Category: Copyright

Udio Answers Record Labels’ Complaint in Gen AI Lawsuit

As mentioned in my last post about the record labels’ lawsuits against GAI companies Suno and Udio, I will generally focus on the latter case. Both cases are almost identical, but because UMG et al. v. Ucharted Labs Inc. is at the SDNY (in the Second Circuit), those proceedings may be followed by other courts with considerably less copyright law precedent.

Udio’s answer filed on August 1 relies substantially on the premise that there is no cause of action whatsoever. The developer intends to show that “This lawsuit…seeks a genuinely unprecedented result: a ruling that it is actionable copyright infringement, not fair use, to have copied Plaintiffs’ works as part of the process of developing a new technology, even though the ultimate outputs of that new technology are themselves non-infringing.”

The list of counter-factual evidence Udio forecasts is too long to summarize, but the heart of its legal argument at this juncture is that 1) statute explicitly bars protection of musical style; 2) its AI training process entailed learning about music rather than copying protected sound recordings; and 3) because the product’s outputs are largely (or entirely) non-infringing, the purpose of producing “new” music is “what copyright law is designed to encourage, not prohibit.” The defendant also alleges that the recording industry’s claim is invalid on the basis that it has “misused copyright law” as part of a longstanding tradition of stifling competition. So, we have a ballgame that’s going into extra innings, and there will be plenty to say about the details as they emerge.

For now, I take issue with a few premises inherent to Udio’s answer. The first, which I have already stated a few times, is that even if every song output by Udio is “new” as a matter of law, the fact that none of these outputs is a work of “new authorship” as a matter of law militates against Udio’s implication that its product expands the purpose of copyright in general. And as stated, if that is correct, this should militate against a finding of fair use.

Second, despite the fact that Udio can and does point to uses of its product by creators who are plausibly engaged in authorship is, at best, a difficult basis on which to argue that the primary purpose of the product advances authorship. For one thing, the business model appears to be based largely on providing a music toy for consumers, not a tool for creators. Next, even where Udio may be used by professional music creators, the extent to which this fosters new authorship is a case-by-case consideration—one that relies on still-developing doctrine around the use of AI and authorship.

Third, even if Udio could prove allegations of relevant, anti-competitive practices among all the record labels (and I do not mean to suggest they can), the court must remain focused on the interests of individual creators—especially the next generation of music makers. The labels’ argument that the outputs compete with demand for existing sound recordings could be read as protectionism of existing catalogs but should be considered as to whether Udio competes with, or even obviates, the need for new human authorship in music. If so, this is categorically not what copyright law is designed to foster.

As stated in a few posts, and in comments to the Copyright Office, the unique challenge presented by GAI is that rather than pose a threat to the interests of specific authors’ works, it poses a potential threat to authorship itself. In this light, Professor Jane Ginsburg, in a new paper about the state of fair use jurisprudence, discusses two points that stand out for me at the moment. First, she describes the nature of a use-based fair use analysis (as applied in Warhol), which should not “untether” the fair use protection for a use other than the one narrowly ruled on by the court. Second, she notes that the courts may look beyond the “explicit direction” of the fair use statute to consider a factor like broad effect on authors’ careers—or even the potential for other unlawful uses like forgery or fraud.

With regard to use-based analysis, Ginsburg forecasts the uncertainty in adopting a per se fair use rule for machine learning because the consideration of fair use of the inputs may turn on the nature of the outputs. “If an AI system ingests multiple images of apples, including Cézanne’s depictions (let’s assume Cézanne’s works were still under copyright) its training data will enable the system to “know” both what an apple looks like, and what a Cézanne apple looks like. The fair use inquiry may depend on whether the user asks for an apple, or for a Cézanne apple,” Ginsburg writes.

Perhaps more directly applicable to the labels’ case against Udio, Ginsburg states in regard to image-generating AIs and fair use factor four, the effect of the use on market value:

… even under a solely work-based interpretation of section 107(4), one may observe that the wholesale copying of an artist’s works into training data in order to enable stylistically similar outputs jeopardizes not only the artist’s future employment or commissions, but also devalues the actual works copied, because the image-generation program can produce outputs that compete with already-created works as well.

That same rationale would seem to apply to the labels’ evidence that Udio can output sounds which are substantially similar to famous and protected sound recordings. So, while the defendant is correct to say that copyright does not protect style and that music production relies substantially on mixing and matching a finite combination of styles, arrangements, etc., that premise, both statutory and judicial, is derived from a copyright history that has only ever included human artists in “competition” with one another. Consequently, the courts have latitude to find that it is in fact the AI developer who is seeking the novel conclusion that its machine furthers the purpose of copyright law.

As I say, there will be plenty of details to follow and plenty of considerations to nerd out on, if one is so inclined. And for better or worse, I am so inclined. Stay tuned (pun intended).

COPIED Act Introduced in the Senate with Focus on Content Provenance

copied act

On July 11, Senators Cantwell, Blackburn, and Heinrich introduced a bill called the Content Origin Protection and Integrity from Edited and Deepfake (COPIED) Media Act. One of many AI related bills in Congress, the heart of COPIED is transparency in artificial intelligence through implementation of content provenance information (CPI). COPIED requires development of industry standards to create “machine-readable information documenting the origin and history of a piece of digital content, such as an image, a video, audio, or text.”

The Commerce Committee press release announcing the bill states endorsement by News/Media Alliance, National Newspaper Association, Rebuild Local News, NAB, SAG-AFTRA, Nashville Songwriters, Recording Academy, RIAA, music publishers, artists, and performers. Senator Heinrich, who sits on the Senate AI Working Group, stated, “I’m proud to support Senator Cantwell’s COPIED Act that will provide the technical tools needed to help crack down on harmful and deceptive AI-generated content and better protect professional journalists and artists from having their content used by AI systems without their consent. Congress needs to step up and pass this legislation to protect the American people.”

In a nutshell, the bill calls for advanced, hard to remove, watermarks (or metadata) which would be permanently attached to digital content. In what sounds like a combination of copyright management information (CMI) and a chain of title concept, the development of CPI would enable tracing and validating the source of digital content with a variety of goals, including mitigation of deepfake or modified news stories and use of protected creative content without permission.

The COPIED Act would require the Under Secretary of Commerce for Standards and Technology to oversee the development and implementation of CPI in collaboration with the Register of Copyrights and the Director of the U.S. Patent and Trademark Office. If passed and effectively implemented, the law would prohibit removal, alteration, or tampering with attached CPI for deceptive or adversarial commercial practices; and one part of Section 6 of the bill begins, “It shall be unlawful for any person, for a commercial purpose, to knowingly use any covered content….” [emphasis added] This focus on use of material with attached CPI will be of greatest interest to creative professionals concerned about the myriad ways in which their work is used without permission for the development and commercialization of GAI.

Of course, there are miles to go before we see if and when this bill makes progress, at which point it may provoke some familiar arm flapping by the Electronic Frontier Foundation (EFF) recycling the same rhetoric it used to complain about digital rights management (DRM) technology. EFF lost its campaign to prove DRM under §1201 of the Copyright Act is unconstitutional, while this bill’s proposal for CPI is more reminiscent of §1202 under which it is unlawful to remove copyright management information (e.g., a watermark) for the purpose of copyright infringement. It strikes me that a similar approach would apply to removal of, or tampering with, content provenance information. After all, if it is designed to be as robust and tamper-proof as the bill projects, this would suggest its removal takes some effort and expertise, which itself implies a purpose that is likely to be unlawful.

Stay tuned. We shall see where this goes, but the aims of the COPIED Act strike me as a well-founded good start.

Photo source by:

Music Making Gen AI: A Deeper Dive into Fair Use

fair use

In February 2023, I argued that using copyrighted works for the purpose of training generative artificial intelligence (GAI) products is not fair use. My view in that post was, and remains, that because the purpose of copyright law is to promote authorship, and authorship is human as a matter of doctrine, then a purpose which replaces authorship is facially antithetical to copyright’s constitutional foundation. So, because a finding of fair use, should, as a matter of law, further copyright’s purpose, the GAI developer’s defense under that exception is invalid.

That said, I assume the courts will not rule on this threshold, constitutional question at summary judgment and will instead conduct fair use analyses in the first cases that proceed to at least bench trial. After one or two outcomes, if favorable to plaintiffs, we will likely see a lot of settlements because so many of the AI cases alleging mass infringement for the purpose of machine learning (ML) present the same legal questions.

Turning to the recent lawsuits filed by record labels UMG, et al. against GAI developers Udio and Suno, both complaints anticipate the fair use responses to come and, to an extent, imply the doctrinal view articulated above. Because the two complaints are nearly identical in substance, I’ll stick to Udio because the suit is filed in the Southern District of New York (SDNY), and the Second Circuit is where nearly all—if not all—of the relevant fair use case law has been active or decided. Odds are, the court in the First Circuit, which has a comparatively thin copyright record, will follow the Second Circuit’s lead in resolving Suno.

As discussed in my first post about these cases, the defendants seem unable to present a plausible claim of non-infringement and even signaled that they have planned to argue fair use in the lawsuits they knew were coming. All the action will be focused on factor one, part one (whether the use is “transformative”) and on factor four (potential harm to the market for the works used).

To dispense with factors two and three, the nature of the work(s) used and amount of the work(s) used respectively, these clearly tilt in favor of the plaintiffs. The sound recordings used for ML are highly expressive in nature (factor two); and by all evidence and inferences presented, it seems clear that defendant copied whole works—and many of them—into the AI model (factor three). One way the use of whole works could swing back to favor the defendant would be finding that the purpose of the use under factor one is, indeed, transformative.

To get there, I believe the court would have to find transformativeness under its precedent in Google Books, but in addition to the court itself describing that decision as the outer boundary of fair use, the Supreme Court in Warhol may have at least sharpened, if not narrowed, that boundary. As a factual matter, a GAI like Udio is nothing like Google Books. The latter feeds whole books into a system for the purpose of creating a research tool, while the former feeds whole sound recordings into a system for the purpose of producing other sound recordings—several of which have been presented in evidence as substantially similar to famous sound recordings.

Considering Udio Under Fair Use

As mentioned, the focus will be on factors one and four, which is not uncommon, but these cases highlight the interplay between the two factors. Factor one asks the purpose of the use, including whether that purpose is commercial; and factor four asks whether the use threatens the market value for the work(s) used. Thus, if a court finds under factor one that a use serves a “substitutional” purpose, this suggests that the use will unavoidably cause harm to the market value of the works used under factor four. This is what the labels argue, but a product like Udio does imply new territory for a fair use consideration.

Factor one asks two interrelated questions—whether the purpose of the use is transformative, and whether that purpose is commercial in nature. Commercial use tilts away from fair use but is not determinative, and transformativeness tilts toward fair use, but is also not determinative. In fact, the Supreme Court decision in Warhol, reversed a trend whereby transformativeness too often carried the entire fair use analysis. For instance, Udio’s failure to license the works used for ML is itself a potential market harm under factor four. Thus, even if Udio’s purpose were held to be transformative, its commercial purpose would split factor one, and the rest of the fair use factors would likely still favor the plaintiffs.

But, as the complaint states, “[Udio is] far from transformative, as there is no functional purpose for Udio’s AI model to ingest the Copyrighted Recordings other than to spit out new, competing music files,” states the labels’ complaint. This is properly framed in context to what “transformative” means in copyright law. Transformativeness is not about technological novelty or even innovation that promises to “make the world better” and so on. Notwithstanding the hyperbole in many such claims by various developers, the transformative question in fair use focuses on distinction of purpose from the works used.

As the quote above indicates, Udio ingests (i.e., copies) sound recordings for the purpose of making other sound recordings. And the purpose of both sets of sound recordings is, generally and presumably, listening pleasure for consumers. Even if none of the music produced by Udio were substantially similar to any of the music that went in, the labels contend that the overall purpose is holistically substitutional for all the recordings used to create the product. Udio used the music of human artists to “make” music without human artists, which is a purpose far beyond the Google Books boundary of providing a research tool to humans, including some who will be authors of works.

In response, Udio can argue that the purpose of its product is to produce a plethora of “new” music, which may indeed threaten to replace artists, but which is no more a substitute for the works used than a Beyonce song is a substitute for a Taylor Swift song. This is a tricky moment for copyright, which protects authors’ rights by protecting the use of their property. For instance, if no song ever comes out of Udio that sounds like a copy of an existing song, or if Udio can show that the majority of songs output are “new,” it could argue that its purpose is not substitutional under factor one.

Additionally, if Udio could show that its purpose is substantially providing a tool for would-be music creators, its claim to being “transformative” would be stronger under Google Books. But based on the reported function and market objectives of both Udio and Suno—i.e., mass-market products enabling any consumer to “make music” with a few basic prompts—the “tool” claim, if it were made, seems unpersuasive.

Let me interject that tech developers and copyright antagonists often conflate the economic concept of “creative destruction” with transformativeness, arguing that “copyright stifles progress.” While I personally question whether Udio et al. necessarily represent progress as a cultural matter, even if Joe Schumpeter himself would agree that technological replacement of human music makers is “creative destruction,” that prospect anticipates the nullification of copyright law as a relic of impliedly obsolete human authorship. As such, it would seem preposterous for a court to find that an affirmative defense to infringement should be applied in a manner that would cause copyright law itself to implode.

None of this is to say, as indeed the complaint makes clear, that peaceful coexistence between human authors and GAI cannot come to pass. Where GAI may be used by the human creator to make an expressive work of her own mental conception, the AI product has a much stronger claim to promoting the progress of authorship. But in the case of these music making products, that does not appear to be the intent—either by design or business model. And so, to reprise the doctrinal assertion I advocate, the Google Books opinion itself states:

Courts thus developed the doctrine, eventually named fair use, which permits unauthorized copying in some circumstances, so as to further “copyright’s very purpose, ‘[t]o promote the Progress of Science and useful Arts.’” [Emphasis added]

On that basis, the Second Circuit should find that a use of protected works which is holistically substitutional for human authorship does not further the purpose of copyright and is, therefore, barred from presenting a valid fair use defense.

Image source by: