On May 18, 2023,  the Supreme Court found that artistic changes to a pre-existing work, alone, not necessarily sufficient to make a derivative work fair use. The actual use of the derivative work is important as well, and it is not fair use if the derivative work is used for the purpose of competing with or supplanting the underlying work used to create the new work.

At issue before the Supreme Court in Andy Warhol Foundation for the Visual Arts Inc. v. Goldsmith was whether or not Warhol’s use of Goldsmith’s photograph as a reference and departure point for the creation of an image of Prince constituted fair use or copyright infringement under U.S. copyright law. The Supreme Court narrowed its attention to the only issue before it, the use of the image in a magazine instead of Goldsmith’s photo, as opposed to as a series of paintings/lithographs.

Applying a new lens on how to view the purpose of a derivative work under U.S. copyright law,  the Supreme Court focused on the actual use made, i.e. what the user does with the original work. While Andy Warhol made artistic changes and added new elements by cropping, coloring, and make other artistic additions/changes to Goldsmiths photo, it did not establish fair use since the use of the derivative work before the court (use in a magazine as an image of Prince) was competitive with the underlying work. Therefore, even though Warhol’s efforts turned Goldsmith’s depiction of Prince as a vulnerable, uncomfortable person into presenting him as a larger than life icon, the majority found that this was not sufficient alone to make the work transformative and thereby support a finding of fair use.

Copyright law in the U.S. goes all the way back to our Constitution, which gave Congress the power “To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” Thus, copyright law provides authors with the right to control the use of their creative works. Contrasting, Congress, in enacting our copyright law determined that there are certain uses of another creator’s work that are beneficial to society and promoting freedom of expression. The legal concept of fair use is designed to balance these two conflicting goals — protect the financial and artistic interests of authors while permitting certain positive uses of those same works.

Fair use actually permits the unlicensed use of copyright-protected works in certain circumstances. So, something that might otherwise be infringing — the reproduction of portions of a book for educational or criticism purposes — is determined to be beneficial and not considered infringement. U.S. copyright law identifies certain types of uses — such as criticism, comment, news reporting, teaching, scholarship and research — that are particularly important to our society. The determination of fair use requires assessing four separate factors, only one of which was at issue before the Supreme Court:

“The purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes.”   

The Supreme Court had previously endorsed a “transformation” test to apply to this factor- whether the new work merely ‘supersede[s] the objects’ of the original creation, (‘supplanting’ the original), or instead adds something new, with a further purpose or different character that serves to transform the underlying work into essentially something else.  Allowing Warhol’s work to literally replace Goldsmith’s work as a depiction of Prince in a magazine did not, in the majority’s view, change the purpose of the work as a depiction of Prince, and thereby protected Goldsmith’s right as a copyright owner to authorize derivative works which is one of the exclusive rights granted a copyright owner under U.S. copyright law. The Supreme Court noted how this was different than Andy Warhol’s famous Campbell soup cans. Those works were “artistic commentary on consumerism, a purpose that is orthogonal to advertising soup”. The Court noted also how this contrasted from the Supreme Court’s finding that 2 Live Crew’s use of Roy Orbison’s song “Pretty Woman”  was a clear commentary on the prior song and provided new meaning that was not designed to compete with or supplant the pre-existing work.

Ultimately, it seems that the Supreme Court went to great lengths to minimize the impact of its decision.  Holding in Warhol’s favor could have made the “transformation” test a one factor fair use test — focusing solely on an aesthetic transformation or the purpose of the artist would make it difficult for any work to be deemed derivative, i.e., infringing, as opposed to transformative and therefore protected by fair use. Some artists were concerned that to allow such appropriation without limits would undercut the secondary market for original works of art.   As Goldsmith’s counsel, Lisa Blatt, argued before the Court that  “adding new meaning to a work is simply not a good enough reason to copy for free, ” and that such a result “would decimate the art of photography by destroying the incentive to create the art in the first place.” 

In this way the Supreme Court honors the original purpose of the copyright laws as set out in the Constitution and copyright law, rather than undercut by making free appropriation the rule rather than the exception. This decision also falls in line with other equity-based judicial determinations where financial considerations form an unstated additional fair use factor balancing what is really fair about not paying for the use vs. the profit made from the use.  From that perspective, regardless of Warhol’s creativity it is understandable to view Goldsmith as the owner of rights to her work and see that Warhol had simply made too much use of the photograph in replacing it to ethically and equitably avoid paying licensing fees.

Here, the Supreme Court did not go that far by limiting its application of the “purpose” test to one of direct competition and supplanting of the original work, leaving many questions unanswered, including potentially whether Warhol’s Prince series—as a series of artworks not published in a magazine- would constitute fair use.  

 Most nonpracticing entities are private firms owned by patent attorneys, investors, or a combination of the two. In other cases, NPEs do everything they can to obscure the identify of the true owner. And in virtually all cases, NPEs purchase the asserted patents from another company seeking to divest assets for cash.

Leigh Rothschild is a bit different. Mr. Rothschild is typically both the inventor of the asserted patent and the owner of the company asserting the patents. Some have called him a modern day Thomas Edison given the hundreds of patents for which he is a listed owner—ranging from an internet connected drink mixer to a telecommunications method for transferring media content. Others have accused him of being a “patent troll” based on the broad interpretation of his patents and his litigious enforcement activity.

A fair assessment likely places Mr. Rothschild somewhere in between. His entities have filed a whopping 1,249 patent infringement lawsuits with 23 of those pending. This certainly qualifies as a high volume patent troll. And Rothschild has been hit with fee awards for filing objectively baseless lawsuits against defendants who clearly did not infringe his patents. This suggests at least some of his cases are quite weak. All troll-like behavior, but from an inventor who is creating the patents rather than mining for them.

Rothschild is likely better described as someone who is capable of slipping a patent application by the Patent Office in very broad terms, then asserting it against technology that is completely different than the invention in the patent. Take the drink mixer invention above. Mr. Rothschild argued that invention was broad enough to cover the concept of remote configuration. Here is claim 1 of that patent:

1. A system for customizing a product according to a user’s preferences comprising: a remote server including a database configured to store a product preference of a predetermined product for at least one user; and

a first communication module within the product and in communication with the remote server;

wherein the remote server is configured to receive the identity of the predetermined product and the identity of the at least one user, retrieve the product preference from the database based on the identity of the predetermined product and the identity of the least one user and transmit the product preference to the first communication module.

The patent examiner was clearly fooled by the invention of the application (a remote drink mixer) and did not mean to allow a claim that was so broad. But that’s the game – and Rothschild plays it well.

Fast forward to today and you’ll find Rothschild filing cases in a time when IP Edge has effectively halted its patent enforcement activities. Of course, IP Edge reduced their filings in response to Judge Connolly forcing the disclosure of the owners of IP Edge entities. Rothschild has no issue offering up his ownership and, in fact, enjoys the perception of the Thomas Edison-like status that comes with it.

Patent applications are often filed at the last minute. Priority considerations force the application to be filed expeditiously to avoid statutory bars or catastrophic public disclosure issues. But filing too quickly—or in the wrong country—may cause the eventual patent to be invalid for failure to obtain a foreign filing license.

What is a Foreign Filing License?

A foreign filing license is a communication from the U.S. government granting the applicant the right to file in a country other than the United States. These licenses are granted only after the Patent Office determines the invention does not relate to subject matter that would be a threat to national security as set forth in 35 U.S.C. § 181. As an extreme example, a patent application directed to a nuclear weapon should not be filed in a foreign patent office because it would explain to the foreign country how to build a nuclear weapon.

But what if you’re not in the nuclear weapons business? Unfortunately the requirement for a foreign filing license applies to every type of invention. Regardless of the invention, the resulting patent will be deemed invalid if the application is filed in another country prior to the grant of a foreign filing license. It is therefore critical to first file in the United States and then file elsewhere only after receiving the foreign filing license (typically provided with the filing receipt just a few weeks later). Or, the applicant can file a PCT application in the U.S. receiving office and nationalize in the PCT member countries within the allotted time period.

An important consideration is that a foreign filing license is required only for inventions first made in the United States. Many other countries have similar rules. An invention made in China must first be filed as a patent application in China, for example, and only afterwards filed in the United States. It is important to determine where the invention was first made to then determine where to first file the application.

What if You File an Application Without a Foreign Filing License?

Mistakes happen and applications are often filed abroad prior to the issuance of a foreign filing license. If left uncorrected, the patent would be invalid. But all is not lost—the PTO permits the applicant to petition for a retroactive foreign filing license to effectively reverse the error. The requirements for such a petition are (a) a listing of the countries in which the application was filed and the dates of filing; and (b) a verified statement that (i) the invention was not subject to a secrecy order, (ii) the retroactive foreign filing license was diligently sought immediately after the discovery of the error; and (iii) the foreign filing was done through error and without deceptive intent.

By far, the most onerous portion of the petition is the showing of error and no deceptive intent. The Patent Office is quick to dismiss petitions that provide “conclusory statements” and will only grant a petition that is supported by substantial facts. Why was the error made? Does the law firm have a policy that prevents this issue from happening?  Why did the attorney not contact the docket clerk before instructing the foreign associate to file? Was the applicant itself aware of the issue?

The factual record must also be supported by documentation. Often times, it may be necessary to have the applicant or attorney sign an affidavit. It may also be necessary to provide emails between the attorney and client with important information redacted. It is, to be blunt, a pain.

At bottom, the Patent Office makes the petition process difficult because they want to prevent lawyers from using the process as a default solution. Perhaps the Patent Office scrutinizes these types of petitions because it believes attorneys were well aware of the requirement to wait for a foreign filing license but ignored it out of convenience. It is therefore best to carefully analyze whether a foreign filing license has been granted prior to filing a patent application—the alternatives are not kind.

Last week, a joint statement was issued by four federal agencies expressing their apprehension regarding the use of AI for discriminatory or anticompetitive purposes and outlining their plans for regulation. This comes on the heels of Elon Musk requesting a “pause” in AI development and meeting with Senator Chuck Schumer to guide the statutory framework of AI management. It also comes as China and the EU unveiled their own proposed laws surrounding the regulation of AI. How will governments across the globe respond to this pivotal technology? And for the United States, how will our government respond?

1. Joint Statement

The Consumer Financial Protection Bureau (CFPB), the Justice Department’s Civil Rights Division, the Equal Employment Opportunity Commission (EEOC), and the Federal Trade Commission (FTC) released a Joint Statement on Enforcement Efforts Against Discrimination and Bias in Automated Systems (“Joint Statement,” found here). The Joint Statement expressed the collective concern that AI would be used for discriminatory or anticompetitive purposes.

  • The CFPB is concerned about the discriminatory use of AI in lending decisions. “[T]he fact that the technology used to make a credit decision is too complex, opaque, or new is not a defense for violating these laws.”
  • The EEOC similarly cautioned against the use of AI “to make employment-related decisions about job applicants and employees” and made clear that there is no technology exception to its rules.
  • The Civil Rights Division followed suit, explaining “the Fair Housing Act applies to algorithm-based tenant screening services.”
  • The FTC predictably focused on the potentially anticompetitive impact of AI and “warned market participants that it may violate the FTC Act to use automated tools that have discriminatory impacts, to make claims about AI that are not substantiated, or to deploy AI before taking steps to assess and mitigate risks.”

This is all to say that one cannot outsource its liability to AI. In the realm of decision-making and data-driven solutions, pleading ignorance to the ways in which AI is utilized is not, nor will it ever be, an adequate defense, whether presently or in the future. Many AI algorithms are a black box to the user, so the views of the Joint Statement will be difficult to mesh with the practical use of artificial intelligence tools.

2. International AI Laws

Countries across the globe are rushing to institute their own regulation of AI in response to how it is implemented in their societies. China’s law would require AI models to pass a security assessment before the tool is publicly released. This development is in response to the recent unveiling of numerous generative AI tools by large Chinese companies, which have the capability to simulate human speech and generate text and image-based content.

The EU AI Act would assign three categories to AI tools based on their risk level. The most egregious tools are assigned “unacceptable risk, such as government-run social scoring of the type used in China.” These would be outright banned. Less egregious tools would be deemed “high-risk applications, such as a CV-scanning tool that ranks job applicants” and such tools would be regulated, but not per se banned. Those models not falling within either category would be “largely left unregulated,” whatever that means.

3. What will the United States do?

The Joint Statement makes clear that there is no AI exception to current regulatory rules. It also implies the regulatory agencies believe its framework is already set. For example, using an AI tool to discriminate on the basis of gender or race is illegal—just as it always has been. Anticompetition is anticompetition—AI or not. It appears the agencies are awaiting congressional action for more aggressive oversight on AI.

How Congress reacts will be interesting to watch. The stakes could not be higher and legislation will certainly take time. It is helpful that Congress is seeking guidance from the private sector and understanding the ways in which this revolutionary technology is being used. Perhaps they should seek guidance from AI itself?

4. What AI Has to Say About…Regulation of AI

I had the following exchange with ChatGPT to see what it felt would be effective AI regulation in the financial lending industry:

Pat: I work for the government. We are concerned that artificial intelligence may cause financial firms to discriminate in their lending. We are also concerned that companies will use artificial intelligence to gain an anticompetitive advantage. What are some policies that may prevent such abuses of artificial intelligence?

ChatGPT: There are several policies that could help prevent discrimination in lending and anticompetitive behavior resulting from the use of artificial intelligence (AI):

  1. Algorithmic transparency: Requiring financial firms to disclose the data, models, and algorithms used in their lending decisions would make it easier to detect and address discriminatory practices.
  2. Ethical guidelines: Developing ethical guidelines for the use of AI in financial services could help ensure that algorithms are designed and implemented in a way that is fair and non-discriminatory.
  3. Oversight and regulation: Government agencies could establish oversight and regulation to ensure that AI applications in financial services comply with ethical and legal standards, and do not lead to unfair market practices.
  4. Anti-discrimination laws: Enforcing anti-discrimination laws can help prevent lending decisions based on factors such as race, gender, or religion, even if those factors are indirectly inferred from other data.
  5. Competitive safeguards: Promoting competition by preventing the accumulation of market power by large firms could help prevent anticompetitive practices enabled by the use of AI.
  6. Data protection and privacy: Ensuring the protection of customer data and privacy rights could also help prevent discriminatory and anticompetitive practices enabled by AI.

It will be interesting to see how government responds and who—or what—they listen to. We will certainly keep you updated as developments arise.

Don’t worry, machines haven’t completely replaced humans as artists—at least, not yet. But the U.S. Copyright Office is considering the possibility.

The Copyright Office recently declared that it will not grant protection over AI-generated works, upholding its longstanding rule that non-human authors cannot own copyright. At the same time, the Office is well aware that AI technology is changing everything. It is seeking public input on AI-related issues throughout the month of May, starting with a listening session on AI and visual arts from 1 to 4 pm Eastern today. A session on audiovisual works is scheduled for May 17, and another for music and sound recordings on May 31. Members of the public can access Zoom links through the Office’s website.

As it stands, under U.S. law there must be “some element of human creativity” for material to be copyrightable. In a registration decision issued February 21, 2023 regarding a comic book entitled “Zarya of the Dawn,” the Copyright Office explained that it would only issue a registration covering the compilation of images and text created by the human author. The registration would not cover AI-generated images contained within the compilation.

The decision relied on the 1884 U.S. Supreme Court case Burrow-Giles Lithographic Co. v. Sarony, in which the Court grappled with what was then the newly emerging technology in art: photography. Although the process of the camera capturing the photograph was purely mechanical, the Court explained, the photographer’s artistic choices in composing the photograph were protectable because they were instrumental in the ultimate output. The Copyright Office drew a parallel to Zarya of the Dawn, explaining that “the Work is the product of creative choices with respect to the selection of the images that make up the Work and the placement and arrangement of the images and text on each of the Work’s pages.” In contrast, the images themselves were created using Midjourney, an AI tool that generated visual depictions based on text prompts. “While additional prompts applied to one of these initial images can influence the subsequent images, the process is not controlled by the user because it is not possible to predict what Midjourney will create ahead of time,” the Office concluded. Therefore, the process of creating the images “is not the same as that of a human artist, writer, or photographer.”

The Office followed up with formal guidance published in the Federal Register on March 16, 2023, explaining that the term “author” as used in the Intellectual Property Clause of the U.S. Constitution and in the Copyright Act “excludes non-humans.” Therefore, when applying for a copyright registration of a work that contains AI-generated content, the AI-generated content must be explicitly excluded from the application. If the applicant fails to disclose which portions of the work were AI-generated, and the Office later becomes aware of that fact, the resulting registration may be cancelled.

In fact, this is what happened in connection with Zarya of the Dawn. After originally obtaining a registration covering the entire comic book, the book’s human author boasted about it on social media, generating interest among journalists. The Copyright Office saw the attention its original decision had gotten, and decided to revisit the issue sua sponte, revoking part of the registration.

The Zarya decision does provide helpful guidance on the current state of copyright law and AI, but the lines between human-generated art and AI-generated art remain blurry. For example, although Midjourney actually created the output of images in Zarya, a human had to come up with the text prompt to initiate the process. Should the human author be allowed to claim copyright in the prompt? And, if so, would that right actually be meaningful?

Generally, it is not possible to obtain copyright in short phrases. The Copyright Office has concluded that short phrases do not contain a sufficient amount of creativity to be copyrightable. Even so, depending upon the length and complexity of a given prompt in an AI tool, there may be room to argue that authors are making creative choices affecting the ultimate output generated by the tool. If a particular tool generates consistent output based on the same prompt, copyright could help artists exclude others from using identical prompts to obtain identical outputs.

At the same time, under the Copyright Office’s current policy, there would be no copyright in the output itself. In other words, no one would own rights in that output, so copying it wouldn’t constitute infringement. Therefore, even if the artist were able to prevent another individual from using a particular prompt to re-generate an output, there would be nothing stopping that same person from simply making a digital copy of the output itself. Copyright in the original prompt would be toothless in this context.

It will be interesting to see how all of these dynamics play out as artists and authors increasingly rely on AI tools as part of their creative process. It will also be interesting to see how the Copyright Office balances the current legal framework preventing copyright over AI-generated content with the original goal of copyright law as articulated in the Constitution: “To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.”

kaung-myat-min-0SJj3nUIW_0-unsplash

This post was originally published to Seyfarth’s International Dispute Resolution Blog.

There is a little-known provision of the Lanham Act (the US Trademark Act) that packs a potentially big punch. 15 USC § 1051(e) provides that if a non-U.S. entity registers for a trademark in the United States without designating a United States resident for service of “notices or process in proceedings affecting the mark” (a “Domestic Representative”), or if the Domestic Representative cannot be found, then service, including of court pleadings commencing and related to a lawsuit, can be accomplished on the non-U.S. entity by serving the Director of the United States Patent and Trademark Office (USPTO).  The practical effect of this provision is that certain non-U.S. entities who apply for trademarks in the United States could find themselves a party to U.S. litigation if the non-U.S. entity is never directly served with court papers.

Recent Decisions Apply Section 1051(e) to Court Proceedings

Two recent cases suggest that Section 1051(e) may not be little-known much longer.  In San Antonio Winery, Inc. v. Jiaxing Micarose Trade Co., Ltd.,[1] the plaintiff sued a foreign trademark applicant and served the court papers on the USPTO Director, after which the USPTO sent the papers to the foreign applicant.  After the foreign applicant failed to appear in court, the plaintiff sought a default judgment.  The district court rejected that application, holding that Section 1051(e) only applies in USPTO administrative proceedings, not court cases.[2]  On appeal, the Ninth Circuit vacated the district court’s decision.  It found that the plain meaning of “proceeding” includes court cases because court cases can “affect” a trademark.  This includes, among other things, determining whether someone has the right to register a mark, canceling a registered mark or restoring canceled registrations.[3]  The Ninth Circuit also relied on Section 1051(e)’s references to service of notices or “process,” which it held applied to service of process in a court case.[4]

The Ninth Circuit also held that Section 1051(e) does not conflict with the Hague Service Convention, which provides procedures for service on non-U.S. defendants in countries that are signatories to the Convention.[5]  The appeals court reasoned that the Convention applies only if “the method of service at issue ‘require[s] the transmittal of documents abroad.’”[6]  Because Section 1051(e) involves domestic service on the USPTO Director, it “falls outside the scope of the Convention.”[7]  Following the Ninth Circuit’s decision, the district court entered the default judgment sought by the plaintiff.[8]

Soon after the Ninth Circuit’s decision, a federal court in New York followed the same approach.  In Equibal, Inc. v. 365 Sun LLC,[9] two Brazilian entities that had applied for U.S. trademarks had listed U.S. attorneys as their counsel in their applications.  But the attorneys were no longer in contact with the Brazilian entities or authorized to accept service on the entities’ behalf.[10]  Neither entity had designated a Domestic Representative. Under those circumstances, the New York court found that Section 1051(e) applies in court proceedings, and concluded that service via the USPTO Director was proper.

The New York court held that service on the USPTO Director satisfied Federal Rule of Civil Procedure 4(f), which provides that individuals or entities outside the United States can be served: (1) by means set forth in an international agreement, such as the Hague Convention or the Inter-American Convention; (2) if there is no applicable international agreement, by means calculated to give reasonable notice; or (3) by means not prohibited by an international agreement that a U.S. court may order.[11]  The New York court found that the third option was satisfied.  Like the Ninth Circuit, the New York court concluded that neither the Hague Convention nor the Inter-American Convention (to which Brazil is a signatory) bars service on the USPTO Director.  The New York court found that service on the USPTO Director also satisfied due process because the Brazilian entities were “explicitly warn[ed]” by Section 1051(e) that if they availed themselves of U.S. trademark protection, they could be served via the USPTO Director if they failed to designate a Domestic Representative or their Representative could not be found.[12]  The court concluded by finding that service on the USPTO Director was warranted in the case before it in light of the difficulty of service through other means.[13]

Points for Plaintiffs

Section 1051(e) is a potentially useful tool for Lanham Act plaintiffs seeking to sue non-U.S. entities in connection with marks those entities apply to register in the United States.  If accepted in the particular judicial Circuit, service through Section 1051(e) could allow plaintiffs to avoid procedures like those set forth in the Hague Convention, which can often be cumbersome and expensive.  The statute, as interpreted by the Ninth Circuit and New York court, may also permit service via Federal Rule of Civil Procedure 4(f) on entities in countries that are not signatories to the Hague Convention, such as Algeria, Ghana, Kenya, Laos, Lebanon, Nigeria, Uganda, and Yemen.

It is worth noting, however, that the New York court was inclined to permit service in part because the plaintiff had made efforts to serve the Brazilian entities directly, and then sought leave to serve the USPTO Director when those efforts failed.  Plaintiffs who rely solely on serving the USPTO Director, but do not attempt to serve via other means, may find courts to be less receptive to the application of Section 1051(e).

In addition, Lanham Act plaintiffs should remain aware that, simply because service of process could be made through the USPTO Director, that in and of itself does not subject the foreign trademark applicant to personal jurisdiction in a federal district court.  The cases interpreting Section 1051(e) were based on the defendants’ United States use of allegedly infringing marks that they had also applied to register.

Points for Non-U.S. Entities

Non-U.S. entities should consider designating a Domestic Representative when they apply to register a trademark in the United States.  This should be a fairly easy thing to do, as all foreign-domiciled trademark applicants, registrants, and parties to proceedings must be represented by a U.S. attorney,[14] and attorneys may, and quite often are, appointed as foreign entities’ Domestic Representatives.  This said, designees may retire, pass away, or have a change in contact information.  Non-U.S. entities should find a trusted agent they can designate as Domestic Representative and share that agent’s details with their trademark attorneys.  That way, they ensure domestic representation for the long-term and can keep tabs on that agent so that they can update their Domestic Representative if the need arises.  Non-U.S. entities may want to consider using a corporate agent as a Domestic Representative to ensure continuity and predictability, with the caveat that the non-U.S. entity needs to ensure that the corporate agent has up-to-date contact information for the non-U.S. entity.

Non-U.S. entities should also take care to make sure that the USPTO has the most up-to-date contact information for the non-U.S. entity, in the event that service does go through the USPTO director.  In the Ninth Circuit case, the USPTO mailed the court documents to the non-U.S. entity at the address the entity had included in its trademark application.  If that address is out of date, the non-U.S. entity may never receive the court papers even if the USPTO sends them promptly.

Conclusion

Section 1051(e) is an important provision of the Lanham Act that, as recently interpreted by the Second and Ninth Circuit Courts of Appeal, can have a potentially significant impact on both non-U.S. entities registering trademarks in the United States and Lanham Act plaintiffs seeking to sue those entities.  Trademark registrants and litigants should pay close attention to both the statute and decisions interpreting that statute.

[1] 53 F.4th 1136 (9th Cir. 2022). 

[2] Id. at 1139-40. 

[3] Id. at 1141. 

[4] Id.

[5] Id. at 1143. 

[6] Id. 

[7] Id. at 1143-44. 

[8] Case No. 20-cv-9663-GW, ECF No. 57 (C.D. Cal. Nov. 28, 2022).

[9] 2023 U.S. Dist. LEXIS 62759 (S.D.N.Y. Apr. 10, 2023).

[10] Id. at *18-19. 

[11] Fed. R. Civ. P. 4(f). 

[12] 2023 U.S. Dist. LEXIS 62759, at *22. 

[13] Id. at *22-24.

[14] 37 C.F.R. § 2.11(a).

The Supreme Court yesterday declined to hear a case brought by a computer scientist whose “invention” was in fact created by artificial intelligence. Stephen Thaler was appealing a Federal Circuit decision that interpreted the Patent Act to require a human “inventor” for purposes of obtaining a patent. The invention at issue was conceived of by Thaler’s AI model DABUS and not by a human, dooming its chances of obtaining patent protection.

The law of inventorship is now quite murky. Professor Dennis Crouch wrote yesterday about the ethical issues presented to patent attorneys in the wake of the decision. If AI-created inventions are not patentable, then is a patent attorney required to interrogate their client to determine which parts of the invention were human versus AI-created?

Of course, the elephant in the room is this: what inventions will be protectible when every R&D operation utilizes artificial intelligence to get ahead? Extending to copyright law, where Thaler is fighting a similar battle, isn’t there a “modicum of creativity” leading to the AI-created work? History suggests the law will split inventorship down the middle and grant rights to those “portions” of an invention or copyrightable work that were created by a human, and avoid granting protection to the “portions” that were AI-generated.

Copyright law does this well with the protection of derivative works. The term “Derivative Work” is defined in 17 U.S.C. § 101:

A “derivative work” is a work based upon one or more preexisting works…

Examples of creative works include a translation of a novel into another language. Or a screenplay adaptation of a novel. Or—remember in the 90s when rappers would sample 80s music? The end song would be a derivative work of the 80s tune. The “author” of the end work (translation, screenplay, 90s rap song) owns the copyright on the derivative work. But the author must have a license to the original copyrighted work, or the original work must not be copyrighted (e.g., an expired copyright). In this manner, copyright law separates the original creative expression from the preexisting copyrighted work. The same can be done for AI-generated works, with the human expression protected to the extent possible, and the AI-generated expression excluded from protection.

Patent law has also separated protectible and unprotectible portions of the inventions. For example, 35 U.S.C. § 101 controls which inventions are eligible for patent protection. As early as 1852, the law precluded the patentability of abstract ideas that would effectively preempt the abstract idea. Le Roy v. Tatham, 55 U.S. (14 How.) 156, 175 (1852) (“A principle, in the abstract, is a fundamental truth; an original cause; a motive; these cannot be patented, as no one can claim in either of them an exclusive right.”). More recently, the Supreme Court maintained that an abstract idea was not patentable, but a practical application of an abstract idea may be worthy of patent protection. Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 573 U.S. 208, 216 (2014) (“[I]n applying the §101 exception, we must distinguish between patents that claim the ‘buildin[g] block[s]’ of human ingenuity and those that integrate the building blocks into something more”). Look for patent law to handle AI inventions in much the same way and to continue carving out patent protection of AI-generated inventions through the lens of §101. What was created by AI cannot be protected and will be carved out, much the way anything other than a “practical application” of an abstract idea cannot be patented. But the elements of human ingenuity giving rise to the invention will surely be protected to continue the incentive for technological innovation.

Seyfarth Brian Michaelis, Matthew Brekus, and JC Zwisler authored an article titled “Calif. Ruling May Sow Seeds Of Cannabis Patent Precedent” in Law360 on April 19. The Seyfarth attorneys discussed a California ruling from last year, Gene Pool Technologies Inc. v. Coastal Harvest LLC, which states patent owners in the cannabis industry may be able to enforce their rights against infringers in federal court. You can read the full article here.

 A new track by Drake and The Weeknd is going viral on social media and is bound to top the charts. The song “Heart on my Sleeve” details the emotional trauma of The Weeknd’s breakup with pop star Selena Gomez in a tear jerker that extends just over two minutes. The song has been hailed as the “banger of the weekend” and was viewed more than 8.5 million times on TikTok. The only problem—the song is generated entirely by artificial intelligence and not one bar was written, rapped, or produced by either Drake or The Weeknd.                

Several generative AI models now provide the ability to simulate your favorite artist’s voice and lyrical tempo to create a track that sounds indistinguishable from the real thing. All a user needs to do is record reference vocals, add musical background, and voila – it now sounds like it was written by Drake. Or The Weeknd. Or Kanye West. Or, really, anyone in the music industry that the user chooses. This video provides a great example of just how easy it is to create a song that sounds like it was written by Kanye West when in fact it was created in 20 minutes by a lyrically-challenged Youtuber.

The results are shocking and have significant implications for the music industry. Several artists have voiced their displeasure for these impersonations and music giants are now pushing streaming services to ban music created by AI models with some success (the Drake/Weeknd song has been removed from all streaming services). But legally, how can the artists fight back?

I. Copyright

Copyright protection is unlikely to help artists whose voice and musical style are simulated by AI models. The models are trained by the artist’s vocals but do not reproduce any of those vocals when creating the AI-generated song. There is nothing “copied” by the AI models and therefore no copyright to infringe.

II. Data Privacy

Data privacy may eventually create a path to protection but this will not come easy. The creators of the AI models are unlikely to be contractually bound to restrictions on the use of the voices in songs they own, at least not yet. And state or federal laws are unlikely to restrict use of publicly available songs on privacy grounds.

III. Right of Publicity

The most likely source of legal relief may be found in the artist’s right of publicity. A “right of publicity” can be described as a trademark for a person’s celebrity. “In cases involving confusion over endorsement by a celebrity plaintiff, ‘mark’ means the celebrity’s persona.” White v. Samsung Elecs. Am., Inc., 971 F.2d 1395, 1400 (9th Cir. 1992).

The concept was first created when a company released an advertisement with a robot that was meant to imitate TV star Vanna White. Id. White successfully litigated the dispute and was found to own a right in her “persona”, which was later defined as “the aspect of a person’s character that is displayed to or perceived by others.” Roberts v. Bliss, 229 F. Supp. 3d 240, 249 (S.D.N.Y. 2017). And so the “right of publicity” was born!

Courts later went on to limit the right of publicity to extend only to the celebrity’s name, image, or “a symbol or device such as a visual likeness, vocal imitation, or other uniquely distinguishing characteristic.” Waits v. Frito-Lay, Inc., 978 F.2d 1093, 1110 (9th Cir. 1992). The term likeness “does not include general incidents from a person’s life.” Matthews v. Wozencraft, 15 F.3d 432, 438 (5th Cir. 1994). Nor does it extend to “any attribute of an individual.” Martin v. Living Essentials, LLC, 160 F. Supp. 3d 1042, 1046 (N.D. Ill. 2016), aff’d, 653 Fed. Appx. 482 (7th Cir. 2016). So is a celebrity’s voice protected?

One case appears to be directly on point. The case of Waits v. Frito-Lay, Inc. held a celebrity’s right of publicity extends to “a symbol or device such as a visual likeness, vocal imitation, or other uniquely distinguishing characteristic.” 978 F.2d 1093, 1110 (9th Cir. 1992). In this case, the court held singer Tom Wait’s right of publicity extended to his voice, because “[a] voice is as distinctive and personal as a face.” Id. at 1100.

Here, the Drake/Weeknd/Kanye AI models are creating indistinguishable simulations of the real artist’s voice. They are profiting off the celebrity of these artists by imitating the artist’s voice, even though the model does not reproduce any collected snippets of the artist’s voice. The end result is the type of imitation that Waits holds to be violative of the right of publicity.

Generative AI will certainly alter the arts in a manner never seen before. All may not be lost for the music industry, but look for artists to focus more on their celebrity than their artistic talent when monetizing their skillset. Of course, there’s nothing stopping these artists from using the same AI models to provide cleaner versions of their own songs or avoid the costs of recording an album from scratch. If a Youtuber can create a song in 20 minutes, why couldn’t an established musician?

If there is anything movies like The Terminator have shown us, it’s that AI systems might one day become self-aware and wreak havoc.  But until Skynet becomes self-aware, let’s enjoy the AI toy that is quickly becoming a part of our daily lives. Some Samsung employees recently discovered that playing with AI models like ChatGPT may have unexpected consequences. These employees used ChatGPT for work and shared sensitive data, such as source code and meeting minutes. This incident was labeled as a “data leak” due to fears that ChatGPT would disclose the data to the public once it is trained on the data. In response, many companies took action, such as banning or restricting access, or creating ChatGPT data disclosure policies.

First, let’s talk about ChatGPT’s training habits. Although ChatGPT does not currently train on user data (its last training session was in 2021), its data policy for non-API access says it may use submitted data to improve its AI models. Users are warned against sharing sensitive information, as specific prompts cannot be deleted. API access data policy is different, stating that customer data is not used for training/tuning the model, but is kept for up to 30 days for abuse and misuse monitoring. API access refers to access via ChatGPT’s API, which developers can integrate into their applications, websites, or services. Non-API access refers to accessing ChatGPT via the website. For simplicity, let’s focus on non-API access. We’ll also assume ChatGPT has not been trained on user data yet – but, like Sarah Connor warning us about Judgment Day, we know it’s coming. Our analysis will mainly focus on ChatGPT.  As noted below, this analysis may change based on a given usage policy of a chatbot.

This situation brings to mind the classic philosophical question: If a tree falls in a forest and no one’s around to hear it, does it make a sound? In our AI-driven world, we might rephrase it as: If we share our secrets with an AI language model like ChatGPT, but the information remains unused, does it count as trade secret disclosure or public disclosure of an invention?

Continue Reading Spilling Secrets to AI: Does Chatting with ChatGPT Unleash Trade Secret or Invention Disclosure Dilemmas?