewards the creative contributions made by inventors rather than the computational abilities of AI systems.

???????????? Designs created exclusively by AI should be considered art.? If the design is not patent-eligible because it lacks sufficient human contribution, it should similarly not be considered art.? The tools that create the designs are not relevant to this analysis.? Artists will use any available technology to create art.?

?

b. Determining whether there is a rationale to modify the prior art, including the example rationales suggested by KSR (MPEP 2143, subsection I) (e.g., “obvious to try”) or the scientific principle or legal precedent rationales (MPEP 2144)?

?

??????????? As previously described, now is the right time for Congress to eliminate nonobvious requirement design patents.? Recognizing that USPTO must examine the tens of thousands of designs filed each year, it must not use hindsight to render a design obvious.? The use of hindsight in evaluating the obviousness of patent claims is a critical issue addressed by the U.S. Supreme Court. The Court has consistently emphasized that an obviousness determination should be based on objective evidence without relying on hindsight.

??????????? In Graham v. John Deere Co. of Kansas City (1966), the Supreme Court articulated the framework for determining obviousness under 35 U.S.C. § 103. The Court highlighted that the obviousness analysis must avoid the use of hindsight:

??????????? “Under § 103, the scope and content of the prior art are to be determined; differences between the prior art and the claims at issue are to be ascertained; and the level of ordinary skill in the pertinent art resolved. Against this background, the obviousness or nonobviousness of the subject matter is determined. Such secondary considerations as commercial success, long felt but unsolved needs, failure of others, etc., might be utilized to give light to the circumstances surrounding the origin of the subject matter sought to be patented. As indicia of obviousness or nonobviousness, these inquiries may have relevancy.”

??????????? In KSR International Co. v. Teleflex Inc., 550 U.S. 398 (2007), the Court addressed the dangers of hindsight bias in the obviousness analysis:

“A factfinder should be aware, of course, of the distortion caused by hindsight bias and must be cautious of arguments reliant upon ex post reasoning.”

In the design patent case of Whitman Saddle Co. v. United States, 148 U.S. 674 (1893), it reinforces the principle that hindsight should not be used in the evaluation of patents:

“It is not sufficient to say that the things combined were old, known, or used; and that the combination, when made, accomplished a new and beneficial result. This is not enough, unless it can be shown that the things combined did not exist together before in such a manner that their combination would naturally suggest itself to persons skilled in the art, and that they were not used together in the manner claimed, so as to produce a new and useful result.”

These cases collectively underscore that obviousness determinations must be made based on objective criteria, avoiding the use of hindsight. This principle extends to the use of AI in generating proposed combinations of references. AI systems, with their vast data-processing capabilities, can analyze and combine prior art in ways that might not be apparent to a person having ordinary skill in the art (PHOSITA).

??????????? However, relying on AI-generated combinations poses a risk of hindsight bias, as these systems can retrospectively create combinations that seem obvious after the fact but would not have been evident to a PHOSITA at the time of the invention. Therefore, the Supreme Court’s rulings in Graham, KSR, and Whitman Saddle collectively preclude the use of AI to generate a proposed combination of references if it introduces hindsight bias into the obviousness analysis.?

??????????? While AI tools can assist in identifying prior art, their use must be carefully controlled to ensure that the inventive step is evaluated based on human creativity and not retrospective reconstruction, adhering to the guidance provided by these landmark Supreme Court cases.

c. Determining whether the modification yields predictable results with a reasonable expectation of success (e.g., how to evaluate the predictability of results in view of the stochasticity (or lack of predictability) of an AI system)??

??????????? AI systems have a nearly limitless capacity to combine elements of prior designs. This extensive ability is evident in examples such as the Nutella label and the “All the Music” project. The Nutella label generator, for instance, demonstrates how AI can effortlessly create numerous variations of a label design by combining different graphic elements. Similarly, the “All the Music” project showcases AI’s capability to generate an exhaustive range of musical compositions by combining existing musical elements in countless ways?

??????????? In the context of patent examination, particularly for design patents, the element of enablement is rarely a significant concern. Enablement pertains to the requirement that a patent application must describe the invention in sufficient detail for a person skilled in the art to replicate it. However, for design patents, which focus on the ornamental aspects of an item, the concept of enablement is generally straightforward and less complex than for utility patents.

?

??????????? Moreover, the question of predictability is not applicable to design patents. Design patents are concerned with the visual appearance of an item, not its functional aspects. Therefore, the predictability of combining prior art references, a consideration crucial in utility patent examinations, does not hold the same relevance in the context of design patents. The emphasis for design patents remains on the novelty and non-obviousness of the design’s ornamental features, evaluated without the use of hindsight.

??????????? While AI’s vast capacity to generate combinations of prior art presents challenges, the principles of novelty and non-obviousness should guide the examination of design patents, without undue reliance on AI-generated combinations that benefit from hindsight. The focus should remain on the creative contributions of human inventors and the aesthetic innovations they bring to their designs.

d. Evaluating objective indicia of obviousness or nonobviousness (e.g., commercial success, long felt but unsolved needs, failure of others, simultaneous invention, unexpected results, copying, etc.)?

??????????? AI tools have been developed to predict the attractiveness and commercial success of product designs. These tools leverage vast datasets and advanced algorithms to assess consumer preferences and forecast market performance.?

??????????? Adobe Sensei integrates AI into its Creative Cloud suite through Adobe Sensei, which analyzes design elements to predict consumer response. This tool uses machine learning to understand trends and preferences, helping designers create more appealing products?

??????????? Symanto offers AI-driven insights into consumer behavior, using natural language processing to evaluate sentiment and predict how a design will be received by the market. This tool is useful for tailoring designs to specific audience segments based on psychographic data.?

??????????? Alex Burnap, a Yale SOM assistant professor, explores using AI to enhance vehicle design processes in the Yale Insights article: “Can AI Help Design a More Appealing Car?” ?Burnap, with colleagues from MIT and Northwestern, developed a machine-learning model to streamline design experimentation and market focus. This model predicts consumer responses to designs and generates new design concepts, reducing reliance on costly and labor-intensive focus groups.?

??????????? The AI algorithm uses a deep neural network trained on a mix of rated and unrated vehicle images to predict consumer ratings and generate new designs. Despite challenges with small data sets and the complexity of image data, the model significantly improved prediction accuracy and facilitated creative design exploration. The AI tool aims to augment, not replace, human designers by providing early feedback and creative inspiration, improving the design process’s efficiency and innovation. The model successfully anticipated design trends, indicating its potential to guide future vehicle designs. Burnap emphasizes that AI should enhance existing workflows, empowering designers to focus on creativity and iteration.

??????????? These tools illustrate the growing role of AI in the design industry, providing valuable insights that help creators develop products with higher chances of market acceptance and success.? Does this mean that a design with a high consumer appeal score is obvious?? USPTO and the patent system should not discourage utilizing technology to improve commercial success and reduce development costs.? I believe these tools prove the point that everything and therefore nothing is obvious when it comes to designs and Congress should eliminate this requirement, but returning to the focus of the question, consumer appeal as predicted by AI tools should have no part in constructing an obvious rejection.? Guidelines should focus on what the PHOSITA would consider and not the output of AI tools.?

?

10. How, if at all, does the recency of the information used to train an AI model or that ingested by an AI model impact the PHOSITA assessment when that assessment may focus on an earlier point in time (e.g., the effective filing date of the claimed invention for an application examined under the First-Inventor- to-File provisions of the America Invents Act)??

??????????? The recency of the information used to train or ingested by an AI model can significantly impact the Person Having Ordinary Skill In The Art (PHOSITA) assessment, especially when that assessment is focused on an earlier point in time, such as the effective filing date of a claimed invention under the First-Inventor-to-File provisions of the America Invents Act.? When evaluating the PHOSITA standard, the goal is to determine what a person of ordinary skill in the relevant field would have known or been able to do as of the effective filing date. This involves considering the state of the art, including available knowledge and technologies, at that specific time.

??????????? If the AI model is trained on or ingests more recent information, it may incorporate advancements, techniques, or understandings that were not available or known at the earlier point in time. This can lead to an inaccurate PHOSITA assessment in two primary ways:?

??????????? Overestimating Skills and Knowledge: The AI might assume a higher level of skill or knowledge than what was actually common in the field at the earlier date. This can result in an unrealistic portrayal of what the PHOSITA would have been capable of, potentially affecting the evaluation of the invention’s novelty and non-obviousness.

??????????? Incorporating Non-Existent Technologies: The AI might reference technologies or methodologies that were developed after the effective filing date. This could lead to erroneous conclusions about the obviousness of the invention, as these later-developed technologies would not have been part of the prior art at the relevant time.?

??????????? To accurately reflect the PHOSITA assessment at the earlier date, the AI model must be trained on information that was available up to that specific point in time. This ensures that the model’s outputs are consistent with the historical context of the invention, providing a fair and accurate assessment of the state of the art as it existed when the invention was made.

?

11. How, if at all, does the availability to a PHOSITA of AI as a tool impact the enablement determination under 35 U.S.C. 112(a)? Specifically, how does it impact the consideration of the In re Wands factors (MPEP 2164.01(a)) in ascertaining whether the experimentation required to enable the full scope of the claimed invention is reasonable or undue??

??????????? As discussed in question 12 below in more detail, enablement is rarely an issue in design patents.?

?

C. The Implications of AI That Could Require Updated Examination Guidance and/or Legislative Change

?

12. What guidance from the USPTO on the impact of AI on prior art and on the knowledge of a PHOSITA, in connection with patentability determinations made by the Office, would be helpful??

??????????? Enablement under 35 U.S.C. 112(a) for design patents requires that the patent application provides a detailed description of the claimed design sufficient to enable a PHOSITA to make and use the design without undue experimentation. In the context of design patents, the enablement requirement primarily focuses on the sufficiency and clarity of the drawings and, to a lesser extent, the written description.?

??????????? The primary means of satisfying the enablement requirement in a design patent is through the drawings. These drawings must clearly illustrate the ornamental design of the article and include sufficient views to convey the complete appearance of the design.? The drawings must show every feature of the design, and these features must be clearly and accurately depicted. This includes providing multiple views (e.g., front, rear, top, bottom, left, and right sides) if necessary to fully understand the design.? While the emphasis is on the drawings, a written description may be provided to explain the design further. However, the written description cannot substitute for inadequate drawings. It can, however, provide context or details that might not be fully clear from the drawings alone.?

??????????? Design patents have occasionally been found non-enabled due to deficiencies in the drawings or descriptions. Here are a few examples:

·????? In re Rubinfield, 270 F.2d 391 (C.C.P.A. 1959), the Court of Customs and Patent Appeals (CCPA) found that the design patent was not enabled because the drawings were inadequate. The court emphasized that the drawings did not provide a complete understanding of the design as they were inconsistent and did not depict the claimed design sufficiently.

·????? In re Stevens, 173 F.2d 1015 (C.C.P.A. 1949), the CCPA held that a design patent application was non-enabled due to the lack of sufficient views. The drawings did not provide enough detail to fully disclose the design, and the court stressed that all necessary views must be included to meet the enablement requirement.

·????? In International Seaway Trading Corp. v. Walgreens Corp., 599 F. Supp. 2d 1307 (S.D. Fla. 2009), the district court found a design patent invalid due to insufficient disclosure. The drawings failed to adequately disclose the claimed design in its entirety, making it impossible for a PHOSITA to understand and recreate the design without undue experimentation.?

??????????? By adhering to these guidelines, a design patent application can meet the enablement requirements under 35 U.S.C. 112(a), ensuring that the claimed design is fully disclosed and can be reproduced by a PHOSITA without undue experimentation.? It is possible to use AI tools to identify issues regarding enablement, especially for design patent drawings.? Given the dearth of issues in this field, it is not a priority for USPTO.?

?

13. In addition to the considerations discussed above, in what other ways, if any, does the proliferation of AI impact patentability determinations made by the Office (e.g., under 35 U.S.C. 101, 102, 103, 112, etc.)?

?

??????????? The USPTO should provide a method where the inventor of a design can create a portion of the design and then task AI tools to complete the routine, non-inventive duplication. This approach leverages human creativity and AI efficiency in a complementary manner. Below are examples illustrating this concept, followed by an explanation of how the USPTO could implement guidelines to distinguish human-created sections from AI-generated ones.

?

Example 1: Speaker Grill Design

A designer of a speaker grill for an automobile might design the perimeter and the hexagonal openings. This initial design represents the creative input required to ensure the aesthetic and functional aspects of the grill. The AI software then takes over to evenly fill the space within the perimeter with the hexagonal pattern, ensuring consistency and precision across the entire grill.

?

Example 2: Vinyl Fabric Graining???????

A designer creates a small section of graining on a vinyl fabric to simulate a leather texture. The initial design captures the desired look and feel of the texture. The AI software can then extend this pattern to cover an entire sheet of vinyl fabric, replicating the grain’s randomness and variability characteristic of natural leather.

?

Example 3: Complex Geometric Patterns

??????????? Consider a designer working on a complex geometric pattern for a tile. The designer creates a unique section of the pattern, incorporating specific shapes and alignments. The AI software can then replicate and extend this section to form a seamless, continuous pattern across multiple tiles.

?

Guidelines for Identifying Creative and AI-Generated Sections

To implement this method effectively, the USPTO should provide clear guidelines enabling inventors to identify individually created sections and AI replication and extension of these sections. These guidelines could include:

1.???????? Documentation Requirements: Inventors should document the portions of the design they created manually, specifying the details and creative decisions involved.

2.???????? AI Contribution Explanation: Inventors should explain how AI tools were used to complete the design, including the parameters and algorithms employed.

3.???????? Distinction in Filing: Patent applications should clearly distinguish between human-created and AI-generated elements, ensuring transparency and proper attribution of creativity.

4.???????? Verification Procedures: The USPTO could establish procedures to verify the originality of the manually created sections and the routine nature of the AI-extended portions.

?

By adopting these methods and guidelines, the USPTO can foster a collaborative environment where human creativity and AI efficiency complement each other. This approach not only streamlines the design process but also ensures that inventors receive appropriate recognition and protection for their creative contributions.

?

14. Are there any laws or practices in other countries that effectively address any of the questions above? If so, please identify them and explain how they can be adapted to fit within the framework of U.S. patent law.

?

In several jurisdictions, the nonobviousness requirement is not a criterion for design protection. Here are some examples:

·????? United Kingdom and European Union

In both the United Kingdom and the European Union, designs do not need to meet a nonobviousness criterion to be protected. The primary requirements for design protection in these jurisdictions are that the design must be new and have individual character. Individual character means that the overall impression the design produces on the informed user differs from that produced by any design made available to the public before the filing date.

·????? Japan

Japan also does not require nonobviousness for design registration. The Japanese design law focuses on the novelty and the aesthetic aspects of the design. Similar to the EU, the design must be new and create a different impression on an informed user compared to prior designs.

·????? China

In China, the requirement for design protection is focused on absolute novelty. Designs must be new and not previously published or publicly used anywhere in the world. There is no explicit requirement for nonobviousness in the Chinese design law.

·????? Brazil

Brazilian law grants design protection based on novelty and originality. Nonobviousness is not a requirement. The design must be new and must not have been disclosed before the date of filing.

These jurisdictions provide a more accessible pathway to design protection by not imposing a nonobviousness requirement, focusing instead on the novelty and distinct visual appearance of the designs. This approach facilitates the protection of a broader range of designs, encouraging innovation and creativity in design industries.

?

15. Should title 35 of the U.S. Code be amended to account for any of the considerations set forth in this notice, and if so, what specific amendments do you propose, and why?

?

Title 35 USC 171a should be amended as follows:

“Original” means that the design must originate from the inventor or the inventor’s assignees and not be based on a design not created by the inventor or the inventor’s assignees.??

Amending 35 U.S.C. 171(a) to explicitly define “original” as meaning that the design must originate from the inventor or the inventor’s assignees, and not be based on a design not created by the inventor or the inventor’s assignees, would provide greater clarity and integrity to the design patent system. This definition would ensure that design patents are awarded only for true innovations that reflect the creative effort of the applicant, thereby preserving the originality and authenticity of design patents. By emphasizing the origin of the design, this amendment would prevent the misappropriation of designs and discourage the practice of merely modifying or copying existing designs for the purpose of obtaining patent protection.

This proposed definition aligns with the fundamental principles of intellectual property law, which aim to protect and reward genuine creativity and innovation. By requiring that the design must originate from the inventor or their assignees, the amendment would reinforce the idea that patent protection is reserved for those who contribute original ideas and creative efforts. This would help maintain a high standard of innovation within the design community and ensure that patents are not granted for designs that merely replicate or slightly alter existing works.

Furthermore, defining “original” in this manner would provide clearer guidelines for both patent applicants and examiners. Applicants would have a better understanding of the requirements for obtaining a design patent, knowing that their submissions must truly originate from their own creative efforts or those of their assignees. Examiners, on the other hand, would have a more concrete standard to assess the originality of a design, making the examination process more straightforward and consistent. This clarity would reduce ambiguities and disputes over what constitutes an “original” design, leading to more predictable and reliable outcomes in the patenting process.

Lastly, this amendment would foster a more ethical and transparent design patent landscape. By explicitly stating that a design must originate from the inventor or their assignees, it would discourage the unethical practice of copying or minimally altering another’s design and seeking patent protection for it. This would not only protect the rights and efforts of original designers but also promote a culture of integrity and respect within the design industry. Ensuring that design patents are granted only for truly original works would enhance the credibility of the design patent system and encourage genuine innovation, ultimately benefiting consumers, creators, and the broader marketplace.

?

Title 35 USC 171b should be amended as follows:

“Title 35 USC § 103 shall not apply to patents for designs.”

Design patents, which protect the ornamental design of a functional item, are currently subject to the same patentability requirements as utility patents under U.S. law. One of these requirements is non-obviousness, outlined in 35 U.S.C. 103. This means that a design patent must not be obvious to a PHOSITA in view of the prior art. However, the application of this non-obviousness standard to design patents is ill-suited to the unique nature of design patents, which focus on aesthetics rather than functional innovations. Therefore, 35 U.S.C. 171(b) should be amended to exclude 35 U.S.C. 103 for design patents, aligning the evaluation criteria more closely with the specific characteristics of design patents.

Firstly, design patents are inherently different from utility patents because they protect the visual appearance of an object rather than its functional features. The subjective nature of aesthetic judgment makes the application of the non-obviousness standard, which is inherently objective, problematic. Determining what constitutes an “obvious” design often involves a significant degree of subjectivity, leading to inconsistent and unpredictable outcomes. Excluding design patents from 35 U.S.C. 103 would eliminate this incongruity, allowing for a more straightforward assessment based on the novelty and distinctiveness of the design, which are more appropriate measures for design patents.

Secondly, the current non-obviousness requirement for design patents can stifle innovation and creativity in design fields. Designers often build upon existing designs, creating new variations that might not be deemed “non-obvious” under the strict standards of 35 U.S.C. 103. This can discourage designers from pursuing new design patents, knowing that their designs might not meet the stringent non-obviousness criteria. Amending 35 U.S.C. 171(b) to exclude the non-obviousness requirement would encourage more design innovation by allowing a broader range of designs to be patented, thus fostering a more vibrant and diverse design landscape.

Moreover, the exclusion of 35 U.S.C. 103 for design patents would streamline the patent examination process. Patent examiners currently face challenges in assessing the non-obviousness of design patents due to the subjective nature of aesthetic judgments. This can lead to longer examination times and higher costs for both the Patent Office and applicants. By removing the non-obviousness requirement, the examination process for design patents could be simplified and expedited, reducing administrative burdens and making it more efficient to obtain design patent protection.

Amending 35 U.S.C. 171(b) to exclude the application of 35 U.S.C. 103 for design patents would address the fundamental differences between design and utility patents. It would eliminate the subjective and often inconsistent application of the non-obviousness standard to design patents, encourage greater innovation and creativity in design fields, and streamline the patent examination process. This amendment would better align the patentability criteria with the unique nature of design patents, ultimately benefiting designers, the Patent Office, and the broader innovation ecosystem.

?

Title 35 USC 171b should be amended as follows:

“Title 35 USC § 102(a)(1) “printed publication” shall not apply to designs that are not widely read by humans or that would not be patentable under 35 USC 101.

Amending 35 U.S.C. 171(b) to specify that printed publications created by computers that are not widely read by humans do not qualify as prior art would address significant concerns about the evolving nature of information dissemination in the digital age. As artificial intelligence and automated content generation become more prevalent, there is a growing volume of computer-generated publications that may not necessarily be intended for or accessed by a human audience. These publications, while technically accessible, do not contribute meaningfully to the public domain of knowledge and should not be considered in evaluating the novelty or obviousness of design patents.

The essence of prior art lies in its availability and accessibility to the public. Prior art serves the purpose of preventing the patenting of known information and ensuring that patents are granted only for true innovations. Publications that are generated by computers but not widely read by humans fail to meet the core criterion of public accessibility. They do not enhance the knowledge base of the relevant community or contribute to the state of the art in a meaningful way. By excluding such publications from qualifying as prior art, the amendment would ensure that the evaluation of patent applications remains rooted in the practical and accessible body of knowledge that truly informs and guides the relevant field.

This amendment would mitigate the risk of an overwhelming and unmanageable volume of prior art references. The increasing capability of computers to generate vast amounts of content could inundate the patent examination process with irrelevant or low-quality references that are not actually utilized by professionals or the public. This could complicate and delay the patent examination process, making it more difficult for examiners to identify genuinely relevant prior art. By excluding computer-generated publications that are not widely read, the amendment would streamline the process, focusing on meaningful and impactful references that truly reflect the state of the art.?

The amendment would protect the integrity of the patent system by ensuring that prior art remains a reliable indicator of public knowledge and technological progress. Computer-generated content that is not read by humans does not contribute to the dissemination of knowledge or the advancement of technology. Including such content as prior art could undermine the value of the patent examination process, leading to unfair rejections of patent applications based on references that do not genuinely represent the knowledge or practices of the field. By clearly defining what qualifies as prior art, the amendment would uphold the credibility and fairness of the patent system.

?Amending 35 U.S.C. 171(b) to state that printed publications created by computers that are not widely read by humans do not qualify as prior art would align the criteria for prior art with the principles of public accessibility and meaningful contribution to the field. This change would streamline the patent examination process, maintain the integrity of the patent system, and ensure that patents are granted based on genuine advancements in knowledge and technology. It would address the challenges posed by the proliferation of computer-generated content, ensuring that the evaluation of design patents remains grounded in the practical and accessible body of knowledge that truly informs and advances the relevant art.

???????????? I thank you for considering my comments and I welcome any further dialogue or opportunity to provide additional information to assist your efforts.? This comment was researched and prepared using the latest AI research tools.?

?

Sincerely,

?

Damian Porcari