Artificial Intelligence and Intellectual Property: A Potentially Explosive Mixture
Prepared for Proceedings of the American Institute of Chemical Engineers 2025 Spring Meeting and 21st Global Congress on Process Safety Dallas, TX April 6-10, 2025
I. Introduction
“As a human being, one has been endowed with just enough intelligence to be able to see clearly how utterly inadequate the intelligence is when confronted with what exists.”1
What exists today is an emerging technology that appears to have the capacity to revolutionize any human activity not completely executed by hand. Artificial Intelligence (“AI”) is attracting mind-numbing magnitudes of investment in the promise of relativistic returns. It is being developed in a variety of ways to address the common groupings of myriad use cases.
Intellectual Property (“IP”) is a body of law that developed first in common law, and then by statute, over hundreds of years to protect rights of inventors, authors, and actors in commerce with rights to that have been roughly analogized real property rights. Intellectual Property law has developed into four distinct bodies of law, which may not have all kept up with the rapid rate of development and adoption of AI.
Patent law, which protects novel and nonobvious inventions directed to “any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof…”2
1Albert Einstein, letter to Queen Elizabeth of Belgium, September 19, 1932.
235 U.S.C. §101.
Copyright law, which protects “original works of authorship fixed in any tangible medium of expression… Works of authorship include the following categories:
literary works;
musical works, including any accompanying words;
dramatic works, including any accompanying music;
pantomimes and choreographic works;
pictorial, graphic, and sculptural works;
motion pictures and other audiovisual works;
sound recordings; and
architectural works.”3
Trade secret law, which generally protects information used in business which is maintained as a secret, and which is valuable based on its not being generally known. Trade secrets can include recipes, customer lists, repair manuals, and process descriptions.
Trademark law, which protects generally any word, name, symbol, or device, or any combination thereof used or intended for use by a person to identify and distinguish his or her good from those manufactured or sold by others and to indicate the source of the goods.4 Since AI has not as of yet generated any particularly thorny relevant issues, trademark law will not be further addressed in this paper.
This paper will address certain interesting questions that arise when AI meets the laws of patents, copyrights, and trade secrets, and some of the surprising consequences of such commingling. Would Einstein find our intelligence adequate to confront the coexistence of Artificial Intelligence and Intellectual Property?
317 U.S.C. § 102.
4See 15 U.S.C. § 1127.
II. AI and Patents
Are AI Models Patentable?
The five primary statutory requirements for a U.S. patent are subject matter eligibility5, novelty6 and nonobviousness7 over the prior art, and written description and enablement in the specification8.
Subject matter eligibility presents a threshold question: Is this the type of invention that the patent statute was intended to protect? The statute recites as eligible subject matter “any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof…” Courts have provided a two-part test to determine whether the claimed invention is within the scope of patent-eligible subject matter. First, does the claim fall into one of the judicial exceptions of laws of nature, natural phenomena, and abstract ideas which have been held to be generally not patentable?9
Second, if the claim is directed to a patent-ineligible concept, does it contain an inventive concept sufficient to transform the claimed abstract idea into a patent-eligible application?10 The court will consider the claim elements individually and as an ordered combination to determine whether any additional limitations amount to significantly more than the ineligible concept.11
Courts have long held that mathematical algorithms for performing calculations, without more, are abstract ideas, and thus patent ineligible. “Without additional limitations, a process that employs mathematical algorithms to manipulate existing information to generate additional information is not patent eligible.”12 Simply storing information and providing it upon request does not alone transform the abstract idea into patent eligible subject matter.13 Additional claim elements that improve the functioning of the computer itself or improvements to any other technology or technical field have been cited as examples of “inventive concepts” for purposes of satisfying the second step.14
5“Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.” 35 U.S.C. § 101. 6“A person shall be entitled to a patent unless— (1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the eAective filing date of the claimed invention…” 35 U.S.C. § 102(a) (text abbreviated). 7“A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the diAerences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the eAective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.” 35 U.S.C. § 103. 8“The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.” 35 U.S.C. § 112. 9Ass'n for Molecular Pathology v. Myriad Genetics, Inc., 569 U.S. 576, 589 (2013), citing Mayo Collaborative Servs. v. Prometheus Labs., Inc., 566 U.S. 66, 70 (2013). 10 In re Killian, 45 F.4th 1373, 1379 (Fed. Cir. 2022), citing Mayo at 72, 80. 11 See Alice Corp. Pty. Ltd. v. CLS Bank Int'l, 573 U.S. 208, 217-18, 221 (2014). 12 In re Schrader, 22 F.3d 290, 294 (Fed. Cir. 1994) (holding that a data gathering step of entering bids was "insuAicient to impart patentability to a claim involving the solving of a mathematical algorithm".
A claimed process is patent-eligible under § 101 if: (1) it is tied to a particular machine or apparatus, or (2) it transforms a particular article into a different state or thing.15 This “machine-or-transformation test” can provide a useful clue in the second step of the Alice framework, but it is not the sole test governing § 101 analyses.16 Where the only mention of a particular machine was in the claim's preamble (rather than in the claim elements), a court found the process claim ineligible, since “we have routinely held that a preamble does not limit claim scope if it merely states the purpose or intended use of an invention.17
Patent claims to AI models will face this same challenging eligibility landscape. To maintain scope (and potential commercial value) of patent claims, patent draftsmen routinely seek to employ the broadest language possible, and to thus avoid limitations thought to be extraneous to the core of the invention. However, in the field of AI, this may cause an examiner at the U.S. Patent & Trademark Office (“USPTO”) who is applying the above rules to find that the invention is drawn to an abstract idea (Alice step 1), lacks an additional element sufficient to provide “inventive concept” (Alice step 2), and is thus patent-ineligible. Even if the examiner does not come to this conclusion allowing the patent to grant, when the patentee later seeks to enforce the AI patent against an accused infringer, a much more heavily-resourced legal examination in litigation may come to that same conclusion, after great expenditures of cash and tears.
Despite these challenges, a recent study indicates that there is a burgeoning interest in filing patent applications directed to AI. The figure below shows estimates of the aggregate number of patent applications and granted patents that published from the USPTO since the year 2000.18 Applications generally publish 18 months after the earliest priority filing. The estimates for 2023 are between 155,000 and 175,000 published patent applications and granted patents.
13 See, e.g., In re Greenstein, 774 F. App'x 661, 664 (Fed. Cir. 2019) explaining that the claims at issue only invoked a computer as a generic tool to store information and record transactions. 14 See Alice at 210. 15 See In re Bilski, 545 F.3d 943, 954 (Fed. Cir. 2008). 16 See Ultramercial, Inc. v. Hulu, LLC, 772 F.3d 709, 716 (Fed. Cir. 2014). 17 Digitech Image Techs., LLC v. Electronics for Imaging, Inc., 758 F.3d 1344, 1351 (Fed. Cir. 2014). 18 Pairolero, N.A., Giczy, A.V., Torres, G. et al. “The artificial intelligence patent dataset (AIPD) 2023 update.” J Technol Transf (2025). https://doi.org/10.1007/s10961-025-10189-8
What happens when AI “invents”?
AI models can excel at sorting through large databases to find otherwise obscure relationships among individual record elements. For example, pharmaceutical innovators are using AI models to sort through vast proprietary libraries of compounds and their physico-chemical properties to identify potential candidates for laboratory evaluation. What happens when the AI model selects the compound that winds up in the patent claim on a successful drug? Can AI be listed as an inventor?
A fairly recent case holds that AI cannot be an inventor. “The Patent Act, 35 U.S.C.S. § 100(f), requires that inventors must be natural persons; that is, human beings. Congress has determined that only a natural person can be an inventor, so artificial intelligence (AI) cannot be. Nothing in the Patent Act indicates Congress intended to deviate from the default meaning. The Patent Act uses personal pronouns to refer to an "individual" under 35 U.S.C.S. § 115(b)(2). While the court did not decide whether an AI system can form beliefs, nothing in the record showed that one could.”19
What happens if the patent applicant uses AI to conceive and/or reduce to practice, but represents herself as the inventor?
The USPTO imposes a “duty of candor” regarding the filing and prosecution of a patent application.
“Each individual associated with the filing and prosecution of a patent application has a duty of candor and good faith in dealing with the Office, which includes a duty to disclose to the Office all information known to that individual to be material to patentability as defined in this section. The duty to disclose information exists with respect to each pending claim until the claim is cancelled or withdrawn from consideration, or the application becomes abandoned. * *.
19 Thaler v. Vidal, 43 F.4th 1207, 1209 (Fed. Cir. 2022).
Individuals associated with the filing or prosecution of a patent application within the meaning of this section are:
Each inventor named in the application;
Each attorney or agent who prepares or prosecutes the application; and
Every other person who is substantively involved in the preparation or prosecution of the application and who is associated with the inventor, the applicant, an assignee, or anyone to whom there is an obligation to assign the application.”20
After a patent has granted, and the patentee then seeks to enforce the patent against an accused infringer, that defendant can raise a defense of inequitable conduct by plausibly alleging that during filing or prosecution of the patent application, the applicant failed to observe the duty of candor.
The remedy for inequitable conduct is unenforceability of the patent. If a trial court considers that the case is an exceptional one, which is often found when the patent has been improperly procured, attorney fees can be awarded.21 The court may find that the inequitable conduct infected related patent applications and patents, and thus hold continuations, continuations-in-part, divisionals and reissue patents unenforceable. Moreover, the Supreme Court has held that "enforcement of a patent procured by fraud on the Patent Office may be violative of § 2 of the Sherman Act provided the other elements necessary to a § 2 case are present."22
After the Thaler decision (which held that AI cannot be an inventor), and in view of the law of inventorship (regarding the entity that first conceived the invention and reduced it to practice), it would appear that failure to disclose to the USPTO during prosecution of a pending patent application that an AI model first generated the inventive concept (as recited in the claims) could risk a later finding of inequitable conduct. This possibility may also come to light in the due diligence process conducted by a prospective third party licensee/assignee.
III. Copyright
Can AI Be an Author for Purposes of Copyright?
A recent court decision directly addresses this question. Dr. Stephen Thaler is a computer scientist who creates and works with artificial intelligence systems, and who invented the Creativity Machine. (Dr. Thaler is apparently an ardent and persistent AI champion, as he is indeed the plaintiff in the patent case referenced supra.) On May 19, 2019, Dr. Thaler submitted a copyright registration application to the Copyright Office for an artwork titled "A Recent Entrance to Paradise." On the application, Dr.Thaler listed as the author of that work the "Creativity Machine", an AI model of his making. Under "Copyright Claimant," Dr. Thaler provided his own name. In the section labeled "Author Created," Dr. Thaler wrote "2-D artwork, created autonomously by machine."
20 37 CFR 1.56 (a),(c). 21 Zenith Elecs. Corp. v. Exzec, Inc., 182 F.3d 1340, 1349 (Fed. Cir. 1999). 22 Walker Process Equip., Inc. v. Food Mach. & Chem. Corp., 382 U.S. 172, 174 (1965).
"A Recent Entrance to Paradise" generated by the "Creativity Machine", and claimed by its developer, Dr. Stephen Thaler.
The Copyright Office denied Dr. Thaler's application because "a human being did not create the work." Dr. Thaler sought reconsideration of that denial through the Office’s Registration Program, which again denied registration because the work lacked "sufficient creative input or intervention from a human author." Thaler then filed for a reconsideration by the Review Board, arguing that he should own the copyright under the work-made-for-hire doctrine because "non-human, artificial persons such as companies can already be authors under this doctrine." The Review Board affirmed the denial of Dr. Thaler's copyright application based on the human-authorship requirement.
Not to be deterred in his quest for recognition of AI authorship, Dr. Thaler then filed a complaint in the U.S. District Court for the District of Columbia. On summary judgment, that court affirmed the Copyright Office's denial of registration. Based on the caselaw and the Copyright Act's text, the district court concluded that "[h]uman authorship is a bedrock requirement of copyright."23 The court also held that Dr. Thaler could not rely on the work-made-for-hire provision because that provision "presuppose[s] that an interest exists to be claimed."24 Dr. Thaler then filed an appeal.
The appellate court examined the language of the Copyright Act of 1976. Based on provisions in the statute regarding the author’s capacity to hold property25, the author’s lifespan and date of death26, succession in ownership by the author’s widow or widower27, the signature of the author28, the author’s nationality or domicile29, and the author’s intentions30, the court held that the statute must be construed as requiring that an author be a human being.31
23 Thaler v. Perlmutter, 687 F. Supp. 3d 140, 146 (D.D.C. 2023).
24 Id. at 150.
25 17 U.S.C. § 201(a).
26 17 U.S.C. § 302(a).
27 17 U.S.C. § 203(a)(2).
28 17 U.S.C. § 204(a).
29 17 U.S.C. § 104(a).
Regarding Dr. Thaler’s argument that the registration was proper as a work for hire, the court noted that the Copyright Act only protects "original works of authorship." The authorship requirement applies to all copyrightable works, including works-made for-hire. The word “authorship,” like the word “author,” refers to a human being. As a result, the human-authorship requirement necessitates that all “original works of authorship” be created in the first instance by a human being, including those who make works for hire. 32
Interestingly, the court noted that “the human authorship requirement does not prohibit copyrighting work that was made by or with the assistance of artificial intelligence. The rule requires only that the author of that work be a human being—the person who created, operated, or used artificial intelligence—and not the machine itself. The Copyright Office, in fact, has allowed the registration of works made by human authors who use artificial intelligence.33
Can the Operation of AI Infringe Third Party Copyrights?
Generative AI models use huge databases of individual records to train statistical algorithms that will generate predictive outputs based on user prompts. Many (but not all) of these databases have been built by “scraping” the web for content. Much of that content is subject to copyrights of third party authors (e.g., anyone posting on a social network account). Generally, that scraping is conducted without the consent of the authors. This process results in potential copyright violations at both the scraping and output stages.
On the input side, AI developers have raised the statutory fair use defense, which states, “… the fair use of a copyrighted work… for purposes such as criticism, comment, news reporting, teaching…, scholarship, or research, is not an infringement of copyright. In determining whether the use made of a work in any particular case is a fair use the factors to be considered shall include—
the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
the nature of the copyrighted work;
the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
the effect of the use upon the potential market for or value of the copyrighted work.”34
At the pleading stages in some relevant cases, AI developers have contended that, under the first factor, the use of copyrighted works for the purpose of training an AI algorithm is completely unrelated to the uses of those works by their authors, and that, under the fourth factor, their use to train their models does not negatively affect the authors' potential market for the copyrighted works.
30 17 U.S.C. § 101.
31 Thaler v. Perlmutter, 2025 U.S. App. LEXIS 6294, *9 (D.C. Ct. Ap. 2025).
32 Id. at 27.
33 See Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16,190, 16,192 (March 16, 2023) (Whether a work made with artificial intelligence is registerable depends “on the circumstances, particularly how the AI tool operates and how it was used to create the final work.”)
34 17 U.S.C. § 107.
On the output side, a copyright plaintiff must demonstrate that the allegedly infringing work and the copyrighted expression "share substantial similarities."35 This will require detailed legal analysis and extended legal proceedings, which may have the practical effect of leaving many aggrieved authors without a remedy.
DMCA Liability
Distinct from the Copyright Act of 1976 is the Digital Millenium Copyright Act (“DMCA”), specifically, 17 U.S.C. §§ 1202(b)(1) and 1202(b)(3). It addresses the removal of "copyright management information" ("CMI"), which includes, among other material, a work's title, author, and copyright notice.36 Correspondingly, § 1202(b)(1) provides that "[n]o person shall . . . intentionally remove or alter any [CMI]," and § 1202(b)(3) provides that "[n]o person shall . . . distribute . . . works ., knowing that [CMI] has been removed or altered." Liability attaches only to those who removed CMI or distributed works without CMI while "knowing, or, . . . having reasonable grounds to know, that it will induce, enable, facilitate, or conceal an infringement of any right under this title."37 When this element is paired with the requirement of an "intentional[] remov[al]" of CMI under § 1202(b)(1) and "know[ledge]" of CMI's removal under § 1202(b) (3), the statute is said to impose a double scienter requirement.38
In a recent case of some note, an AI developer was accused of generating training sets using an algorithm (“Dragnet”) that omits CMI from scraped content, and then, at least on occasion, regurgitating verbatim copies of the scraped content without its CMI. At the dismissal stage, the court allowed certain claims to proceed, while dismissing others.39
Could AI Users Be Held Liable for Copyright Infringement?
Although the Copyright Act does not expressly render anyone liable for infringement committed by another, the doctrines of secondary liability emerged from common law principles and are well established in the law. One infringes contributorily by intentionally inducing or encouraging direct infringement, and infringes vicariously by profiting from direct infringement while declining to exercise a right to stop or limit it.40 Query the potential liability of an AI user who inputs a prompt that returns output that is substantially similar to the scraped content of a third party author.
Some AI user agreements disclaim any warranties of noninfringement of third party intellectual property, while strictly limiting their own potential liability, and requiring the user to indemnify the AI developer for any legal liabilities and legal costs incurred.41
35 Tangle, Inc. v. Aritzia, Inc., 125 F.4th 991, 997 (9th Cir. 2025).
36 See 17 U.S.C. § 1202(c).
37 11 U.S.C. § 1202(b).
38 See Mango v. BuzzFeed, Inc., 970 F.3d 167, 171 (2d Cir. 2020).
39 See Intercept Media, Inc. v. OpenAI, Inc., 2025 U.S. Dist. LEXIS 30147 (S.D.N.Y. 2025).
40 See MGM Studios Inc. v. Grokster, Ltd., 545 U.S. 913, 930-931 (2005).
41 See, e.g., “Copilot AI Terms and Conditions”, updated January 28, 2025, §§ 8-10. https://www.copilotai.com/terms-of-service.
IV. Trade Secrets
Trade secret law is a creature of both state and federal statutes (as well as common law in some instances). The Uniform Trade Secrets Act, published by the Uniform Law Commission (ULC) in 1979 and amended in 1985, is a model law that has been adopted (sometimes with variations) by most U.S. states. Generally, it states that a trade secret is information related to a business that is maintained as a secret, and is valuable because it is not generally known. The federal Defense of Trade Secrets Act (“DTSA”) defines “trade secret” as:
“all forms and types of financial, business, scientific, technical, economic, or engineering information, including patterns, plans, compilations, program devices, formulas, designs, prototypes, methods, techniques, processes, procedures, programs, or codes, whether tangible or intangible, and whether or how stored, compiled, or memorialized physically, electronically, graphically, photographically, or in writing if—
(A) the owner thereof has taken reasonable measures to keep such information secret; and
(B) the information derives independent economic value, actual or potential, from not being generally known to, and not being readily ascertainable through proper means by, another person who can obtain economic value from the disclosure or use of the information…”42
Thus, it is not unexpected that much of the trade secret case law addresses the question of whether the plaintiff had taken “reasonable measures to keep such information secret”. In one DTSA case, a company that shared its purported trade secrets with a third party without a confidentiality agreement and without other policies or practices for safeguarding the information was held not to have taken reasonable steps to safeguard its trade secrets, and thus to have no trade secret claim to enforce.43 However, a plaintiff was found to have taken "reasonable measures" to maintain secrecy by encrypting and compiling its source code and requiring licensees to agree to confidentiality.44 Providing alleged trade secrets to third parties may not undermine a trade-secret claim, so long as the information is provided on an understanding of confidentiality.45
Consequently, when one shares her trade secret with a third party, the conditions under which that information is shared can determine whether an enforceable trade secret survives the communication. Some AI user agreements require the user to grant a very broad license of her content to the AI developer, including rights to host, reproduce, distribute, communicate, publicly display, and use user supplied content. The AI developer’s stated purposes may include operating and improving its services, targeting users with ads, and promoting its services.46
42 18 USCS § 1839 (3).
43 Farmers Edge Inc. v. Farmobile, LLC, 970 F.3d 1027, 1033 (8th Cir. 2020).
44 See InteliClear, LLC v. ETC Glob. Holdings, Inc., 978 F.3d 653, 660 (9th Cir. 2020).
45 VBS Distribution, Inc. v. Nutrivita Lab'ys, Inc., 811 F. App'x 1005, 1009 (9th Cir.), cert. denied, 141 S. Ct. 454 (2020).
46 See, e.g., Google Terms of Service, License, effective May 22, 2024. https://policies.google.com/terms.
What is the consequence of loading otherwise confidential or secret information into an AI system as a user prompt?
The case law is not well developed at this juncture. We are left with a series of questions, some factual, some legal.
What actually happens when a user enters an AI prompt? Is the prompt saved for any of the above-stated uses? Is it subject to reproduction in whole or part for sharing with third parties (e.g., prospective purchasers of ad space)? Does the developer harvest user prompts for inclusion in its database to train its algorithms? If so, can the user prompt be regurgitated verbatim or sufficiently to destroy secrecy of the user’s prompt in response to other users’ prompts?
On the legal side, what assurances will courts need to see in AI user agreements to find that the holder of the trade secret has taken reasonable measures to maintain the secrecy of the information he shares with an AI developer in the form of a user prompt?
V. Final Thoughts
AI is a black box, and the courts and legislatures have not yet resolved (or even fully formulated) the emerging relevant issues that arise when AI meets the law of patents, copyrights, or trade secrets. Uncertainty invites caution. Seek the advice of competent counsel if questions arise.
Biography
Gerard Reinhardt is an attorney with over twenty-five years of experience in Intellectual Property, including law firm experience litigating large, complex patent cases, experience as in-house patent counsel with large branded pharmaceutical developers, and a broad palette of IP experience as a solo practitioner.
He is well-versed in drafting and prosecuting US and foreign patent and trademark applications, litigation, drafting patentability, noninfringement and invalidity opinions, licensing, due diligence, and Hatch-Waxman matters. He has written papers on a variety of legal topics, most recently regarding the intersection of artificial intelligence and intellectual property.
Gerard is licensed to practice law in New York, Texas, Florida and the District of Columbia, and he is registered by the US Patent and Trademark Office as a patent attorney. He holds BS and MS degrees in Chemical Engineering from the University of Virginia and Georgia Tech, respectively, a JD from the University of Houston, and an MBA from the University of South Florida. Prior to starting his law career, Gerard practiced Chemical Engineering for fifteen years for Raytheon and United Technologies.
Disclaimer: Nothing in this paper shall constitute legal advice, and the provision of this paper does not establish an attorney-client relationship. Consult competent counsel in case of questions.