While Meta may be confident in its legal strategy despite the new torrenting wrinkle, the social media company has seemingly complicated its case by allowing authors to expand the distribution theory that's key to winning a direct copyright infringement claim beyond just claiming that Meta's AI outputs unlawfully distributed their works.
As limited discovery on Meta's seeding now proceeds, Meta is not fighting the seeding aspect of the direct copyright infringement claim at this time, telling the court that it plans to "set... the record straight and debunk... this meritless allegation on summary judgment."
The allocation of the IP in what AI produces is pretty much zero-sum: one party’s loss is another’s gain. These are the conditions for a messy fight, which will be adjudicated in the first instance by judges and later, most likely, by lawmakers. There will be claims and counterclaims in the meantime; companies and individuals will try and secure authorship using contracts, terms of use, and the other usual tactics. But in the long run, when an economically significant question arises that is not foreseen by any current law, it tends to be resolved only by new law.
This law should avoid basing itself on a fundamental misconception that has dogged thinking about authorship and ownership in the modern era. Intellectual property law does not, in truth, exist to defend natural rights that individuals hold over their works. This isn’t how IP law came to be, when it was first developed as an extension of guild regulation in eighteenth-century England. Nor is it how it is legally justified in the United States, whose Constitution’s first article grants to Congress the power to make laws “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.” As the first subclause makes clear, the purpose of copyrights and patents is to incentivize authors and inventors by promising them a set of time-limited rights, akin to a temporary monopoly, over the dissemination of their works.
In other words, copyright and patent ownership are instrumental rather than intrinsic goods. The intrinsic good that IP law was established to serve is the widespread availability of inventions and works of art, and in the majority of cases it is clear—whatever the armies of lawyers employed by corporations with vast holdings of copyrights and patents might argue—that this intrinsic good is best served by those works being in the public domain, as free as the air for common use.
In fact, as a mass of historical scholarship has shown, the conceit that copyright law was ever a recognition of a natural right was generated and sustained above all by the nineteenth-century Romantic cult of authorship, itself a rearguard action in an epoch when writers threw off patronage for a new market system, becoming literary workers for the first time, directly dependent for their livelihood on the products of their labor. Straining to preserve their prestige in an era where the need to sell their works on a marketplace seemed to put it in doubt, writers invented the modern idea of author–ownership, at a stroke redefining themselves as property-owning bourgeois and throwing the sociopolitical stakes of intellectual property law into a mystified confusion from which it has never recovered.
If AI’s ability to generate works of art and to spark progress in science does nothing else, it detonates once and for all the Romantic myth of authorship as a special, organic, spiritual connection between “artist” and “work” that confers a privileged claim to authority over the way human creations circulate and are used. Roland Barthes may have been too early in proclaiming the death of the author in 1967. Now that LLMs can produce haikus and sonnets by the thousand—and may soon be able to do the same for novels and photographs and who knows what else—he may at last be vindicated. It would be worse than ironic to allow the inventors of the technology that has dealt a death blow to the era of Romantic authorship to use its very ideological apparatus—the identification of “authoring” with “owning”—to reap the spoils of what comes next.
...
There is another reason to be wary of retrospective redistribution as the answer to AI’s economic consequences. A century’s experience teaches us that intellectual property law has tended to operate like a ratchet. Since the emergence of international copyright with the Berne Convention of 1886, successive revisions to the global copyright regime have only ever moved in the direction of stronger protections for rightsholders: longer terms, stricter conditions on use, expansions of the amount of protected material. In the current era, which began with the ratification of the TRIPS agreement in the mid-1990s, international copyright and patent protection for “creators” and “inventors”—or more typically the corporations to whom creations and inventions accrue—is stronger than ever. It is precisely this historically unprecedented regime that powerful interests will call upon to divide up the economic spoils of AI.
Lemley posits further, "AI isn't competing with authors or artists. Instead, it is using their work in an entirely different manner. [...] ML systems generally copy works, not to get access to their creative expression (the part of the work the law protects), but to get access to the uncopyrightable parts of the work— the ideas, facts, and linguistic structure of the works." He proposes 'fair learning' as a principle that the use of copyrighted works to train ML systems should be fair even if fair use factors—the nature of the work, and the amount taken would otherwise weigh against fair use.
Take a language AI model trained on millions of books. It's not interested in the stories, characters, or themes; instead, it aims to learn linguistic patterns - things like grammar rules, sentence structures, and word relationships. Similarly, for an AI model to learn what a dog looks like, it needs to analyze millions of dog photos. The system isn't interested in the artistic composition or the specific dog in each photo - elements that might be protected by copyright. Instead, it's learning to recognize general features like fur, four legs, tails, and typical dog shapes. In fact, "verbatim copying" is the necessary intermediate step toward accessing the unprotectable "ideas and functional elements" of works that allow AI systems to learn generalizable patterns and concepts rather than simply memorizing specific content. AI models instead encode patterns from training data into parameters, generating responses using learned probabilities and not by referencing stored content.
...
As we navigate the complex landscape of AI and copyright law, a nuanced understanding is emerging. Legal scholars suggest two key points: the input data used for AI training may often be permitted under "fair use" or "fair learning." At the same time, purely machine-produced output is typically not copyrightable. This perspective recognizes that ML, at its core, is about extracting patterns and facts rather than copying creative expression. Ultimately, fair use is about more than transforming existing works. It's about preserving our collective ability to create, share, and build upon ideas. Or it's about preserving the ability to learn—whether the entity doing the learning is a human or a machine.
Six out of the high court’s nine justices have published books with the publishers involved in the case. Justices Sonia Sotomayor, Neil Gorsuch, Amy Coney Barrett and Ketanji Brown Jackson have all published books or signed book deals with Penguin Random House. HarperCollins has published books by justices Clarence Thomas and Gorsuch. And Justice Brett Kavanaugh is signed to a book deal with Hachette. (None of the publishers responded to requests for comment.)
The case involves a digital lending library operated by the nonprofit Internet Archive that it expanded during the early days of the COVID-19 pandemic. The publishers challenged the archive’s practice of copying and lending out digital copies of library books with no limit, through what the archive calls its National Emergency Library, as a violation of copyright that threatens authors’ earnings. A district court and the appeals court both ruled in favor of the publishers, finding that the archive’s digital lending practices violated copyright law.
The Internet Archive has not appealed to the Supreme Court yet. A spokesperson for the Internet Archive told HuffPost the nonprofit is still reviewing the appeals court decision. But if the case were to reach the high court, it would raise serious questions about the self-enforcement of conflict of interest rules by the individual justices at a time when the court has been embroiled in ethics controversies, particularly around Thomas’ receipt of gifts from friends and wealthy conservative benefactors.
I reached out to Jonathan Band, a copyright lawyer who represents the Association of Research Libraries whom we previously interviewed on the first ruling. As he explained, “Neither the district court decision, nor the affirmance, will have any impact on the controlled digital lending projects of research libraries, which focus on older, out of commerce titles which are not available from publishers in any format.” The Internet Archive, in contrast, offered popular titles that were also available on OverDrive, creating market overlap. But Band notes that “this decision does foreclose the possibility of using CDL as a means of countering the price-gouging public libraries experience with respect to ebooks lent by OverDrive.”
The final part of the ruling really is a tragic insult to the public. The Internet Archive made a final argument that the library should be allowed to continue given its substantial public benefit. But the Court, ridiculously, claims that the public gets a greater benefit from the library being shut down, and mocks the idea that expanding access to the public is all that important.
...
I imagine the Archive will seek cert at the Supreme Court, but who knows what will happen there. The Supreme Court hasn’t been great on fair use lately. The better answer is that Congress should fix this and make it clear that copyright law blesses this kind of open digital lending, though the copyright industries would throw a shitfit if anyone even dared propose such a bill.
This ruling is a huge loss for public access to knowledge and for libraries.
Pending any appeal, I also believe our response to publishers should remain consistent: This issue is not yet fully determined as it affects only a limited number of actors within a single jurisdiction. An appeal to the U.S. Supreme Court is still a possibility, and thus, this decision is not yet the “law of the land.” Nor is it global law.
As for legal next steps, the Archive has two pathways it may choose to take, according to Cara Gagliano, senior staff attorney for the Electronic Frontier Foundation, which is representing the Archive in court. They can either petition for a rehearing, or petition the U.S. Supreme Court to hear the case, an even more intense process, since the court only takes a handful of cases per year.
“We think the Internet Archive really was serving a transformative purpose and doing what libraries have always done: Loaning out books that they owned to one person at a time,” Gagliano said. “Our take is that it’s absurd that the Internet Archive is allowed to mail me a physical book it owns. The physical publishers can’t stop that. But [the Archive] can’t give me the same content in digital form.”
Legal experts are uncertain how much this latest court decision will affect colleges and universities, though they expect institutions to tread carefully. The programs at Carnegie Mellon, Michigan State, the UC system and other institutions—including the University of Florida and the California Institute of Technology—all appear to be operating, according to their respective websites. The institutions either could not be reached or did not respond to requests for comment.
“A lot of people in the academic space and the business space would rather operate as cautiously as possible,” said Stephen Wolfson, assistant general counsel and copyright adviser for University of Pennsylvania Libraries.
...
While the appeals process upheld most of the district court’s ruling, there was one deviation.
... [T]he appeals court found that the Internet Archive’s digital library wasn’t a commercial activity.
“To hold otherwise would greatly restrain the ability of nonprofits to seek donations while making fair use of copyrighted works,” the judges wrote.
Jonathan Band, a copyright lawyer who represents the Association of Research Libraries, said if the district court’s entire ruling had been upheld, the decision could’ve had potentially large ramifications for higher education libraries, many of which are nonprofits. (Note: This article has been updated to correct the library association Band represents.)
“If you start saying what they did was commercial, at that point anything engaged by any nonprofit would be found to be commercial,” Band said.
The American Library Association and the Association of College and Research Libraries both filed briefs stating the Internet Archive’s activity was “clearly not commercial,” though they did not take a further stance on either side of the lawsuit.
Penn’s Wolfson agreed with Band.
“If it had come out otherwise, it could look like practically everything we do is for commercial use,” Wolfson said.
Wolfson and Band did differ slightly on the impact of this latest ruling over all.
Band said the latest ruling—whether it was in favor of Internet Archive or not—wouldn’t have affected higher education libraries, given they work with research papers and scholarly monographs and not the popular titles that were targets of the Internet Archive.
“In this decision, we’re talking about trade books, the mass market books, like best sellers by Stephen King that are in print and available right now for commercial licensing,” Band said. To the contrary, many of the books seen in research libraries are typically not available, either digitally or physically, to the mass market. “These are just older, out-of-print books. They’re not available digitally through some easily accessible platform.”
Jennifer Urban, co-director of the Berkeley Center for Law and Technology, said university libraries’ lending programs differ from Internet Archive in that reader privacy is at the forefront.
In an amicus brief she wrote on behalf of the University of California Berkeley School of Law, along with the Center for Democracy and Technology and the Library Freedom Project, Urban pointed out that libraries minimize data collection and data transfer (transferring only a student’s library card number and book barcode, for example), as well as maintain data security.
“Library-led controlled digital lending incorporates longstanding library values and practices that protect reader privacy and intellectual freedom,” the briefing said. Urban added that commercial aggregators like Overdrive, along with the Internet Archive, “differ sharply from libraries in their incentives and practices regarding reader privacy.”
Wolfson expects higher education to feel minimal, if any, impact because of the small amount of digital lending programs just starting at institutions. But, as students increasingly demand access to online or digital materials, the ruling could stifle further program creation.
“This decision could be used down the road to challenge that sort of activity,” Wolfson said. “It creates an environment where previously you felt OK with lending some things through controlled lending programs—but not everything—but now there’s at least a couple decisions that show it’s problematic for this activity.”
Metro Campus Library: 918.595.7172 | Northeast Campus Library: 918.595.7501 | Southeast Campus Library: 918.595.7701 | West Campus Library: 918.595.8010
email: Library Website Help | MyTCC | © 2025 Tulsa Community College
