Thursday, April 16, 2026
Executive Summary
A Berkeley Technology Law Journal symposium panel examined how the Supreme Court’s unanimous 2025 decision in Cox Communications, Inc. v. Sony Music Entertainment overhauled a half-century of contributory copyright infringement doctrine, sharply narrowing secondary liability to two specific fact patterns—inducement and tailored material contribution—while leaving platforms, AI gatekeepers, and lower courts without coherent guidance on how the new framework interacts with the broader architecture of copyright enforcement.
Instructor(s)
Christopher Sprigman, Professor of Law and Co-Director of the Engelberg Center on Innovation Law and Policy, NYU School of Law
Oren Bracha, Professor of Law, University of Texas School of Law
Justin Hughes, Distinguished Professor of Law and Distinguished Chair in Business and Economics, Loyola Law School
Laura Heymann, Professor of Law, William and Mary Law School
Erik Stallman, Clinical Professor of Law, UC Berkeley School of Law
Keywords
Cox Communications, Inc. v. Sony Music Entertainment • contributory copyright infringement standard 2025 • secondary liability copyright platforms • substantial similarity infringement test • fair use affirmative defense copyright • derivative work right Andy Warhol Foundation v. Goldsmith • vicarious liability copyright 1976 Act • Section 512 DMCA secondary liability after Cox • AI platforms direct infringement copyright output • how does Cox Communications change contributory liability • what is the new test for contributory copyright infringement • Gershwin Publishing Corp. v. Columbia Artists Management
Legal Analysis
The Derivative Work Right and Substantial Similarity After Andy Warhol Foundation v. Goldsmith
The scope of copyright’s exclusive rights under the Copyright Act of 1976, Pub. L. No. 94-553, 90 Stat. 2541, remains contested terrain defined as much by judicial common law as by the statute’s text. Christopher Sprigman opened the symposium panel by situating this hybrid character as a feature, not a defect, of the Act’s design. Drawing on his role as a reporter for the American Law Institute’s forthcoming Restatement of Copyright, Sprigman observed that “Congress left ample room for judicial interpretation, even where … there is statutory text that must be understood and implemented,” and that the statute’s openness has kept copyright viable across successive waves of technological disruption. The derivative work right under 17 U.S.C. § 106(2) illustrates this dynamic most vividly. Sprigman argued that the Supreme Court’s decision in Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith, 598 U.S. 508 (2023), resolved a persistent tension between that right and the fair use doctrine by clarifying that § 106(2) is expressly “subject to fair use” and that “an over broad concept of transformative use … would narrow the copyright owner’s exclusive right to create derivative works.” Rather than weakening transformativeness as a fair use mainstay, Sprigman contended, Warhol “reaffirmed and strengthened” it by requiring that a transformative use exceed a “merely distinguishable variation” and rise to a level of significance sufficient to affect the purpose and character of the new work. The doctrinal architecture Sprigman described—fair use as a scope doctrine rather than a formal affirmative defense—found resonance in legislative history: the 1965 report of the Register of Copyrights specifically declined to place the burden of proof on either party, treating the question as one of law once the underlying facts are established. Sprigman predicted that a textualist Supreme Court, “unashamed about” reversing settled precedent, might one day revisit the affirmative-defense classification, and he counseled practitioners to “prepare for what might come next” rather than assume existing doctrine is permanent.
The Substantial Similarity Test’s Erosion and the Fair Use Overburden Problem
Oren Bracha offered the symposium’s most searching critique of the infringement test, arguing that four decades of case law have “consistently” hollowed out the improper-appropriation prong of the two-part infringement standard. The organizing principle that originally animated the test—substitution, meaning that a defendant’s work must function as an “expressive substitution” for the plaintiff’s work—has been gradually displaced, Bracha contended, leaving “a nebulous, frictionless idea of substantial similarity that indeed means very little.” He identified two subsidiary doctrinal distortions as particularly damaging. First, certain courts have conflated the de minimis exception with the substantial similarity inquiry, producing what Bracha characterized as “a race to the bottom” in which de minimis becomes a test of unrecognizability rather than a distinct threshold. Second, the Second Circuit’s alternative quality-quantity test—applied only when the standard test fails to produce an infringement finding—exemplifies, in Bracha’s view, the proliferation of ad hoc frameworks that undermine doctrinal coherence. The predictable consequence, he argued, is “fair use creep”: because the infringement test offers little friction, parties and courts have made fair use do the front-end scope work for which it was neither structurally designed nor conceptually suited. Bracha pointed to Andy Warhol Foundation litigation as emblematic: “the whole substantial similarity infringement analysis became a footnote” at the district court, received only brief treatment from the Second Circuit, and was largely absent from the Supreme Court’s analysis. The remedy Bracha proposed involves fixing all three dimensions simultaneously—restoring a substitution-anchored infringement test, cutting fair use “down to size,” and clarifying the boundary between the reproduction right under 17 U.S.C. § 106(1) and the derivative work right under § 106(2), whose internal frontier remains, in his characterization, “fused or unclear,” creating a capacious “super-right” of indeterminate outer boundary. Sprigman independently proposed an antitrust-inflected analogy: treating literal or near-literal copying as per se infringement while requiring plaintiffs to demonstrate a likelihood of substitution for more creative derivative-work appropriations, an approach he acknowledged would be “very divisive” but believes would better align doctrine with copyright’s underlying social-welfare rationale.
Cox Communications v. Sony Music and the New Contributory Infringement Framework for Platforms and AI Gatekeepers
The panel’s most consequential doctrinal territory concerned the Supreme Court’s unanimous decision in Cox Communications, Inc. v. Sony Music Entertainment, handed down weeks before the symposium, which reversed a Fourth Circuit judgment upholding a $1 billion jury verdict against Cox for continuing to provide internet service to subscribers known to be repeat infringers. Justin Hughes identified the decision’s most significant move in its second operative sentence: contributory liability can be shown only where the defendant induced the direct infringement, or where the defendant provided “material contribution tailored to that infringement”—meaning a contribution “not capable of substantial or commercially significant, non-infringing uses,” a test the majority derived from Justice Ginsburg’s concurrence in Metro-Goldwyn-Mayer Studios Inc. v. Grokster, Ltd., 545 U.S. 913 (2005), rather than from the Court’s Sony Corp. of America v. Universal City Studios, Inc., 464 U.S. 417 (1984). The Gershwin Publishing Corp. v. Columbia Artists Management, Inc., 443 F.2d 1159 (2d Cir. 1971) standard—which had governed contributory liability for over fifty years—was jettisoned without engagement with a half-century of lower court precedent. Sprigman, who filed an ACLU amicus brief urging the narrower nonfeasance approach adopted by Justice Sotomayor in concurrence, recounted that upon reading the majority opinion he “was like, no, not great, weird,” expressing concern that the Court “make[s] it way too simple.” Laura Heymann noted that Cox—written on “an easy” fact pattern—left unresolved whether the framework applies differently to hosting platforms versus connectivity providers, to say nothing of emerging AI architectures. Bracha pressed the point further, arguing that generative AI producers “occupy structurally a very similar position to other platforms in being gatekeepers,” and that the existing secondary liability framework, even before Cox, was “much too individualistic, specific, tied to specific knowledge” to capture the “systemic, broad social risk imposed by AI gatekeepers.” Heymann added that as AI output cases develop, courts will face granular questions about who—the model or the user—is responsible when a prompt elicits substantially similar output, and that Cox’s brevity “isn’t helping to … categorize those sorts of applications at all.” Hughes, meanwhile, highlighted a transatlantic divergence that predates Cox: the European Union’s e-commerce framework permits no-fault injunctions against third parties to suppress infringement without financial liability, a remedy unavailable under the current U.S. system, which he characterized as “a major divergence between the United States and other developed economies with sophisticated copyright systems.” He anticipated that Cox may prompt rights holders to test whether 17 U.S.C. § 502’s injunctive relief provision could be read more expansively than courts have thus far allowed.
Generated by AI based on the Interview/Transcript below.
Key Takeaways
- Cox Narrows Contributory Liability: The Supreme Court’s Cox Communications decision replaced the fifty-year-old Gershwin standard with a two-path framework requiring either inducement or untailored material contribution, a move Sprigman described as making secondary liability “way too simple.”
- Gershwin Is Now Dead Letter: The Gershwin Publishing formulation—“one who with knowledge of the infringing activity induces, causes or materially contributes”—was discarded without engagement with the lower court precedent it had generated, leaving practitioners without the analytical scaffolding on which copyright enforcement had rested since 1971.
- Substantial Similarity Needs Substitution Anchor: Bracha argued that the improper-appropriation prong has been emptied of meaning over four decades and that restoring a substitution-centered conception is prerequisite to doctrinal coherence: “we need to restore a coherent, meaningful infringement test that actually means something.”
- Fair Use Is Overburdened as a Scope Tool: Both Bracha and Hughes agreed that fair use “is bearing too much of the weight” of copyright scope analysis, a burden ill-suited to a doctrine whose structure “was designed to do” back-end safeguard work rather than front-end threshold filtering.
- AI Gatekeepers Require Rethought Doctrine: Bracha cautioned that existing secondary liability frameworks are too individualistic to address systemic AI risk, arguing that “we need to rethink the doctrine” toward a standard focused on “social risk” rather than specific knowledge of specific acts.
- S.–EU Injunctive Relief Gap Is Widening: Hughes identified a structural fault line: European systems permit no-fault injunctions against third-party infringement enablers, while U.S. doctrine, reinforced by Cox, routes enforcement through damages-only secondary liability theories that threaten existential liability or provide no remedy at all.
- Fair Use Burden of Proof Remains Unsettled: Sprigman traced the affirmative-defense classification to unreflective judicial repetition rather than statutory command, citing the 1965 Register of Copyrights report that expressly declined to assign the burden to either party; an audience member further noted that Sony’s language placed the burden on plaintiffs, a reading “lost in Harper and Roe.”
- Derivative Work Right and Fair Use Are Reconciled by Warhol: Sprigman read Andy Warhol Foundation v. Goldsmith as resolving—not deepening—the tension between § 106(2) and fair use, holding that the degree of transformation required for a fair use defense must exceed that required to qualify as a derivative work.
- Cross-IP Doctrinal Borrowing Demands Justification: Heymann cautioned that the Court’s repeated importation of patent concepts—“staple article of commerce,” the Aro Manufacturing framework—into copyright law has proceeded without explaining “whether that phrasing is tied to or particular to the IP doctrine that it is borrowing it from,” embedding path-dependent error into doctrine.
- Legislative Reform Is Difficult but Not Impossible: Sprigman rejected fatalism about comprehensive copyright revision, invoking historical precedents of democratic self-correction: “I think we are hopefully near the bottom of a period of decline, and we are on the precipice of a period of reform.”
B-CLE Recording (CLE: FREE) | Youtube Recording | Resource(s) | Speaker Bio(s) & Contact Info
Download the interview/transcript and slides here!
Interview/Transcript
This interview/transcript was based on a conversation on April 16, 2026 about 29th Annual BTLJ-BCLT Spring Symposium: Origins, Evolution, and Possible Futures of the 1976 Copyright Act, hosted by Berkeley Center for Law & Technology, UC Berkeley School of Law. The topic on “The Scope of Exclusive Rights and Modes of Enforcement” was presented by Christopher Sprigman, Professor of Law and Co-Director of the Engelberg Center on Innovation Law and Policy, NYU School of Law; Oren Bracha, Professor of Law, University of Texas School of Law; Justin Hughes, Distinguished Professor of Law and Distinguished Chair in Business and Economics, Loyola Law School; and Laura Heymann, Professor of Law, William and Mary Law School; and was moderated by Erik Stallman, Clinical Professor of Law, UC Berkeley School of Law.
Erik Stallman 00:00
My name is Eric Stallman. I’m a clinical professor of law here at Berkeley, and I’m very excited to be moderating this panel on the somewhat broad topic of the scope of exclusive rights and modes of enforcement under the 76 Copyright Act. Our panelists today will focus on discrete issues within that sort of broad scope, and ideally, we’ll explore the ways that those issues presented in one area have interconnection with issues presented in another. For example, how issues with the formulation of the derivative work right in turn influence our tests for infringement of that right, which in turn influence who is responsible and can be held liable for that infringement, which then in turn also affects how we think about the appropriate remedies for that infringement. Throughout, one question we’ll be asking ourselves is how much, and this is very appropriate in view of our last speaker, how much these issues are attributable to the text of the 76 Act itself, versus judicial interpretations that that may have made too much or too little of that text. And then hopefully, we’ll arrive at a general discussion with all of you at the end that also touches those questions. So each panelist will give a short introduction, and the other panelists as much as possible, will try to avoid initial interventions. Once we do all of that, we will, we will have a conversation amongst the panelists, and then when we’ll open it up, hopefully for a discussion with all of you, so I will briefly introduce all of the panelists in a quick word on what they’ll talk about. First, Christopher Sprigman, Professor of Law and at the NYU School of Law and co director of the Engelberg Center on Innovation Law and Policy. Chris is going to talk about the scope of exclusive rights, focusing a bit on the derivative work right and the related and important issue of limitations and exceptions that police the boundary of that right. Then Oren Bracha, professor of law and legal historian at the University of Texas School of Law, will discuss what we’ll charitably call evolutions in the test for infringement. Then Justin Hughes, Distinguished Professor of Law and Distinguished Chair in Business and Economics at Loyola Law School, will discuss secondary liability with a focus on the path to the language in Section 106 to do and to authorize, and how that language has fared in subsequent secondary liability cases. And then finally, Laura Heymann, Prof essor of Law at William and Mary Law School, will talk us through remedies and enforcement, with a focus on issues of doctrinal overlap between different intellectual property regimes and how that overlap might influence the way we do or should think about fashioning appropriate remedies. So with that, we’ll hand over to Chris.
Christopher Sprigman 02:59
All right, so I’ve been asked to talk about some of the elements of the statute and of the judicial development that both preceded it and grew out of it that helped to determine the scope of copyright. And I’m going to keep this brief, some of these doctrines may come up later in discussion and questions. I probably won’t focus as much as some others do on the period of the copyright acts, negotiation, drafting and enactment, although obviously events during that period are relevant to the discussion of scope. Instead, I’m going to talk about scope for a few minutes through the particular framework that I’ve been obliged to use for the past decade and more, as one of the reporters for the American Law Institute’s restatement of copyright, which has recently been completed and is about to be published in its final form. And I want to read you a key paragraph from the ALI’s own summary of the nature of a restatement. It captures what restatements are meant to do, as well as the direction of my brief remarks today. So the ALI writes in its so called style guide, which is really a lot about substance, a restatement assumes the perspective of a common law court, attentive to and respectful of precedent, but not bound by precedent that is inappropriate or inconsistent with the law as a whole. Faced with such precedent, an Institute reporter is not compelled to adhere to what Herbert Wechsler called quote, a preponderating balance of authority. Close quote, yay, Herbert Wechsler for using the word preponderating, but is instead expected to propose the better rule and provide the rationale for choosing it. So obviously I would add to this formulation the role of the statute, and where the statute, the Copyright Act, speaks clearly, as far as I’m concerned, it determines the question. But as many have pointed out, the statute, in many of its central provisions, is far from clear or far from fully prescriptive. Congress left ample room for judicial interpretation, even where, as in the case of 102B or fair use or the definition of a joint work, there is statutory text that must be understood and implemented. And then, of course, there are central doctrines in copyright, the rules governing secondary liability, for example, which we’ve just seen the Supreme Court opine on and perhaps significantly alter, we’ll see the definition of an author, even the standard by which infringement is judged. These are central doctrines. These are not the periphery of copyright law. These are the viscera, and these are doctrines on which Congress has had nothing or next to nothing in the statute. In all these instances, courts have built over the last half century an intricate architecture of common law doctrines around the copyright Act’s statutory skeleton, defining originality, structuring the infringement analysis, developing secondary liability theories, shaping and reshaping fair use to account for changing technologies of knowledge, production, of distribution and consumption. The result is today’s hybrid US, copyright law, part statute, part common law. In my view, it is copyright law’s hybrid nature enabled by the spaces that Congress purposefully left open that have kept the Copyright Act viable through waves of technological change over our lifetimes. So just one word on legislative history. What’s the role of legislative history in this hybrid system? On that question, I’ll come down somewhere between the positions Peter and Jessica outlined in the first panel. The restatement makes substantial use of legislative history, but to do a particular narrow job that is, the Restatement uses legislative history as a window on the meaning of statutory language, principally where courts have disagreed, and even then, we understand that the window can be clouded, and we approach the enterprise with caution. Now, treating legislative history with caution is not the same as rejecting it. Like Jessica, I’ve read warnings from judges about how unreliable and ultimately useless legislative history supposedly is. And then, to be perfectly frank, I’ve watched some of those same judges use legislative history when, perhaps not too surprisingly, they find it useful. So I’ll talk a bit about specific doctrines, and they how they’ve developed in this hybrid system. Which, again, is not some mistake or abdication of legislative responsibility, but a central element of Congress’s design. And I want to start with an important scope doctrine that exemplifies the approach Congress took in many parts of the statute, an approach that provides statutory direction but leaves ample room for and indeed invites judicial development, and that doctrine is Fair Use. Now, there are a couple of ways in which Fair Use is often misconstrued. First, it was not intended to be and in its best use by judges is not a quote for factor test unquote, it is a framework for analysis. As a line in the legislative history puts it, it is an equitable rule of reason. Congress has laid out, in Section 107, four elements that must be considered in every case, but it’s also made clear that those elements do not exhaust the analysis. By legislating in this way, Congress left open room for other considerations to enter, as the circumstances of particular cases suggest they might be useful. And I would add that even with respect to the four statutory factors, Congress has created significant interpretive space take just as an example, the first statutory factor the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes. So that text makes clear that the commercial nonprofit educational distinction is just one facet of a wider analysis of purpose and character. It’s difficult for me to imagine a formulation that more clearly invites judicial development. And of course, that’s exactly what we’ve seen. We’ve seen this, for example, in the introduction growth and subsequent refinement of the idea of transformativeness. Now I consider the development to have been a significant in advance, a significant advance in our law of Fair Use. Special solicitude for transformative works is, I believe, very helpful in aligning the copyright owner’s exclusive rights with the goals of copyright law. I’m aware that there are those who disagree, but I don’t want to fight that out today. My goal is simply to point out to those who say that transformativeness is not in the statute and so should not be used in the Fair Use analysis, that that argument is just a category error. It mistakes the nature of what Congress gave us in Section 107. So let me talk about another aspect of our current understanding of Fair Use, which is arguably also a mistake. For many years, courts have treated Fair Use as a defense, or specifically as an affirmative defense. I think that both of those formulations are too simple. The statutory text, specifically, the text stating that a Fair Use is, quote, not an infringement of copyright creates a substantial question regarding whether fair use is a defense in the formal sense, or rather whether it may best be understood as an element regulating copyright scope. The statutory language of section 107 could be understood although no court has so held to place the burden on the plaintiff of proving as part of its prima facie case that the defendant infringed, that the defendant’s use is, in fact, not fair. I actually don’t think that’s right either. And if you look at the copyright Act’s legislative history, it suggests very strongly that Congress did not mean to place the burden of proof for proving or disproving Fair Use on either party. In 1965 during a lengthy revision process that led ultimately to the 76 Act, the head of the US Copyright Office issued a report explaining the provisions of the most recent revision bill introduced on Congress in Congress in February 1965. In that report, the register described the debate over treating Fair Use as an affirmative defense, and concluded that the draft bill refused to adopt that position. And I’m going to quote from this report the author, publisher, interests, says the register of copyright have suggested that Fair Use should be treated as a defense with the statute placing the burden of proof on the user, the educational groups have urged just the opposite, that the statute should provide that any nonprofit use for educational purposes is presumed to be a Fair Use with the copyright owner have the having the burden of proving otherwise. We believe that it would be undesirable to adopt a special rule placing the burden of proof on one side or the other when the facts as to what use was made of the work have been presented, the issue as to whether it is a Fair Use is a question of law. Statutory presumptions or burden of proof provisions could work a radical change in the meaning and effect of the doctrine of Fair Use. The intention of section 107 is to give statutory affirmation to the present judicial doctrine, not to change it. So in the restatement, we explain that courts have held uniformly, although largely without analysis, that Fair Use is an affirmative defense. We acknowledge as any restatement must that this is the current state of the law, but we also note the difficulties that both the statutory text and the legislative history present for this view. And so now I’m going to set aside my role as reporter and reclaim my freedom as an academic to speak my mind. Reconsideration of the judicial framing of Fair Use as an affirmative defense is overdue. The Copyright Act does not codify Fair Use as a limitation. Fair Use is a scope doctrine that is Fair Use is one of the doctrines that determines the scope of the copyright owner’s rights in the first instance Like other scope doctrines, like the distinction between expression and ideas, it is ordinarily, for good practical reasons, the responsibility of the defendant to identify the scope issue as relevant in the first instance, but it is not necessarily the defendant’s burden to prove it, the parties are equally responsible for presenting the evidence they have to produce on the issue, and then the scope issue should be decided by the court as a question of law. Now, there are a lot of other scope issues that others on the panel will address, and still others we can discuss after everyone’s had their say. Let me just close with a few words about the relationship between Fair Use and the author’s right to prepare derivative works. For years, one of the attacks on the centrality of transformativeness in the Fair Use analysis has been the assertion that giving Fair Use solicitude for transformative works is inconsistent with the author’s 1062, right. That argument has always struck me as proving too much, and the Supreme Court’s recent opinion in Warhol puts it in a pretty tight box. Specifically, the Warhol court made clear that the copyright owner’s exclusive right to prepare derivative works is, quote, subject to Fair Use, unquote. The court further noted that while the two Fair Use and the Derivative Works Right are not mutually exclusive, quote, neither do they completely overlap, and I’ll continue to quote them in saying an over broad concept of transformative use, one that includes any further purpose or any different character, would narrow the copyright owners exclusive right to create derivative works. Okay, so in order to reconcile the two statutory provisions, the Warhol court held that the degree of transformation required to make transformative use of original for Fair Use analysis must go beyond that required to qualify as a derivative. Note that some transformative uses involve changes in purpose without changes in the material copied. An example would be the copying of complete, unauthored, unaltered books in the Authors Guild case to build a research database. Such a use can be transformative, the Court’s opinion suggests based on its distinct purpose, even though the use does not involve altering the content of the work. With regard to those transformations that do alter the content of the work, the Warhol court’s distinction suggests that to qualify as a transformative use that can weigh in favor of Fair Use, the defendant’s use of the plaintiff’s work must involve more than a merely distinguishable variation. It must rise to a level of significance sufficient to affect the purpose and character of the new work and to distinguish it from the purpose and character of the plaintiff’s work. That seems to me broadly correct and far from weakening transformativeness as a Fair Use Mainstay, the Warhol court seems to me to have reaffirmed and strengthened it by structuring the interaction between Fair Use and the Derivative Work Right.That preserves derivative works while not significantly limiting the scope of Fair Use. Okay, I’m going to stop there. Thank you for listening.
Oren Bracha 17:22
Good afternoon to you. So I will be speaking about the infringement test, and its interaction with a few other doctrines that determine the scope of copyright. I have a lot to say about that subject, and my time is limited, so I’ll be shooting rapidly, if you can forgive that. As you know, the infringement test, as well as some of the doctrines I’ll be talking about, were largely left by the 1976 Act for the case for judicial development. And therefore a lot of my comments will be about that. Let me warn you right at the outset, or maybe give you a spoiler, with apologies to everybody, and Justin in particular, this is not going to be a happy story. I’m sorry about that, and it’s going to unfold in three acts as follows. So first act is about the infringement test itself and its development, specifically the improper appropriation prong of the test, specifically the substantial similarity criterion within it. And you know, who amongst us has not, on occasion, and maybe more than occasion, succumb to the temptations of vilifying the substantial similarity test? You know, it’s subjective, uncertain, unpredictable, arbitrary, pick your favorite judicial quotation, right? Maybe Learned Hand from Nichols. Nobody has ever been able to fix the boundary, and nobody ever will. I paraphrased a little bit. Perhaps go even a century earlier to Justice Joseph Story saying this is the true metaphysics of the law, etc, etc, you know the drill. Well, here’s what I have to say about it. I think that actually what happened here is that for the last 40 years, the infringement test has been subjected to a consistent process of erosion and dilution, and it has been vacated, unfortunately, within the developing case law from a lot of its substantive content, and most importantly, it has been vacated from its central conception, the conception that basically supplied coherence and meaning with the within this field, which actually existed beforehand, and that conception was the idea of substitution, the idea that in order to infringe a defendant has to work as a substitution, as an expressive substitution for the plaintiff’s work. And of course, that doesn’t solve all problems. Of course, there are hard cases, but it’s an organizing concepts that give meaning to the test. What happened throughout the decades is that code gradually again hollowed out the test, and the idea of substantial similarity from that notion of substitution, leaving us with a sort of a nebulous frictionless idea of substantial similarity, that indeed means very little. You know when is it substantially similar? You know when it’s substantially similar, right? That’s not very helpful. That’s correct. And once that happened, then the test became uncertain, unpredictable, maybe even arbitrary, et cetera, but it is us right, meaning the case law, and of course, I’m over generalizing that has done that to a large extent, and therein lies the problem, I think. Now, once that happened, there were a few other unfortunate developments within this area, I’m only going to quickly mention two of them. One of them is, I think, the very confused and unfortunately more and more widespread tendency of some courts to describe the infringement test as an exception for de minimis uses, it’s a de minimis uses exception to infringement, an exception for particularly trivial cases. It’s really a confusion of relevant concepts, and once that happened, other other courts naturally, very quickly, were fast to understand this de minimis characterization of the test as a criterion of unrecognizability. When don’t we have substantial similarity? Well, when it’s de minimis. When is it de minimis? Naturally, only when one cannot even recognize the plaintiff’s work within the defendant work. That’s a sort of a classic race to the bottom, and it’s really deeply confused, at least the conflation of a de minimis exception with the substantial similarity improper appropriation test. So that’s one support in development. The other one is the proliferation I think. Although I will only mention one key example, the proliferation of alternative infringement tests that basically in specific cases, specific patterns, specific situations, create alternative forms of analysis. The worst kind of example, in my opinion, anyway, is the Second Circuit alternative quality quantity test that applies when the regular test is in Act read when it does not produce the outcome of infringement. So that is not very helpful either. Act Two, after we hollow out the meaning of the infringement test, there’s a vacuum, and what steps into the vacuum? Of course, what steps into the vacuum is the Fair Use doctrine. Okay, so it’s a sort of act two is about the Fair Use creep, you might say. And the idea here is simple. If the infringement test indeed means very little, it’s very hard to apply. It’s subjective, et cetera, you don’t need to worry about it. Why? Because we can always Fair Use it. Don’t worry. We’ll Fair Use it. We don’t even have to do the infringement test. We’ll take care of it at the back end that only too often becomes the front end. A classic example of that. There are many, but a classic example of that is the litigation of The Andy Warhol Foundation case where it turns out that the whole substantial similarity infringement analysis became a footnote, basically ignored at the district court level. Then you get a short passage by the Second Circuit because they couldn’t avoid it. And then again, abundant at the Supreme Court level. That’s all there was to it. It’s a classic example of, don’t worry, or maybe you should worry, we’ll Fair Use it. And although I’m not an enemy of Fair Use, and indeed setting aside the debate whether it should be an affirmative defense or not, I think this is an unhappy and indeed abnormal situation. We basically, we’ve been putting too much burden on, in my opinion, the two narrow shoulders of the Fair USe doctrine. It’s an important doctrine, but I think it’s a doctrine that whose structure and design and conceptual content is about being a back end safeguard of last resort, basically again, without necessarily arguing that it should be an affirmative defense. And what has happened is that we’ve shifted a lot of the burden of copyright scope analysis to Fair Use. Because remember, don’t worry, we can Fair Use it. And because I think there’s a basic mismatch there between the concept and the structure and what it was designed to do and the burden that is being put on its shoulders. I think it was inevitable that at the end, it’s not going to work very well. And I think it has become more and more evident, certainly since the Andy Warhol backlash period. So that’s very useful. Finally, third act is about the shall we say, complicated relationship between reproduction, derivative works, entitlement and scope in general. So again, as most of you probably know, we actually have two very deep, problematic conceptual issues in that relationship. The first one is probably nobody knows where the borderline is between derivative works and reproductions. Indeed, many have argued that there is no such border line. Indeed, courts daily ignore that borderline because they keep analyzing cases without telling us whether they’re doing reproduction or Derivative Works analysis. Second at the outer boundary, since the derivative works entitlement is often understood in very capacious, free, willing concepts, as you know, any secondary valuable use of the work, the outer boundary is also unclear, nebulous, floating out there. What do we get as the bottom line of those two moves? We get that huge Bob, sorry for the expression, right? Or, if you prefer a super right, a reproduction, derivative work, super right, the internal boundary is fused or unclear, and externally, nobody knows the boundary, so it keeps potentially expanding. That’s not a good situation for scope. So to sum up, what should we do about it in an ideal world? You know, if it’s broke, fix it along the three dimensions. We do need to restore a coherent, meaningful infringement test that actually means something. My idea would be along the lines of the substitutes conception. Once we’ve done that, and only once we’ve done that. I actually think we should cut Fair Use down to size to do its very useful work, but the very useful work it’s capable and designed to do. And finally, we should fix the derivative reproduction situation. How can we do that? Well, the derivative work right should be what it has always should have been, which is a right of making adaptations, which is something very similar then a free willing no boundary. Thank you.
Justin Hughes 29:09
All right well, thank you Eric, and thanks to the Berkeley team for including me in today’s wonderful gathering. The last time I was in this magnificent room was for a conference Pam and colleagues organized on a new European law. So I like the notion of being in this room now to discuss an old American law. But looking out in the room, I realized if Jessica Silbey were here, I would hear something from her about Hogwarts. I’m assigned to discuss secondary liability. I’m not sure how that happened, which after Cox Communications versus Sony Music, I will summarize as what the hell. I found the emoji appropriate for this. I looked for the shocked emoji most reminding me of Edvard Munch. So, but to talk a little bit more intelligently about secondary liability, I will have to talk as fast as Bob Brauneis and Peter DiCola did this morning, which I will try to do. Remember, copyright remedies come in two main branches, damages and injunctions, and that’s important because for assumptions inherited from the 1909 act and from the Federal Rules of Civil Procedure, the assumption has been in copyright that when it comes to many and many, many, very important, potentially secondarily liable parties, only damages will be available. And that I mentioned because that is actually producing and has been producing for a number of years, a major divergence between the United States and other developed economies with sophisticated copyright systems. So what was the lay of the land on secondary liability at the time of the 1976 act? With exposure to damages in mind two distinct branches of secondary liability had developed, and we’re going to see if I can work this clicker, I can watch that yes, vicarious liability and secondary liability or contributory liability. Now you all know vicarious liability is rooted in respondeat superior, but it was the second circuit’s 1963 Shapiro, Bernstein versus H.L. Green Company case that was and is now widely regarded as the landmark case, crystallizing the standards for vicarious liability outside the employer employee context. And you can read the language of the court and the test that is repeatedly quoted, at least until our Supreme Court gets its hands on it. While the 1976 Act was silent on vicarious liability, the legislative history shows that the drafters were well aware of vicarious liability, and indeed, while the Shapiro court had said direct, right it made it the direct financial benefit, you’ll see that in the legislative history of the 76 Act, the House report talks about pushing back some amendments that would have limited vicarious liability, and they talk about vicarious liability being available for indirect financial benefit as well as direct financial benefit. On the other side, the doctrine of contributory infringement we all know was or had its origins in the Supreme Court’s 1911 decision in Kalem Company versus Harper Brothers. In Harper Brothers, the exclusive right to dramatize the novel Ben-Hur was violated by the showing of an unauthorized motion picture reenacting scenes from the novel, and the issue before the court was whether Kalem, who had prepared or at least distributed the motion picture, was also liable for the exhibitors, public performance. So in that circumstance, evoking quote principles recognized in every part of the law, Justice Holmes imposed liability, concluding that if the defendant did not contribute to the infringement, it is impossible to do so except by taking part in the final act. What was interesting about that 1911 case is Justice Holmes did not use the phrase contributory liability or contributory infringer. Although, Kalem was quickly recognized as embodying that doctrine, and indeed it was initially recognized as embodied in that doctrine in a Sixth Circuit decision on patent law. Over time, the fact patterns that supported a finding of contributory liability expanded and In the midst of the final run toward the 1976 Act, the second circuit’s 1971 decision in Gershwin publishing versus Columbia Artist Management gave us what was, until few weeks ago, arguably the classic formulation of contributory liability. One who with knowledge of the infringing activity induces, causes or materially contributes to the infringing conduct of another may be held liable as a contributory infringer. As almost everyone in the room knows, all this was up ended a few weeks ago when the Supreme Court handed down their Cox Communications versus Sony Music decision. In that decision, the Court unanimously reversed the Fourth Circuit’s upholding a jury award of $1 billion for contributory liability that had been based on Cox continuing to provide internet service to customers who it knew to be repeat infringers. The Cox Communications majority opinion, for those who have not read it is notable for its extreme brevity and for its lack of citations to lower court contributory liability decisions to which the court had previously given meaningful consideration, eschewing lower Court precedent as apparently not worthy of discussion and relying. On only its own three Supreme Court decisions on copyright law and two on patent law, Justice Thomas provided us with a new, more limited framework for contributory liability in copyright law. Now I’d give you these two sentences, because I think these are the critical sentences. In the first sentence, the provider of a service is contributorily liable for the user’s infringement only if it intended that the provided service be used for infringement. That first sentence is unremarkable if you understood 110 years of contributory case law as using adequately specific knowledge of infringement as a placeholder, a kind of sufficient evidence of bad intent. It’s the second sentence that sharply curtails the range of contributory liability fact patterns can only be shown by two specific fact patterns where the defendant induced the direct infringement, or where the defendant provided material contribution tailored to that infringement. And the majority identifies the first pattern inducement with the court’s Grokster decision as to the second fact pattern, a material contribution is tailored to the infringement the majority reasoned, if it is not capable of substantial or commercially significant, non infringing uses a test we all know it draws from Sony. But, to make things more complicated, Justice Thomas draws it from Justice Ginsburg’s concurrence in Grokster, which would mean, perhaps it means only what Justice Ginsburg thought the phrases meant in Grokster, and maybe not what the court thought it meant in Sony. The brief majority opinion is also remarkable in its utter imprecision as to what kinds of knowledge the defendant could have for any kind of liability, and may be even more remarkable for its willingness to abdicate the court’s role in the development of common law. Is this what Congress intended? Okay? Remember that under the 1976 Act, the exclusive right given to the copyright owners are to do and to authorize each of the six enumerated activities. Now that to authorize language was intentionally added. It was copied from other common law copyright systems. In particular, I think, the UK Copyright Act of 1911 the Imperial Copyright Act, which provides, under section one, one of the Imperial Copyright Act, a set of rights the copyright owner has, and then at the end and to authorize any such act as a foreset. Of course, they misspell Authorize. It’s properly spelled with a Z authorize, but we got that from their system. Okay, now, what does that to authorize language mean? Well, it was intended by both according to the House report and the Senate report, It’s intended to avoid any questions as to the liability of contributory infringers. For example, a person who lawfully acquires an authorized copy of a motion picture would be an infringer if he or she engages in the business of renting it to others for the purposes of unauthorized public performances. Now that’s interesting, because the example given in the legislative history would under some fact patterns be preserved by Justice Thomas’s Cox communication reasoning. However, even an authorized copy of a film does have substantial non infringing uses. So according to the legislative history, there was an example of a contributory liability situation that would have involved distribution of something that had potentially authorized uses. Anyway, but remember, that begins with just for example, raising the legitimate question whether to authorize was intended to include the entirety of the existing contributory case law at the time. Now I’ll just say that I think Jessica was correct this morning in noting that statutory damages for single infringement when applied to these intermediary platforms and services is not a palatable remedy, and that’s a big problem for our system okay, that we’ve only been able to address these secondarily liable parties, many of them through a system that threatens them with damages only. I think that comes from a big mistake in the 1976 Act, or a big mistake from our understanding of the 1976 Act, and for that, I need a lot more time to walk you through what the 1909 Act said about injunctions, and what section 502 of the 1976 Act says about injunctions. But clearly we’ve drawn a separation between us and the European Union and other systems of complicated, sophisticated copyright who do permit these injunctions against third parties without holding them financially responsible for damages. So the last thing I think on the list Eric assigned me, which leads to Laura, is to say I’m fascinated by the Supreme Court’s obsession of making patent and copyright law into kissing cousins, and I understand it’s rooted in the copyright and patent clause. But what I don’t understand is that is, I don’t think we’re we’re going to, at some point the problem is going to arise of drawing the connection between contributory liability in trademark law, as set out in Inwood, and contributory liability as it’s now been defined by the Supreme Court in Cox Communications. But I think that’s the world of Laura’s topic, and I’m happy to turn it over to her for that.
Laura Heymann 41:11
Thanks Justin, and thanks to all the organizers for putting together this great conference, and thanks also to Eric for getting us very well organized here. So As Justin mentioned, I want to talk about the question of scope through two somewhat related topics, most of which I’m going to talk about is the idea of doctrinal overlap, and then I’ll talk about remedies a little bit at the end. And the doctrinal overlap question itself actually has sort of two components to it. So one is whether a party can bring claims under more than one IP doctrine on one set of facts, relatedly, whether and when it can use one IP doctrine when another is unavailable. And then secondly, how courts, and the Supreme Court in particular, as Justin was just talking about, borrow from one doctrine to set out rules in another. The unifying question, I think, is whether intellectual property law as a whole should have some consistency among the doctrines. I know we’d settle for consistency in at least one, or should at very least courts and perhaps Congress, if there are revisions, be encouraged to put forward an intentional and conscious acknowledgement of inconsistency. All right, so let’s talk first about doctrinal overlap. There’s some interesting questions around that. I think that the court’s attempts at what some have called channeling in the copyright context. I’m thinking about Dastar and Star Athletica here have not been, shall we say, especially nuanced in Dastar, the court concluded that the right to use a work in the public domain included the right to use it without attribution, whatever the sort of policy rationale or justifications for that might have been that has struck me as an over broad reading of the cases on which it relied, such as Sears. Komko in the shredder we case. Star Atletica still remains somewhat impenetrable. The court’s language in describing the test doesn’t provide clear guidance on how it conceptualizes the dividing line between the expression potentially eligible for copyright protection and the useful article potentially eligible for patent protection. Now I think, as some folks have acknowledged earlier, both of those opinions are partly the result of their facts. They involve sort of odd fact patterns to decide these larger issues. And the court, perhaps, in some ways, could be seen as saying, Look, we’re just going to resolve the question we have before us, and we are not going to worry too much about how this rule gets applied in other cases, CEG Cox versus Sony. Conversely, Cox versus Sony was an easy or fact pattern that allowed the court to do the same thing. The court does seem though to be clear about at least its intention or its belief that it is important to have a clear division between these doctrines. And yet, there is nothing that prevents a rights holder from claiming rights under both copyright and trademark, for example, such as for a logo. And indeed, the court in Walmart versus Samara Brothers, specifically invited rights holders in those situations to use copyright or design patent protection while building up the acquired meaning that it would need for trade dress rights. And there are multiple other lower court opinions that also acknowledge that overlap as well. And of course, there is nothing. If we now move from the idea of a single party asserting both rights just to the idea of overlap in general, there’s of course, nothing that prevents a company from creating trademark rights around a work that in the public domain or that eventually ends up in the public domain. United’s continued use of Gershwin’s Rhapsody in Blue, which it once paid $300,000 for the rights for, but has entered the public domain in 2020 the Harlem Globetrotters, to take another example, have maintained a registered trademark to the music to Sweet Georgia Brown since 1992, that song went into the public domain in 2021. Those songs, that music, at least a portion of it that’s being used are source identifying and so a competing airline or a competing trick basketball team likely could not use them as their own marks without facing some sort of legal action. But of course, nothing prevents anyone from giving a concert featuring either of those songs. So if the question comes down to whether or not there should be election, I don’t personally favor a system that would require election, as I’ve previously written. My preference is for courts to do that work on the back end, either through application of a misuse doctrine, if there’s a strong sense that the rights holder is trying to do an end run around the non existence of a right or or through more carefully calibrated remedies that are tailored to the harm that’s articulated in each case. The tide does seem to be turning a little bit on the question of disclaimers, which are think coming back a little bit more into favor, particularly if the alternative is an outright injunction against use, and maybe that we should see some additional work on that. Finally, on that front, there might be one additional way that we might encourage Congress if we’re thinking about revisions to Copyright Act, to address this directly is not to talk about sort of overlapping doctrines as such, but perhaps we need a more intentional, positive description of the public domain that would tell us how then that concept interacts with the other IP doctrines. But there is another way in which we can think of this sort of doctrinal overlap, and this is what Justin concluded with, which is the court’s own reasoning in copyright cases when it borrows from patent law, contributory infringement is probably the most prominent example here. The court hasn’t really made clear whether it is looking at the purposes behind the doctrine that it’s borrowing from thinking about these doctrines holistically. We call the entire field intellectual property. Some of us teach survey courses that attempt to provide some unifying themes, but it’s not at all clear that either court, the court, or, of course, Congress, is keeping those things in mind when it offers either decisions in the former or legislation in the latter. And so I don’t think the court has, in my view, given sufficient attention to whether, say, tributary infringement standards as Justin just talked about, should be the same across all the doctrines, or should it use they use the same language to describe those doctrines. The court has been borrowing the phrasing in those instances, but without explaining whether that phrasing is tied to or particular to the IP doctrine that it is borrowing it from. So when the court uses the phrase staple article of commerce, should we really see that is that, at its root really more of a rhetorical move than a doctrinal move, it’s a phrase that sort of resonates in terms of what the court is trying to describe, rather than a sort of core doctrinal move, by which I mean that the concept that’s sort of being discussed in the court cites, for example, Aro Manufacturing versus Convertible Top Company, when it cites that, when it uses that phrase, in Sony, what’s happening in Sony is not quite the same as what would have been happening in Aro, right? Which is the idea of, sort of supplying fabric to an end user that might or might not be used in the creation of something that would infringe a patent, primarily because in Sony, what made up the substantial non infringing use was a judicial determination of Fair Use, which Sony obviously could not have had certainty about prior to the court’s decision, and the court and Sony took pains, in a footnote, to distinguish Inwood the contributory infringement standard there saying, we have consistently rejected the proposition that a similar kinship exists between copyright law and trademark law, and in the process of doing so, have recognized the basic similarities between copyright and patent. Justice Blackmun, in his dissent, took issue with that right. I’m saying I do not agree that the technical judge made doctrine of patent law based in part on considerations irrelevant to the field of copyright should be imported wholesale into copyright law despite their common constitutional source, patent and copyright protections have not developed in a parallel fashion. But it’s not clear if, in fact, they these doctrines across the IP regime should be substantially different, or conversely, whether the fact that the Constitution’s intellectual property clause encompasses copyright and patent and so at least at root, a similar structure of incentives and limited rights is alone justification enough to engage in doctrinal borrowing. But the problem is that we have path dependence and precedent, and so once the court has borrowed that phrase in Sony, it then becomes entrenched, which allows the court and Cox not to re-examine that question of the borrowing right, the fact that the court once said it is good enough. Then there’s a question of whether we can have sort of better kind of coherence and thinking about the regimes, even within a doctrine right and so remedies is an example of that. Remedies, of course, are also a scope question. Right holders are going to be incentivized to push the boundaries of scope in the areas that are likely to get them the most remedies, as Jessica suggested in her remarks this morning. And here, I think I’ll just point to to Pam’s work with Tara Wheatland on statutory damages and her timely and much needed co-authored work with Erik and with Jennifer Urban on section 1202 particularly in the notion that the possible damages under that section, if that section is interpreted broadly and wrongly, could far exceed damages under what’s contemplated in the copyright scheme as a whole. Post Cox, we’re not sure what work exactly section 512 is doing anymore at a statutory level, as others have mentioned, the definition of copy In the Copyright Act and its embedded reference to fixation, which has its further requirement about the work be fixed by or under the authority of the author, works to talk about rights, but doesn’t work as well when we think about the use of the word copy in 106, ones reproduction rights. So sort of similar to the definition of use in trademark law. So I’ll just conclude by saying I think that the court’s opinion in Cox, since that’s going to be I think a topic for discussion for us on this panel, leaves a lot of questions, but I think one thing it demonstrates is that not every member of the court seems to be committed to the process of legal explanation and justification. So the patchwork nature of the more recent amendments to the Copyright Act, reacting to perceived threats from a digital environment suggests that if it’s not time for a holistic review of the Act, that at least there should be more attention going forward at each revision stage, more attention paid to how those revisions fit into and relate to the existing statutory scheme as a whole, and then other IP doctrines as a whole. So I’ll close there and look forward to our discussion.
Erik Stallman 53:29
Great. So I wasn’t going to start here, but having Judge McKeown here, and having Judge McKeown talk about amicus briefs sort of brought something to mind, which was the case of Dr Seuss Enterprises versus ComicMix. There were a couple of amicus briefs filed in that case. One was filed by Peter Menell and some of his colleagues, and the Court cited that brief. There was another excellent brief filed in that case, amicus brief that I worked on with Pam Samuelson and Rebecca Tushnet, which the court completely ignored. And one part of that brief argued, I think, very persuasively, that the Fair Use, that was wrong to conceive of Fair Use as an affirmative defense on which the defendant had had the burden of proof. And we laid out a lot of reasons why that was was wrong, and Judge McKeown and her decision dispensed with all his arguments in a single line that I have tried to forget, but it goes something like this. Not, not much is clear about Fair Use, but one of the very clear few, one of the very few clear things, is that it is an affirmative defense on which the defendant bears the burden of proof. So the question I’d like to pose to the panel after hearing Chris and Oren particularly is, is whether one whether the Andy Warhol foundation really requires us to revisit the sort of the clarity of that statement, but also, if it matters, Oren you said you were sort of agnostic on how to conceive of Fair Use, but if we were to fix the doctrinal problems related to the derivative work, right, and the test for infringement, does that put, so you say less pressure on fair use, and then does it matter less where that goes?
Christopher Sprigman 55:16
Okay, let me just say something about you know, what is the law, right? And then a little bit about lawyers and how they proceed in the face of what the law is because I think those things are distinguishable, and I think the second is very important. So what is the law? I think it’s very clear that the law is fair use as an affirmative defense. I have no quarrel with that. Every court that has addressed it basically has said that. Now I will emphasize that analysis is lacking. And you know, nonetheless, it’s pretty firmly I mean, not pretty. I’ll remove the word it’s firmly established. So what do you do? Okay, so a lot of things are firmly established in the law until they aren’t. So Gershwin was firmly established as the basis for contributory liability until it wasn’t. And, you know, I think the that is actually good. I actually think I don’t agree that It’s good in Cox. I wrote a, colleagues, and I wrote a brief for the ACLU where we argued for the approach that Justice Sotomayor took in her concurrence, and I thought that that was pushing the outer envelope of what was possible. And when the opinion came down, I was explaining to some colleagues earlier, I read reversed, and I thought, oh, great, and then I started reading, and within five minutes, I was like, no, not great, weird. Okay, so lawyers, in the face of that, you know reality, I think we have to prepare like the world for what might come next. And what might come next one day is that a court like our current Supreme Court, which is very textualist, not particularly bound by or respectful of precedent, to be perfectly honest, might look at the Copyright Act and see the language that I talked about. Might look at the legislative history, although they’re, you know, equivocal about legislative history, there’s a range of views on the court about it, but might look at those things and say actually, no, that’s not how it works, and they’re unashamed about doing that. And so I think, as lawyers, you know, in the face of that reality, it’s slightly, I don’t know, it’s wish casting, or maybe even irresponsible to insist that the law is what it is. We can simultaneously recognize what the law is and then prepare for tomorrow.
Oren Bracha 57:28
I’ll just say very quickly, I’m not agnostic at all. I don’t think it doesn’t matter. In fact, let me come clean. I think it would be a good thing if Fair Use is not true as an affirmative defense, and indeed, there’s room for it, you know. Fair Use has traveled long distances along 200 years. People in the room wrote about it. Matt Sag, Tony Reese, yours truly, right. It started its career as part of the infringement. So it could be done, and it would be an improvement. I still think that even if we had such an improvement, which is hard to imagine at the moment, although perhaps we should imagine, it still wouldn’t be a good fit to have fair use as basically the central plank of scope analysis, a doctrine which is based on four factors which I get that there’s precedent and there are patterns, and I read all the smart stuff that people in this room wrote about that still a doctrine that that’s its structure should not be the main plank that regulates scope. Right? It has a role to do, and we do it better if it’s not an affirmative defense, but I think it should be a more secondary role.
Laura Heymann 58:41
Not to get involved in that debate, although I think I tend toward more the agnostic, but it occurred to me just now, as you were talking that sort of an interesting parallel with, and again post Cox we don’t know how relevant 512 is anymore, but what happened after with Lenz in 512 F, right, where the court sort of shifts the burden back to the copyright owner to say that the notice that they’re serving, they’ve taken Fair Use into account, right? So, you know, that is not unprecedented, right to shift it. I just, I think I’m aligned with the idea that I’m not sure it’s going to matter that much as it cashes out.
Justin Hughes 59:24
Well, I agree with Oren that it’s pace is bearing too much of the weight, and more should be going to substantial similarity, but as long as an affirmative defense can be subject to a motion to dismiss or motion for summary judgment, I’m a little more agnostic.
Erik Stallman 59:37
Moving to question of secondary liability. One thing that was interesting in the in the panel on origins this morning, was some of the conversation around what you know, how, how Congress was thinking about computers and intermediaries in 1976, where a lot of the intermediaries were part of the content industry where, like publishers, are intermediaries. But also the way they were thinking about computers was perhaps more as sort of storage and retrieval systems. I mean things that were barely they they were not directing much of of the of the action. And we now live in a world of platforms, and the computers we’re interacting with are doing lot, a lot more than just allowing you to store and retrieve information. And Jessica Litman was sort of pointing out that one reason why maybe decisions like like Sony and maybe decisions like Cox came out the way they did, is not wanting to, sort of to to, you know, once we realize that, you know, sort of personal copies can nonetheless have existential consequences, what do you do about these intermediaries but and holding liable and not liable for all of those personal copies? Seems like it could be, you know, as much as an exceptional existential problem for the rights holders, one that can be existential for the platform. So in view of that, do we just need to think about secondary liability different when we think about platforms?
Justin Hughes 1:01:13
Yes. We need to think about secondary responsibility, not secondary liability. And that’s where the European Union got way ahead of us in their equivalent of the DMCA, the E Commerce Directive that they permit no fault injunctions against third parties in order to suppress infringement without any financial liability. And arguably, there is a significant loosening of injunctive power between the 1909 Act and the 1976 Act just if you look at the statutory language, but whether or not it’s loose enough to overcome federal rule of civil procedure 65 is not clear, but it’s untested. So perhaps, perhaps one of the things that will come out of Cox Communications is some of our content owners will test some of these issues that aren’t tested.
Oren Bracha 1:02:07
Let me, let me say yes, and then raise actually in the following, by using, sorry for all the poker metaphors, by using the AI code, right? So let’s put that on the table, and if we ever get beyond that training debate, we will still be faced with a question of infringing or similar output. The system actually produces similar output to copyrighted works, and many people actually think that’s an easy question. Of course, that’s infringement, and indeed, infringement by the AI producer. Personally, I don’t think it’s an easy case, and this is exactly, exactly the connection. I think it’s harder to see, but I think AI producers actually occupy structurally, a very similar position to other platforms in being gatekeepers. And once you understand that, it becomes a much more complicated situation. I’m not sure that sweeping direct liability is a good idea, because they function as gatekeepers, and that means they don’t have the same interest as their users, or the same interest as the public interest, whatever that is, and there is the concern of too much liability leading to chilling effects. On the other hand, I think that existing secondary liability doctrines before the Supreme Court gave us the cherry on top in the form of Cox, even before that is much too individualistic, specific tied to specific knowledge, specific ability to supervise specific activities to capture the systemic, broad social risk imposed by AI gatekeepers. We need to rethink the doctrine in that direction. It would be a doctrine about social risk. We need something new there well,
Laura Heymann 1:03:55
And then the question, I think, then becomes, yeah, we starting to talk then about negligence, right? Right, right, of course, right? And then, so, you know, if this is a sort of tort based regime right now, of course, that that raises important questions about duty, right, and all those kinds of things, right? And so, you know, is there a duty to the copyright owner, as opposed to their users? But I think the other direction, and maybe this is implicit in what you were saying as well, that the AI cases are going to put on, this is not just sort of towards negligence, but also towards direct infringement as well. Part of the question is going to be, well, who exactly, if the if the substantially similar copy is resulting from the prompt right? Is that the platform sort of model contributing to it right? Or is the, is it the user, you know, or they, in fact, do it right? And so, you know, that’s going to be a very nuanced question. And I think one of the many challenges, of course, with with Cox, is that, because it was sort of, well, this is an easy question, and what was answer the question that’s before us? And I think. You know, many people would agree that on the facts, it was the right result, but the court then doesn’t, sort of, you know, there’s no guidance, of course, into not only does it apply to platforms that host, as opposed to platforms that simply provide a connection, but then, you know, taking a step back, sort of, what is the distinction between any other utility, the electric company, or, you know, as many of the briefs sort of pointed out, right? And the court’s opinion isn’t helping to kind of categorize those sorts of applications at all.
Christopher Sprigman 1:05:29
Yeah, the court’s opinion makes it way too simple, obviously. But I guess for me, there’s a caution here, which is as difficult it is as it is to say this in 2026 when our Congress basically doesn’t function, I don’t think copyright should be judges through secondary liability doctrine should be balancing broad social harms. I think copyright should stay in its lane. And the particular social harms that we’re talking about are the copyright social harms are kind of difficult to calculate, right there we run into the same problem. So I think what’s going on is we’re testing the limits of what courts can actually do as well. And at some point, Congress is going to have to get into the game. And when they get into the game, it may not be copyright that gets into the game. It may be labor law, it may be it may be something like tax law. It may be subsidies, liability, products liability, exactly, tort law, generally, it may be a whole range of things, negligence, other than copyright. And as someone in this inside the copyright bubble, I think sometimes copyright looks like the hammer we have, but I don’t think it’s a particularly good one here.
Oren Bracha 1:06:37
I agree on both counts by the way.
Erik Stallman 1:06:41
Well then in terms of what, what copyrights lane should be, and I think this gets a little bit to the question of kind of overlapping IP regimes in something that you pointed out in your article, Laura, was that, like there was a, you know, like you shouldn’t have to elect which regime you’re you pursue your rights under but they should. It should influence the type of remedies that are available now, prior to the the the 76 Act there, I think there was maybe more of a case to be made that that copyright and patents were pursuing similar goals, because, because in both cases, protection sort of hinged on disclosure. It hinged on other you, sort of you, sort of make patent available, or you, you publication. But in the 76 Act we, we moved to fixation as a basis of protection. So are we starting now to is there more sort of justification for divergence in the IP regimes that they’re sort of like, the basis for protection is no different.
Laura Heymann 1:07:47
Yeah. I mean, I think, you know, maybe I kind of align myself, at least initially, with Justin’s point from earlier, which is, it’s just not clear why we necessarily have to have convergence, you know, just because they both appear in the same part of the Constitution, right? On the other hand, right? If you think about sort of the underlying goals. So again, getting back to contributor liability, for example, right? If the idea there is like this is just all rooted in a sort of torque regime, right? Then, why shouldn’t it be the same basic standard across all of the so I think it just, you know, that question is an important one, but I think answering it depends on asking, sort of, what are we talking about? What are we doing here? Right? And it’s so it’s not just about, well, copyright and patent, as I said, sort of have the same at their root, the same goals of incentivization and limited times and all that kind of stuff. But the underlying, sort of, well, why are we going to apply it and and sort of, is there, is there compatibility in regard to the function of why we’re going to sort of draw from one from the other? I think is an important question.
Oren Bracha 1:08:55
I echo in the same point, actually, I think there are deeper, even deeper reasons why there might be distinctions, which is, yes, those are all intellectual property regimes, which means they’re all property regimes in information, something intangible, but those are under the different regimes, very different kinds of information, right? Applied technological knowledge is not the same as expression. They’re very different. A lot follows from that. So it’s there might be deep distinctions on the other and I think that makes it all the more important that each regime actually have strong, that anticipates, perhaps the next panel, very strong subject matter rules that actually channel the right kind of information to the right kind of regime, and maybe to no legal protection regime.
Justin Hughes 1:09:45
This is statistical question I’d like to answer to but I my gut guess is that there are far more cases where copyright and trademark claims are brought together than copyright and patent claims, copyright and patent claims software, but copyright and trademark claims ubiquitous, and that’s for direct liability. But if it’s also true for secondary liability, then it would make more sense that you want the trademark and copyright rules to line up than the patent and copyright rules.
Christopher Sprigman 1:10:14
I think it’s less true now than it used to be for secondary liability, simply because the seconds are kind of push secondary liability and trademark away sufficiently from the copyright baseline that I think it’s complicated plaintiff’s approach.
Justin Hughes 1:10:29
But this is, this is something that I think is worth studying and figuring out what’s actually happening, and when I assume after Cox, more claims will be brought for trademark contributory liability, and that’s once that happens, we need to ask, Well, should we be lining these two up? More than patent and copyright.
Laura Heymann 1:10:48
You only take one action, right? So yes.
Oren Bracha 1:10:50
On the other end, maybe all the more reason to be careful this indistinguishing distinctions are properly because they’re easy.
Erik Stallman 1:11:01
Since I want to ask one more question and then open it up to the to the audience, buut and this will also echo back to something that we talked about this morning, which was sort of the sense that that something like the 76 act can’t really happen again, that we just live in an era of historically broken Parliament, and something like this is inconceivable, but, but also it’s I really wonder it was great that Judge McKeown pointed out that there have been 44 bills introduced since and none since the MMA, and none have been enacted, and even very minor changes when I, when I was a staffer, one thing I was charged with doing was, was helping through just an unremarkable three year extension of the license in 17 USC, 119, for, for the the compulsory license for, for distant carriage, satellite signals. It was harrowing, just because the the moment that you open up title 17 at all, everybody wants to run through that door and and anybody running through that door can make the whole deal blow up entirely. So I think I set some sort of personal record for cartons of cigarettes, but with that in mind, Justin, in terms of, like, what is actually achievable if you could make a distinct legislative intervention, and this could be also legislatively overturning a single case what is the intervention doing?
Christopher Sprigman 1:12:39
Could I just say, like, just as a preamble, I do not buy the idea that we could never do something like the 76 Act again. First of all, because I don’t valorize the 76 Act to the extent that some people do. And second, because I think throughout the American history, there have been these great waves of decline and then reform. I think we are hopefully near the bottom of a period of decline, and we are on the precipice of a period of reform. And why do I think that it’s just it’s happened again and again and again. The country has pulled itself out of just miserable situations, whether it be slavery or the civil rights movement or, you know, the Gilded Age, which maybe we’re having our analog to that. We’ve always managed, democracy has always managed to find a way. So I’m maybe just, I don’t know, maybe I’m just getting stupid as I get old, but I just feel more optimistic than that. You know, in terms of what I would suggest. You know, it’s funny, Oren said something that really resonated with me earlier. He said, you know, we could rely less on Fair Use if we had a substantial similarity test that that had some anchor to it, and the anchor that Oren proposes is substitution. So back when I was a much younger person, I wrote an article that basically said, hey, let’s, let’s like, let’s think about copyright a little bit like we think about antitrust. Let’s, let’s break copyright up into two different kind of lenses. One is a per se. So I said, like copying, like literal copying, or something close to literal copying, we treat that as per se. That’s, that’s, we do the typical infringement test, but copying, that is part of creating a derivative work, something distinguishably distinguishable, we should make plaintiffs show some kind of harm, okay? And that harm would be the possibility of substitution, or the likelihood of substitution. So people have been thinking about this for a long time. It’s not just me, like many others, have been thinking about this, and I do think that you know whether you believe in that or not, really gets back to what you think copyright is for. Since the beginning of my career, and to this day, I think copyright is a social welfare tool. I do not think it’s an author’s rights thing, at least not essentially, because I don’t really like believe in natural rights, and I don’t think rights talk gets us very far. I actually think we we make I’m a positivist. We make law to advance human welfare, and we think we’re addressing a market failure here, that would be an intervention that I think would be very divisive, but I think nonetheless, would do a lot of good.
Justin Hughes 1:15:11
I will say I agree with Chris that you shouldn’t assume we can’t do it. On the other hand, each person in the room should ask themselves, would they both the next 25 years of their life trying to do it. You know, remember that comprehensive copyright reform almost happened in 1930 Congressman Vestal got a comprehensive copyright revision package passed in the 1930 Congress. Not only did he get out of committee, he got it passed out of the house, he got it over to the Senate. Had Senate sponsors, and it was to come up to the floor for a vote, and low and behold, there was a filibuster on another issue, and Congress ended. Vestal had a heart attack the next year, bad luck, and the bill didn’t get reintroduced. So it’s all very historically contingent. And you know, in the last few years, there’s been observation that tech companies are not the golden children that they were, at least the non AI tech companies, although I guess that. I’m not sure that. So, you know, some something might be possible. I’m open to that.
Oren Bracha 1:16:20
I will just sign up to stupidity, I guess. And add briefly that you know, a time of great upheavals is also a time of opportunities you can apply, perhaps to copyright, maybe beyond.
Laura Heymann 1:16:37
So why not, four for four, but yeah, just to respond Eric, to your question. I mean, I, you know, it’s when building on what other people have said and written about. I mean, remedies is a is a ripe for reevaluation. I think right, it’s, I think it’s not necessarily that we need a sort of upheaval and a sort of complete rewriting of it, but to sort of better calibrate it to the either the harms or the inability to show harm. But, you know, we’ve just seen multiple examples of how the sort of, you know, just enormous damage, threat of enormous monetary damages, has caused, you know, change incentives.
Erik Stallman 1:17:21
Great. At this point, we’d like to turn it over to the audience. Andrew Bridges, I see you in the back.
Audience 1:17:36
Thank you for a great panel which has managed to push just about every button in me that it could push. First, just to the last point, I think statutory damages is the elephant in the room. I think it distorts all of copyright law. It distorts Fair Use, it distorts questions on substantive rights. Also, I think the other issue is we’re talking about consistency within intellectual property. But remember, in Sony, the Supreme Court said secondary liability here is just an instance of a general property of secondary liability across the law, not just IP. And I think that the Cox decision is explainable, and I’ll come to this in a second, both as a direct implementation of Grockster and of Twitter versus Taamneh, which was like Cox, also written by Justice Thomas. But I’d like to come to the to the Cox case, because Justin was criticizing it for the brevity of the opinion. I think the brevity of the opinion reflected it was a simple decision. It was a simple decision because Grokster said something that people refused to believe. Grokster replaced the standard it created for the first time at the Supreme Court, a standard for contributory infringement liability. Justice Ginsburg’s concurrence was useful because it actually stated the totality of contributory as being either inducement or providing something that was not capable of substantial, non infringing use. And if you go back and read Grokster, it’s very clear it never adopted the Gershwin standard. It never acknowledged it. And back in Sony, Sony referred to Gershwin, but did not adopt it, and said, well, if there is to be a liability in this case, is because of this type of knowledge. So that’s that’s the point I want to make about Grokster. I’d like to pick a bone with Chris, if I may also though on the question about burden of of proof. In fact, the Supreme Court made the burden of proof absolutely clear in the Sony decision. And I’m going to read a quote, if I may, respondents, respondents for the studios, the Supreme Court said respondents failed to carry their burden with regard to time to home, time shifting, and the Supreme Court also said it’s necessary to show some meaningful likelihood of future but future harm exists. If not presumed, the likelihood must be demonstrated. Those verbs all referred to what the plaintiff had to do, not to what the defendant had to do. But that got lost in Harper and Roe, which said it was an affirmative defense, which could have been a pleading matter, not a proof. It was only in Campbell that it became a proof issue. So I think one can go back to first principles in Sony to argue that the plaintiff has the burden. Thank you.
Christopher Sprigman 1:20:33
So I don’t disagree that we can go back to first principles in Sony and find the Supreme Court saying something that reads that way. It’s just that in other cases, the Supreme Court’s been like, straight up clear that Fair Use is an affirmative defense. I mean, what do we take from this? One is that the Supreme Court contains multitudes, and we’ve known that for a long time. And the other is that, you know, those multitudes kind of filter down into the lower courts, and the lower courts pick up on something and don’t pick up on others. So what have the lower courts picked up on? They’ve picked up on? They’ve picked up on the Supreme Court’s clear statements about it being affirmative defense. They haven’t read the tea leaves of Sony the way you have. Now I have to, I have to say, like in my life, you’re the only person who ever told me how Cox was going to come out. I never believed you. I still don’t believe you, even though, clearly you are correct. And so, you know, like, I address this question with a great deal of respect, the only thing I will say about, you know, if you, if you go your argument about, you know what Grokster meant, Grokster didn’t squarely involve the question of, like, what is contributory liability as a whole. It involved one particular thing, which was whether Sony applied. You know, the briefs are arguing over whether Sony applies, and the Supreme Court basically says, surprise, there’s this kind of third way. It’s inducement. And inducement goes way back in the common law, and we’re going to apply it here, because these people were convincing. They were jawboning other people into so I understand Justice Ginsburg’s concurrence, and I understand your argument about it, but knowing Justice Ginsburg a little bit and her views on copyright, it’s hard for me to believe that that’s what she meant, right? So I think she must have been speaking within the framework of what the parties were arguing about, and that’s where I leave it. But clearly you were on to something, so I must acknowledge
Audience 1:22:20
Grokster invoked Sony and claimed to be applied and said that Sony was an intent case when we all know it was a knowledge case.
Christopher Sprigman 1:22:29
I hear you.
Justin Hughes 1:22:35
I’m not going to say anything. I’ll just say that I find the reference to Ginsburg’s concurrence interesting because Ginsburg and Breyer take such different views of what substantial non infringing uses is, and to cite Ginsburg opinion is to point in one direction for how to interpret that. But…
Christopher Sprigman 1:22:58
So my question is, was there a narrower ground that the Supreme Court could have decided the Grokster case? I mean, sorry, the Cox case under? I think that, that that was the that was the surprising thing to me. But if you read the amicus briefs, the amicus briefs basically all went one way or the other way. There wasn’t any sort of exploration about middle ground. And I think that’s actually different from the Grokster case, where, in Grokster there were extreme briefs on both sides, but there was also a lot of briefs about inducement, even though inducement had not been in the complaint right, it was contributory and vicarious, and they never argued inducement. But the reason that the Supreme Court could say that inducement was was a viable thing is to say, oh, it’s part of contributory liability, when, in fact it has different elements so and it’s a separate statute in the patent law. And so we’re in this kind of conceptual craziness a little bit because the Supreme Court kind of fudged the thing by saying that that inducement was part of contributory just because Gershwin had said that. But I’d love some comments on what, what are the narrow grounds that the court might have been able to reach instead of kind of the one extreme or the other?
Justin Hughes 1:24:34
Well, I’m going to make a plug for co-panelists Laura and Fred Yen filed an Amity brief recommending exploration of tort principles more fundamentally.
Laura Heymann 1:24:44
I’m not sure if that’s narrower, but at least a different, right, and I you know, I can see why the court, given what we’ve seen in Cox, wasn’t sort of willing to take that on right but, but I think we might see that as lower courts try to figure out what they should be doing with this, I also agree, though. I mean, the inducement is interesting, right? And Gershwin does use the word right, but conceptually, it’s a sort of different kind of activity from the other kinds of activity. So whether that should all fit under sort of a contributory umbrella. So that’s why, you know, I mean, one of the things we always teach our students right is sort of precision and in the way we speak and and I think part of the problem with the development of that area of law has been that gerswin kind of mush, those, you know, a bunch of different concepts into one sentence, and there has been a sort of repetition of that sentence without an attempt to really unpack that.
Christopher Sprigman 1:25:39
Yeah. So one narrower thing would have been just to do what Justice Sotomayor chided Justice Thomas for not doing, which is like follow his own opinion on Twitter versus Taamneh, right? So Twitter versus Taamneh basically says that most of the time there has to be some affirmative act, but not all of the time. There are some circumstances in which nonfeasance can amount to evidence of like an intent to further the scheme, right? It’s not a it’s not a black and white rule, although the thumbs on the scale, if it’s nonfeasance against liability, but the thumbs not determined is not definitively on the scale. So I think that would have been narrower. And I think if that were the way that it came out, courts would be lower courts would be parties would be fighting over like, what kind of nonfeasance do we have here? Does it? Does it amount to intent? Under what circumstances? There’d be a whole bunch of detail to develop, but that would have been narrower. I mean, that’s essentially what the ACLU brief recommended, and, you know, just Justice Sotomayor essentially picked it up and she she made the point that it’s kind of weird that the majority doesn’t follow justice Thomas’s own approach only a couple years old.
Audience 1:26:54
Yes, great conference, but I’m noticing that so far as a preface comment, the First Amendment hasn’t really come up at all, not in terms of how it interfaces with Fair Use. But the point I want to make relevant to this discussion about secondary liability is secondary liability, and particularly as exemplified in the Cox case, is so deeply connected to the internet and how the internet is architected in order to facilitate speech where we are dependent on intermediaries in order to have this happen, and when there is secondary liability, that has been such a censoring pressure on platforms where we’ll protect them with section 230 and to some extent, with the DMCA. But the real pressure that’s on them is kind of a jawboning pressure of fear of secondary liability. And as Thomas noted in Cox, secondary liability isn’t actually in this 1976 law that we’re celebrating. This is entirely judge main and it has a censoring quality. Now they wrote their whole decision, not mentioning the First Amendment either, and not really mentioning the communication and expressive ecosystem that’s at stake. But I read this case in combination with Taamneh as the court substantially limiting secondary liability as a common law principle as applied to the internet. And although I think Sotomayor is probably right that maybe Thomas made it too narrow. Even if you take Thomas’s decision and glue on aiding and abetting liability, we have a much narrower universe of the type of secondary liability that could attach, and that’s critically important, and given that all of our communication these days is happening with the effect, with the help of intermediaries, who are now much freer to be able to do that, because when they were facing this sort of enormous, unmeasured, unlimited, no statutory language to guide them, legal risk. They were having to say no to so much of it, and we were seeing censorship. I mean, this, Cucks didn’t even wrestle with terminating users, but I do think it’s notable that all that both Thomas and Sotomayor were like, you can’t get knowledge out of an IP address. It doesn’t tell you that anything happened wrong, and that is, I think, a very important thing that came out of this decision. So I’m sorry that was more of a rant than a question, but, but I think also in terms of what Justin was saying with trademark liability, I think that one would get tamped down in the same ways of saying that common law under secondary liability principles is narrow, and we’ve been over expanding it, and it sounds like the court is really done with that.
Oren Bracha 1:29:50
I’ll say two things quickly. The first one, I think just broaden the scope. I think basically the First Amendment, not necessarily as a hard. Constitutional limit, but as a set of values and purposes. You can call it open public sphere, if you wish. It’s basically hovering over everything we’ve been talking about, right? Not just secondary liability, but everything. It’s about the internal balancing mechanisms of copyright taken seriously, not just hand waving, copyright takes care of it, but actually, how should we do it within copyright? That’s one. The other point is, I just want to repeat again. Sorry for that. At least in the AI context I repeat. I think the bigger structural danger for from the point of view of gatekeepers and children, effect is actually, at the moment, less secondary liability and more direct liability with respect to similar output.
Erik Stallman 1:30:50
Okay, well, we would we have one minute left. There is one question I would like to pose to the panel, just because it’s been top of mind for me and we talked about this a little bit earlier, which is it relates to the test of infringement. So, so we have, we have the recent panel decision in Sedlik versus Von Drachenberg, where the panel is basically begging their colleagues to take the case and bank so they can revisit what is the test for for infringement. And also bound up in that, in that, in that case, because it also involves overturning a jury determination on substantial similarity is who gets to make that decision. And a less noticed aspect of that litigation was that there was also a challenge to juries having the ability to to answer the to issue general verdicts on the question of Fair Use. So is that so one, I take it from the panel that most of you are okay with, Sedlik going en banc, but should those two issues be bound together in terms of what is the test and who is the decider?
Christopher Sprigman 1:32:02
I mean, sadly, going on bunk it’s funny. So people tried to game out soon after that came down with like, what would that likely end up doing? And I mean, it’s weird because it would likely end up in a verdict for the plaintiff in that case. But I think more broadly, it would likely end up in a substantial similarity test that is quite possibly on the margin, less friendly to plaintiffs. So I think, you know, it’s a little bit like you focus on short term and long term.
Oren Bracha 1:32:38
Look, we could add layers of complications to the analysis right of the of the infringement, just throw in experts opinion right on top of everything, my bottom line is, whatever the layers of complications and complexity, we should have clear guiding concepts understanding what this is actually about. And I’ve already stated my opinion, right.
Erik Stallman 1:32:59
Well, in the interest of not adding layers of complexity, we will go ahead and end there. Thank you. Thank you all.