In recent years, a lively scholarly discourse has emerged about whether and how the U.S. First Amendment protects generative AI outputs. Some have argued that such outputs are protected at the very least by the rights of AI users to receive information and to create their own speech. Others would not recognize gen AI outputs as receiving such First Amendment protections. The issue of whether gen AI program creators are entitled to free speech protections for such outputs has also spurred a variety of reactions.

As this discourse on the First Amendment and gen AI unfolds, it is also important to reflect on what the global freedom of expression standard has to say on the matter. This standard will affect discussions about national and regional regulatory approaches to gen AI throughout the world. In addition, global corporate responsibility standards call on companies to respect international human rights norms in their operations, which may also impact how businesses that provide gen AI services approach their activities.

Part I of this Article explores the scope of the existing global free expression standard. Part II considers the standard’s application to gen AI outputs, including in various governmental and corporate contexts. Ultimately, this Article maintains that the global free expression standard protects the rights of individuals to seek and receive information of any kind, including gen AI outputs.

In addition, if human speakers share gen AI outputs as part of their own speech, this global standard also protects those speakers’ right to impart information. Governmental attempts to restrict gen AI outputs are therefore subject to the standard’s safeguards on how this human right can be limited. And companies providing general-purpose gen AI services should also respect human rights, including freedom of expression, in their operations.

Starting in the 1980s under President Reagan, the FCC curtailed or abandoned most of the content regulations of broadcasting that it had earlier promulgated. That pattern continued through the Biden Administration: Such regulation was largely dormant. Statutes regulating content existed (although, with the episodic exception of indecency, they were interpreted narrowly). But regulations created by the FCC were either repealed (notably the fairness doctrine), never applied (e.g., the broadcast hoax rule), or applied so rarely and weakly as to be toothless (e.g., the broadcast news distortion policy).

Broadcast television may seem like a relic of the past to some readers of this Article. But it garners more than 20% of all television viewing, and the legal issues it raises are important. The developments presented in this Article highlight important questions about the desirability of regulation and tell an instructive story about roads taken and not taken.

There are two key elements of the legal landscape for broadcasting. First, every license is for a term of years and then is subject to renewal under the statutory “public interest, convenience, and necessity” standard. The same public interest standard applies to any application to transfer a license from one entity to another, which means that any corporate merger or acquisition involving licenses requires the FCC finding that the transfer is consistent with the public interest. And, as judicial opinions and Commission practice make clear, the public interest standard extends well beyond concerns about competition.

How can the FCC exercise such pervasive control over a means of communication? That implicates the second key element of the legal landscape—Red Lion Broadcasting Co. v. FCC and FCC v. Pacifica Foundation.

Concerns about the public interest standard and Red Lion/Pacifica are not new. What, if anything, changed with the coming of the second Trump Administration?

First, the second Trump Administration has been more interested in regulating broadcast content than any of its predecessors since Reagan….

Second, the Supreme Court’s likely overruling of Humphrey’s Executor v. United States will mean that “independent agencies” like the FCC are subject to presidential control. …

Third, Trump has imposed more control over the executive branch than any President in at least a half century….

The FCC’s self-restraint on content regulation of broadcasting from Reagan through the Biden Administration could have been rejected at any time. One question is why the second Trump Administration has been more aggressive than its predecessors in pressuring broadcasters. Another is why its predecessors (including the first Trump Administration) did not make similar choices. There is of course no easy answer to either question, or to the related question of the role played by regulatory philosophy, political considerations, powerful interests, and the like in producing the restraint from Reagan through Biden or the apparent waning of such restraint in the second Trump Administration….

Although the story of broadcast content regulation from Reagan through Biden was largely one of continuity, outside of indecency regulation the biggest supporters of a broad and flexible application of the public interest standard have generally been on the political left, and the biggest detractors have generally been on the political right. Thus a question going forward is whether the second Trump Administration has discredited, or will discredit, broad governmental power for those on the political left—particularly with respect to speech—and has led, or will lead, those on the political right to be more supportive of such power….

This Article is … a case study of the easiest form of communication for a presidential administration to effectively exert some control over, but broadcasting is not the only medium an administration can influence. In this way, this Article implicates broader questions about regulation: What, if any, forms of content regulation are normatively desirable? How meaningful are speech protections in a world where the government can exert unseen pressure? How effective can laws be in constraining presidential administrations that maximally exercise the levers of power available to them?

Part I lays out the FCC’s strikingly broad statutory authority over broadcasting and the Commission’s early history of aggressive content regulation.

Part II focuses on the FCC’s curtailment of its regulations. This curtailment began in the Reagan Administration, and many commentators expected it to be reversed under a Democratic President. But, with the exception of indecency regulation in 1987 and the early 2000s, the light-touch regulation initiated under Reagan largely persisted through Biden.

Part III turns to the question of how this regulatory regime can be consistent with the First Amendment. The answer lies in the reasoning and holdings of Red Lion and Pacifica. Those cases could have been generative, but the Supreme Court has limited their application to broadcasting.

Part IV notes that not only can a new presidential administration revive old doctrines but it also can create new ones, because administrative law doctrines allow it to do so. These administrative law doctrines, combined with the public interest standard, thus greatly empower the government.

Part V considers possible ways of preventing or limiting a reinvigoration of content regulation of broadcasting. Overruling Red Lion and Pacifica or repealing the public interest standard are obvious options, but other options are available—notably lawsuits to stop jawboning.

Part VI, however, highlights the limits on these options. A President who makes it clear that he will impose maximum costs on those he dislikes by berating them can effectively achieve the results of jawboning without having to issue any threats; his targets will know that failing to do the President’s bidding will lead to costly castigation. This highlights an important point that transcends broadcasting: Norms do a tremendous amount of work, so a breakdown in norms has massive implications.

Since Hamas’s October 7, 2023 attack on southern Israel, the start of the subsequent war in Gaza, and an outbreak of campus protests expressing vehement hostility to Israel, dozens of universities across the United States have faced lawsuits, administrative complaints, and federal investigations alleging dereliction of their Title VI duty to protect Jewish students from discrimination. These Title VI claims typically include allegations that universities have been deliberately indifferent to the way actions by anti-Israel protesters have created a severe and pervasive hostile environment for Jewish and Israeli students.

Public attention has focused on perceived hateful speech by anti-Israel protesters. This includes slogans seen as calling for the destruction of Israel (e.g., Arabic chants of “From River to the Sea, Palestine is Arab”); speech that is often interpreted as calls for violence against Israel, Jews, and their supporters (e.g., “Globalize the Intifada” and, in Arabic, “Khaybar Khaybar oh Jews, the army of Muhammad will return”); expressions of support for Hamas in general (e.g., “Glory to the Martyrs”) and its October 7 crimes in particular (e.g., flyers that feature paragliders, glorifying the Hamas terrorists who massacred Israelis after flying into Israel on paragliders); and pro-Hamas events on October 7, 2024, the anniversary of the massacre.

Judicial decisions and scholarly commentary have tended to frame the Title VI claims primarily as conflicts over political expression of this sort by anti-Israel protesters. In late 2023, the federal Department of Education advised university officials that in the context of anti-Israel protests, constitutionally protected speech alone can create a hostile environment that universities are obligated to address, albeit without suppressing speech.

Much of the subsequent debate, in courts and otherwise, has been over whether rhetoric widely perceived as antisemitic or as endorsing violence in the abstract can ordinarily support hostile-environment liability without violating the First Amendment. Jewish students’ claims should be rejected, some argue, as objections to rhetoric the students find deeply offensive but that is constitutionally protected. A recent attention-getting article asserted that Jewish students’ claims may not ordinarily rely—even merely as evidence of an overall hostile campus climate—on campus protesters engaging in such incendiary rhetoric.

This Article argues that prevailing analyses of post-October 7 Title VI claims are incomplete and are often legally mistaken because they ignore or misinterpret a central element of hostile-environment doctrine: context. Hostile-environment law asks whether conduct is so severe, pervasive, and objectively offensive that it effectively denies students access to educational opportunities. This inquiry is contextual and depends on the “constellation of surrounding circumstances.”

The context, or surrounding circumstances, neglected by many analysts is the surge of antisemitic assaults, threats, vandalism, and intimidation on and around American campuses. When rhetoric endorsing or glorifying violence occurs against a backdrop of credible threats and actual violence, it may reasonably be experienced as threatening even if the same words would otherwise be understood in a different context as merely abstract or ideological rhetorical excess or overkill. Treating Jewish students’ fear of violence as irrelevant or irrational improperly strips the contextual analysis from the “reasonable person” standard used in hostile environment law.

Yet recent judicial decisions, to the extent they have recognized it at all, have treated Jewish students’ fear of violence as either irrelevant or irrational, even where that fear is grounded in recent incidents of assault, intimidation, and lawless conduct directed at Jewish students and at Jewish institutions. In doing so, courts have effectively assumed away the very conditions that may make certain speech threatening rather than merely offensive.

This does not mean that chants, slogans, flyers and other speech perceived as endorsing violence automatically created or even contributed to a hostile environment. Moreover, even if speech does create or contribute to a hostile environment for Jewish students, such speech is constitutionally protected, so long as it does not rise to the level of incitement or a “true threat.” Universities therefore do not and cannot have a Title VI obligation to suppress protected political speech.

The question, however, is whether Jewish students may have a valid Title VI action for university malfeasance or inaction when it comes to policing behavior by protesters that is not constitutionally protected. The violence-endorsing speech, rather than being the basis of a Title VI claim, serves as relevant context for why the universities had a Title VI obligation to enforce their own rules and the law.

As described later in this Article, since October 7, Jewish students have faced a surge of physical assaults, credible threats, vandalism, and coordinated intimidation against them on campuses nationwide, alongside a broader national wave of antisemitic violence. Against that backdrop, Jewish students’ fear of violence is not speculative or subjective, but empirically grounded. When such fear undermines students’ access to education, universities have a Title VI obligation to address the overall hostile climate. They can do so, without suppressing protected expression, by enforcing neutral rules against trespass, vandalism, threats, and other illicit conduct not protected by the First Amendment.

Failure to account for students’ fear of violence has concrete doctrinal consequences. Courts have dismissed Jewish students’ hostile-environment claims at the pleading stage by characterizing allegations as mere discomfort with opposing viewpoints, even where plaintiffs allege assaults, threats, encampments that block access to parts of campus, and administrative indifference to lawless conduct. By excluding fear of violence from their contextual analysis, these decisions collapse the distinction between ideological offense and intimidation, and they risk turning Title VI’s “reasonable person” standard into an abstract exercise divorced from reality. Worse yet, some courts and commentators have treated illicit actions by anti-Israel protesters as if they are constitutionally protected speech.

This Article proceeds in five parts. Part I documents the surge of physical violence, threats, and intimidation targeting Jewish students on American campuses since October 7, 2023, demonstrating that rhetoric endorsing violence operates within an atmosphere of tangible danger. Part II situates campus developments within a broader national context of antisemitic violence against Jewish individuals and institutions, reinforcing the reasonableness of Jewish students’ fears that they will be the targets of violence. Part III examines the organizational and ideological ties between leading sponsors of anti-Israel campus protests and groups that openly endorse violent “resistance,” explaining why certain chants and slogans may reasonably be perceived as menacing in context. Part IV draws out the implications for Title VI doctrine and university compliance, arguing that courts can and should require universities to address illicit conduct and threatening environments without infringing on First Amendment protections. Part V addresses First Amendment objections to using protected speech as evidentiary context in hostile-environment cases by explaining why universities’ failure to suppress related illicit, unprotected actions has violated Title VI. This Part rejects the notion that universities warrant distinct judicial solicitude in their regulation of unlawful conduct connected to student protest, and also concludes that the chilling effects doctrine does not apply to rulings that allow complainants to rely on hateful, violence-endorsing speech as contextual background in a hostile-environment case.

This Article concludes by emphasizing that courts and administrators must interpret Title VI’s “reasonable person” standard in light of the contemporary reality of antisemitic violence, on and off campus. Courts can demand that universities prohibit illicit behavior by anti-Israel campus constituents, such as trespass, vandalism, and threats, while still protecting the rights of protesters to express their opinions.

John Stuart Mill continues to stir trouble in the twenty-first century. Nowhere is Mill’s liberal legacy more fiercely contested than in academia, where scholarly opinion ranges from something approaching idolatry to icy disapproval. Heterodox Academy invokes Mill as a leading light, with Jonathan Haidt and Richard Reeves publishing an abridged version of On Liberty (entitled All Minus One) as a manifesto for the movement. Other scholars have tried to “update” Mill for a new era by adapting his notion of “harm” to include psychological and civic harm. Still others dismiss Mill as an engineer of and apologist for empire, another hypocritical nineteenth-century liberal who supported freedom for the metropole while advocating despotism for the colonial other.

How to reconcile the various versions of Mill—liberal apostle, Romantic utilitarian, faithful servant of the East Indian Company? The volume of Mill scholarship may hinder any effort to answer the question more than it helps: It is now so vast that it seems impossible to master without the help of artificial intelligence: Even if we restrict the topic to Mill and freedom of expression, JSTOR alone houses some 6,000 books and articles that mention Mill and free speech, enough to intimidate even the most energetic reader. The prospect of saying something new about Mill is perhaps yet more daunting.

To their credit, Drs. Christopher Barker and Fara Dabhoiwala have recently paved a fresh pathway through Mill’s oeuvre, attempting to explain why Mill did not extend his full-throated support for free speech to what he termed “barbarous” countries. Why, they ask, did Mill insist that colonial Indians, for instance, did not deserve the right (or privilege) of free speech? While copious research has been devoted to Mill’s arguments on free speech and to his opinions on empire, surprisingly little has been written on the intersection of the two.

The problem, however, is that the query on India, as posed, est une question mal posée, a complex question that assumes the conclusion and then seeks to explain it. Barker’s claim, for instance, that “Mill consistently supports East India Company (EIC) policy restricting publicity and debate in British India” is belied by a more sensitive examination of the evidence. Dabhoiwala’s forays down the documentary trail in his new book, What Is Free Speech?, are even more misleading. The real question is how two respected scholars could have limned such skewed portraits of Mill.

One of the greatest privileges of being a federal judge is the honor of presiding over a naturalization ceremony. I do it every year in May, to celebrate the anniversary of my own naturalization in May 1982.

I wasn’t born in the United States. I didn’t enter this world as an American. But I wake up every morning thanking God that I will leave this world as an American.

I like to say that I’m Taiwanese by birth, Texan by marriage. But most importantly, I’m American by choice.

If you’ve never attended a naturalization ceremony, there’s nothing more inspiring. People from all around the world come together in one room, for one purpose—to become Americans….

In a nation of over 300 million Americans, we’re bound to disagree on virtually every issue under the sun. And that’s okay. There’s nothing wrong with that….

We’ve all heard the maxim, “I disapprove of what you say, but I will defend to the death your right to say it.” … [But i]f you truly detest what someone says, why on Earth would you want to defend their right to say it—never mind defend it to the death?

Do you know anyone that altruistic? I don’t.

So if the statement makes any sense, it must be because it’s not altruistic at all. In fact, it’s quite self-interested. In fact, I would say that it’s self-interested in at least three distinct ways. With your indulgence and patience, I’ll spend a few moments examining each one.

First: On any given issue where people disagree, you may well be right—and those who disagree with you may well be wrong. But you could always be better.

Or put it another way: You may be right. But you should also want to win. And the way to maximize your chances of winning is to constantly expose yourself to competing viewpoints—and to do so as vigorously and frequently as possible….

[Second, i]n any particular disagreement, you may well be right. But you also could be wrong.

And you need to be okay with that. Because if your goal is not to soothe your ego, but to discover the truth, then you need to be prepared to be wrong at any given moment, on any given issue….

[Third,] regardless of whether you’re ultimately right or wrong on a particular issue, I would submit that respecting disagreement will make your life much better. It will also make the world a far more pleasant place to live….

An enraged King Henry II famously asked several of his barons, while the court met in Normandy at Christmas in 1170, “Will no one rid me of this turbulent priest?!” King Henry was referring, of course, to Thomas Becket, the incumbent Bishop of Canterbury and head of the Roman Catholic Church in England. Becket had steadfastly insisted on preserving the independence of the ecclesiastical courts and authority over bishopric appointments (much to King Henry’s dismay). Subsequently, four of these same barons went from Normandy, in France, to Canterbury, in England, and proceeded to assassinate Becket, on December 29, 1170, while he was conducting a prayer service in Canterbury Cathedral. The perpetrators (shockingly) escaped the king’s justice. Perhaps Becket received some small justice, however, when Pope Alexander III proclaimed him a martyr and saint on February 21, 1173.

Could Henry II’s musing, consistent with the requirements of the First Amendment, serve as a basis for criminal charges in the contemporary United States? On a first cut, one might answer this question negatively. After all, Henry II was in France; Becket was in England. His speech could not have produced an “imminent” threat to Becket’s safety. On the other hand, the barons who assassinated Becket did so only four days after Henry II pondered aloud his abstract desire to be rid of “this turbulent priest”—which, for the Middle Ages, constitutes a rather rapid response.

Even if Henry II might not be on the hook for criminal incitement charges, which under the Brandenburg test require both a clear call to unlawful activity and circumstances in which the speaker’s audience would likely act on the suggestion of unlawful action, a clever prosecutor in today’s United States could evade the Brandenburg rule entirely through the expedient of charging Henry II with solicitation or conspiracy (rather than incitement). Would the First Amendment preclude solicitation charges on these facts? Or, for that matter, conspiracy charges? Moreover, should changing the criminal charges fundamentally affect the governing First Amendment analysis?

Current free speech doctrine offers a less than clear answer—but most likely solicitation charges would stand up against a Free Speech Clause challenge. The controlling Supreme Court precedent on point, Williams, decided in 2008, seems straightforwardly to hold that proposing a criminal action enjoys absolutely no First Amendment protection. The Supreme Court doubled down on this approach in 2023, in Hansen, holding that “[s]peech intended to bring about a particular unlawful act has no social value; therefore it is unprotected [under the First Amendment].”

To be sure, neither of these cases expressly holds that speech proposing a crime, in contexts where the speech activity clearly constitutes expression about a public policy question, may be criminally punished. Moreover, the facts in both cases do not involve advocacy related to matters of public concern. Nevertheless, neither majority opinion bothers to distinguish speech with no relationship to democratic deliberation (for example, an effort to hire a hit man to commit a murder) from hyperbolic speech at a mass public rally (for example, an impassioned call to support illegal immigrants and immigration, and advocating taking direct actions to facilitate it, such as leaving jugs of water in the Arizona desert, because the speaker believes U.S. federal immigration policies are cruel and unjust).

Accordingly, whatever residual doubts might exist about charging Henry II with incitement would greatly diminish if a prosecutor were to charge solicitation or conspiracy instead. This outcome, of course, does not make any sense normatively or as a matter of public policy. First Amendment protection for speech that arguably calls for unlawful action should not depend on the vagaries of the indictment….

[T]his Essay proposes a different, more holistic approach to bringing the criminal law to bear on speech. Just as the Supreme Court’s effort to reconcile tort law with the imperatives of the Free Speech Clause did not end with New York Times Co. v. Sullivan and the tort of libel, free speech principles and values should inform criminal law enforcement too. A careful review of the relevant cases involving speech-crimes shows that, in all save one case (namely Dennis), the speech activity clearly caused, or had an obvious potential for causing, a serious social harm. The federal courts should apply the harm principle to any and all speech crimes—not just to the crimes of incitement and threats. Moreover, switching the crime charged from incitement to solicitation, for example, should not render the First Amendment wholly irrelevant any more than changing torts from libel to intentional infliction of emotional distress would foreclose a meaningful First Amendment analysis. And yet at worst this is unfortunately more or less how things seem to work today and, at best, the Justices have not taken sufficient care to disentangle speech integral to committing a crime from hyperbolic political advocacy.

Part II of this Essay discusses, in some detail, Justice Oliver Wendell Holmes’s iconic dissent in Gitlow v. New York and his argument that all ideas are potentially incitements to action as well. Because ideas and actions are inexorably intertwined, a legal test that turns entirely on a federal judge arbitrarily applying one label or another is really no test at all. Part III then considers the Supreme Court’s treatment of speech crimes and argues that, although the Justices have so far never clearly articulated the harm principle beyond Alvarez, this concept clearly and best explains the outcomes in all of the relevant cases save one (again, Dennis).

In Part IV, I draw upon Justice Anthony Kennedy’s important opinion in United States v. Alvarez to argue that speech can constitute a crime consistent with the Free Speech Clause of the First Amendment only when the government can prove, beyond a reasonable doubt, that it causes or would be highly likely to cause some kind of legally cognizable harm. Consistent with John Stuart Mill’s cogent arguments in On Liberty, and the plurality’s holding in Alvarez, the fact that speech offends or scandalizes many within the community, or might even have bad tendencies, must not be constitutionally sufficient to justify the imposition of criminal punishment. Part IV also explicates the harm principle and how the federal courts should deploy it consistently across the criminal law to limit the potential censorial effects of speech crimes. Finally, Part V provides a brief summary and conclusion.

For present purposes, Justice Holmes’ recognition that all ideas are also incitements constitutes his most important contribution in Gitlow to theorizing how the freedom of speech and punishment of crime should be reconciled. Because every idea is, potentially, an incitement, giving the government carte blanche to punish incitements would necessarily involve giving it a free hand to extirpate ideas that it dislikes from the marketplace of ideas. But the risk of government censorship is not limited to the crime of incitement; it exists too when the government charges a solicitation, a conspiracy, an attempt, a threat, or disorderly conduct. In any given case, the government should have the burden of showing that speech serving as the basis for a criminal charge (of any type) either caused or was highly likely to cause a serious social harm; if the government cannot meet this burden, the federal courts should hold the speech protected under the Free Speech Clause of the First Amendment.

As we celebrate the 100th anniversary of Gitlow v. New York, it is only natural that we focus on the landmark presumption of incorporation in the majority opinion and the dissenting opinion that would foreshadow today’s incitement standard. But perhaps we can spare a moment to consider the dedicated lawyers who represented Gitlow and his fellow radicals throughout the process. That their efforts, at least in the short term, were largely futile is certainly attributable to the historical moment and the prevailing interpretation of the First Amendment, not to any lack of competence or commitment on their part.

Indeed, the lawyers who represented Benjamin Gitlow were among the best of a small cadre of lawyers who represented leftist radicals of various persuasions in the early 20th Century. Some, like Clarence Darrow, would achieve great fame in their own time; others, most in fact, would labor in relative obscurity and are largely forgotten today. And while all were dedicated to their radical clients, their backgrounds, career paths, and personal motivations could be quite different. This article will profile two of Gitlow’s lawyers, Charles Recht and Walter Nelles. Both were involved from the earliest stages of Gitlow’s representation, although only Nelles played a significant role in the U.S. Supreme Court proceedings. Along the way, we will encounter Darrow, Walter Pollak, and others who made substantial contributions to the Gitlow litigation.

In examining the representation of Benjamin Gitlow, let us keep in mind the various interests that powered it. Fundamental, of course, is the interest that all criminal defense lawyers have in the acquittal of their clients. But then some lawyers and clients may have political or personal interests that color their litigation strategy and tactics. And some may have philosophical or ideological goals that transcend mere acquittal and seek to alter the legal landscape as much for the future as the present. All of these factors are present in the Gitlow representation.

On a busy Saturday afternoon in 1940, Walter Chaplinsky took to the streets of Rochester, New Hampshire, to distribute literature promoting the faith of Jehovah’s Witnesses and denouncing all other religions. At one point, Chaplinsky encountered the City Marshall, whom he called a “damned racketeer” and a “damned Fascist.” New Hampshire charged Chaplinsky under a criminal provision restricting “offensive” speech. In upholding Chaplinsky’s conviction a unanimous United States Supreme Court asserted that “[a]rgument is unnecessary to demonstrate that the appellations ‘damned racketeer’ and ‘damned Fascist’ are epithets likely to provoke the average person to retaliation, and thereby cause a breach of the peace.”

When teaching Chaplinsky to law students, I walk over to a nearby student and, in an appropriately loud and threatening voice, call the student a “damned racketeer.” To date, not once have my words triggered a breach of the peace. The obvious point is that context matters. Chaplinsky and I may have spoken the same words, but the meaning of those words is determined at least in part by the context in which they are uttered. This latter observation has generated volumes of work in legal theory, linguistics, and philosophy. I won’t revisit those debates here except to lay my own cards on the table: Meaning is somewhat but not entirely determined by context. To ignore context would mean a rigid fundamentalism; to defer to it entirely would mean an open-ended pragmatism without foundations. Neither of these options accurately describes the social world in which we live. But understanding context gives us a clearer sense of how to understand changed meaning not only among words but also among relationships, politics, and societies.

I think this is what is ultimately at stake in Oliver Wendell Holmes’s famous dictum written a half-generation before the Supreme Court so confidently classified Chaplinsky’s utterance as fighting words likely to breach the peace. That dictum, from Holmes’s dissent in Gitlow v. New York, asserts with no less confidence that “every idea is an incitement” and that “the only difference between the expression of an opinion and an incitement in the narrow sense is the speaker’s enthusiasm for the result.” I want to suggest Holmes is right to assert that every idea is an incitement, but his subsequent focus on the speaker’s enthusiasm neglects other important contextual factors. Chaplinsky‘s context included the generally understood meaning of the speaker’s insults at the time he uttered them and the physical environment in which he uttered them. Without this additional context—in other words, without moving beyond merely “the speaker’s enthusiasm for the result,” we cannot adequately assess the likely harm of a speaker’s words or whether the state should be permitted to limit those words based on that harm.

Most incitement cases also include an additional contextual factor: the audience that hears the words uttered and then chooses whether to act on those words. For this reason, most incitement cases involve not only speech but also assembly. But a speaker’s words do not directly correspond to a hearer’s actions. Hearers have their own agency, which means that they help determine the ultimate meaning, consequences, and effects of the speaker’s words. This is one reason that groups—assemblies—often pose a greater risk of uncertainty and instability than individuals. Having more people in the mix complicates both communication and action.

The relationship between speaker and hearer is further confounded by the variable of time. Chaplinsky’s words were limited to the audience immediately in front of him. But many incitements unfold over time. Think, for example, of a sustained protest, a social movement, or a revolution. In these cases, we will not always know what words will come to mean or what their eventual effect will be—we will not know how listeners separated from the speaker over time might interpret and act on those words.

In the sections that follow, I explore the preceding claims and suggest why Holmes’s aphorisms in his Gitlow dissent insufficiently account for the contingencies of listeners, and expose vulnerable speakers and organizers to liability for downstream and unintended consequences of their words. I illustrate with two contemporary applications. The first is a misguided doctrinal development out of the Fifth Circuit: the concept of “negligent protest.” In Mckesson v. Doe, the Fifth Circuit suggested that a protest organizer could be personally civilly liable for the downstream actions of another protester. Following a remand and the Supreme Court’s intervening decision in Counterman v. Colorado, the district court granted summary judgment to the protest organizer, noting that he “cannot be held liable in negligence for actions taken while exercising his First Amendment freedoms.” While Counterman and the district court’s opinion limit the likely scope of negligent protest, the Fifth Circuit’s novel expansion of liability remains on the books.

The second application focuses on the words of Donald Trump and the ensuing actions of his supporters rioting at the United States Capitol on January 6, 2021. The Capitol riot illustrates the magnitude of the danger of speech uttered in assembly—in this case, the violent assault on a core function of our democratic government. But even here, Trump’s “enthusiasm for the result” of his words is far from clear, and it may well be that the same civil liberties that ought to shield the protest organizer in Mckesson should also insulate Trump from the consequences of his words on January 6th.

Open courts are a cornerstone of democratic government, ensuring not only that justice is done but that it is seen to be done. As the Supreme Court recognized in Press-Enterprise Co. v. Superior Court (”Press-Enterprise I“), “[p]eople in an open society do not demand infallibility from their institutions, but it is difficult for them to accept what they are prohibited from observing.” While public access to court proceedings is indispensable to preserving this openness, access to court records is equally critical. In practice, much of the business of the courts—particularly in the federal system—takes place through written filings: motions, briefs, evidentiary submissions, and judicial opinions. Without access to these records, the public cannot understand and scrutinize the judiciary’s work.

Despite the longstanding presumption that court records are open for public inspection, parties frequently attempt to “seal” documents in order to hide them from public view. Empirical research reveals that the sealing of court records is extensive and increasing, particularly in federal courts and states with broad sealing statutes. Indeed, the sealing of court records is often routine, with minimal opposition or judicial scrutiny. A recent examination of sealing in federal district courts found that motions to seal were among the most common court filings, with over thirty thousand cases containing such motions in the past five years. One federal judge even compared the increasing secrecy in the courts to “kudzu,” a nearly uncontrollable creeping vine that “blocks access to sunlight, slowly strangling fields and forests in its wake.”

The harms that come from this secrecy are far from abstract. In a series of investigative reports titled “Hidden Injustice,” Reuters revealed how secrecy in U.S. courts—particularly the sealing of court records—shields corporate misconduct and conceals information vital to public safety. Their investigation discovered that federal judges had sealed evidence related to harmful products in about half of the 115 biggest product liability cases over the past twenty years and that in 85% of those instances judges provided no explanation for keeping the information from the public. Looking specifically at opioid litigation, Reuters determined that judges allowed litigants to file under seal evidence that could have alerted regulators, doctors, and the public to the dangers of prescription opioids; this concealment, they concluded, played a significant role in prolonging and deepening the opioid crisis.

Under current case law, as discussed in more detail below, the sealing of court records should be rare and permissible only under limited circumstances. In practice, however, the opposite is true: Sealing has become disturbingly routine in the federal courts, often carried out with little meaningful judicial oversight.

Critics of court secrecy point to the absence of clear, uniform rules on sealing as a key factor behind its widespread use in the federal courts. The Federal Rules of Civil Procedure and Criminal Procedure mandate the protection of certain sensitive information, namely Social Security numbers, tax identification numbers, dates of birth, financial account numbers, names of minors, and home addresses. Beyond this specific information, however, the federal rules offer little guidance on sealing to litigants and judges. As a result, when parties seek to seal records containing personal or other confidential information, the requirements and procedures are largely governed by the district court’s local rules.

This project is the first to comprehensively examine whether federal district courts provide clear and consistent local rules for sealing court records in civil and criminal cases. It also evaluates whether these local rules align with established Supreme Court and circuit court precedent concerning the public’s right of access to judicial records.

The analysis of more than seven hundred local rules across all ninety-four federal districts reveals alarming inconsistencies in sealing practices nationwide. Nearly half of all districts lack general sealing rules, many fail to reference the controlling legal standard, and basic procedural safeguards—such as public notice, consideration of alternatives to sealing, and case-specific identification of harms—are frequently absent. Overall, only eleven districts (12%) provide a specific standard for sealing that incorporates their circuit’s requirements for civil cases and a mere six districts (6%) do so for criminal cases. These deficiencies have created a patchwork of inconsistent, often-toothless rules that enable secrecy to spread largely unchecked.

Some might contend that local rules on sealing are unnecessary because the federal courts of appeals already provide sufficient guidance and district courts can be expected to follow those precedents. Empirical studies, however, do not support that assumption: In practice, most judges do not rigorously apply circuit law when evaluating motions to seal, often granting such requests with minimal scrutiny. Furthermore, even if district courts were meticulous in applying governing case law, clear and comprehensive rules would still offer significant benefits. Clear rules on sealing further consistency, transparency, and efficiency by establishing uniform procedures, reducing litigation costs, protecting the public’s interest in open records, and reinforcing confidence in the judiciary through predictable and principled decision-making.

This Article seeks to accomplish three principal objectives. First, it demonstrates that the lack of clear guidance on sealing is not confined to a few outlier districts but is a systemic issue across the federal judiciary. While some districts have adopted rules that provide meaningful direction on sealing, such examples are rare. In fact, forty-one of the ninety-four federal districts—approximately 44%—do not have a general sealing provision in either their civil or criminal rules laying out the procedures for sealing court records. The absence of such rules leaves litigants and judges without a framework for evaluating when and how court records may be sealed, increasing the risk of inconsistent and unprincipled sealing.

Second, this Article offers recommendations for improving the clarity and consistency of sealing rules in the federal courts. The goal is not to impose a uniform substantive standard—such as the First Amendment right of access—on all districts, though I have argued elsewhere that all court records should be subject to the more stringent First Amendment right. Rather, this project assesses each district’s rules in light of the access standard adopted by its respective circuit. Even in circuits that apply only the more-deferential common law standard, the public’s right of access would be better protected through clear, well-defined rules that ensure transparency and accountability in court decisions regarding sealing.

Third, it underscores the judiciary’s responsibility to safeguard the public’s right of access. By allowing the sealing of court records to become a routine, often-unexamined practice, many federal judges have abdicated this responsibility. This court-approved secrecy has, among other consequences, weakened the public trust that the judicial system relies upon. This is a critical time for the federal courts to address this problem. Public confidence in the courts—and in government more broadly—is declining. Strengthening the rules and practices that govern the sealing of court records is essential not only to restoring the legitimacy of the courts but also to reaffirming the judiciary’s role as a guardian of open government.

This Article proceeds in three parts. Part I describes the benefits of public access to court records. It then surveys the principal standards courts have adopted for evaluating requests to seal court records, including the First Amendment, common law, and “good cause” standards. Part II analyzes how district courts implement these sealing standards through their local rules. As Part II shows, local rules on sealing vary widely—not only from one district to another, but even among districts within the same state—creating a patchwork of standards and practices for sealing.

Part III concludes by offering suggestions for how to create clearer and more consistent rules on sealing. It also suggests practical measures to reduce the administrative burdens on courts. Many of the current rules governing sealing are not only confusing but also unwieldy, creating unnecessary work for judges and litigants alike.

In a previous book, the author sought to establish that the likely original meaning of the Fourteenth Amendment’s Privileges or Immunities Clause was that it guaranteed equality in fundamental rights under state law. The central provisions of the first section of the Amendment—due process of law, protection of the laws, and the privileges and immunities of citizenship—had long-established antebellum legal meanings. Due process of law primarily meant there had to be established law before one could be deprived of life, liberty, or property, and any violation of that established law had to be adjudicated according to known and established procedures. The protection of the laws was the other side of the coin: It was the legal protection the government had to extend against private invasions of private rights, principally judicial remedies and physical protection from violence. The protection of the laws was the heart of the social compact: Men exit the state of nature and give up some of their executive power and agree to obey the sovereign—they agree to give allegiance—in exchange for the sovereign’s protection against private violence and private invasions of rights.

The antebellum legal background is crucially important for understanding the original meaning of the Amendment’s privileges or immunities provision because it is that provision that therefore must accommodate the central goal of the Amendment’s drafters of constitutionalizing the Civil Rights Act of 1866. That act guaranteed equality in civil rights under state law. Although many Republicans believed the act was justified under the Thirteenth Amendment’s enforcement clause, many, including John Bingham, the principal author of the Fourteenth Amendment’s first section, believed the Act to be unconstitutional. Moreover, it was necessary to enshrine the civil rights principle in the Constitution itself, lest the Democrats take over and repeal that legislation; Congress needed, in the words of Representative and future President James Garfield, “to lift that great and good law above the reach of political strife, beyond the reach of the plots and machinations of any party, and fix it in the serene sky, in the eternal firmament of the Constitution.”

Yet, if the due process and protection of the laws clauses in the Amendment had their historical legal meanings, they would not accomplish that central objective. Due process does not guarantee equal rights, only that whatever rights one has will not be taken away without established law and known procedures. Nor does equal protection of the law guarantee equal rights, but rather guarantees only that whatever rights one possesses will be equally protected against, say, Ku Klux Klan violence. That leaves only the privileges or immunities provision, whose language does the necessary work. The Reconstruction generation understood that civil rights defined and regulated under state law, including contract and property rights, were fundamental rights that all free governments had to secure. They were, in other words, the “privileges and immunities of citizens of the United States,” shared by all citizens, though states may have regulated the rights differently. A state would “abridge” those rights by giving a lesser set of rights to a disfavored class.

To be sure, the language can also support the conventional, incorporation reading of the clause. The “privileges” and “immunities” of “citizens of the United States” at a minimum include those rights so fundamental that they were constitutionally enumerated, which a state can “abridge” just as Congress can “abridg[e]” the freedom of speech or of the press. Although the book did not tackle the affirmative evidence for incorporation of the Bill of Rights against the states, two other works did so.

One paper, Reversing Incorporation, sought to demonstrate that the antislavery and Republican concern, both before and after the adoption of the Fourteenth Amendment, was equality in civil rights however defined and regulated under state law. Although fundamental rights were routinely mentioned, abolitionists and Republicans relied on state constitutions, on the rights that freedom would bring under state law, and on Congress’s powers to insist on republican governments during readmission. There is very little evidence from the period that anyone understood the Fourteenth Amendment to guarantee a fundamental floor of rights.

A still more recent work, The Antislavery Reading of Article IV, addressed the claim that antislavery constitutionalists had an unorthodox reading of Article IV’s Privileges and Immunities Clause, by which that clause effectively nationalized the rights of citizens. The clause was conventionally understood to require a state to treat citizens of other states on equal terms with its own citizens. This prior work demonstrated that some antislavery theorists did hold unorthodox views of Article IV. The best understanding of the unorthodox view, however, was that it would have merely extended the antidiscrimination work of the clause to discrimination among a state’s own citizens.

Representative Lawrence of Ohio made this equality reading of Article IV explicit in connection with the civil rights bill. Lawrence asked whether the nation was powerless to intervene when a state denies rights to “whole classes of native or naturalized citizens.” He argued that Article IV, Section 2 authorized Congress to enforce “the equal civil rights which it recognizes or by implication affirms to exist among citizens of the same State.”

The aim of this essay, prepared for the symposium marking the 100th anniversary of Gitlow v. New York, the case famously presuming the freedom of speech to be applicable to the states via the Fourteenth Amendment, is to demonstrate that many of the relevant historical debates involved the freedom of speech. Yet these debates do not suggest that the historical actors thought the First Amendment right was “incorporated” against the states. Quite the opposite. The actors presumed that the freedom of speech was a fundamental right guaranteed in many different ways—by state constitutions, by Article IV, by the Republican Guarantee Clause, and, yes, by the First Amendment at least with respect to the federal government.

After the adoption of the Fourteenth Amendment, that right was, also, guaranteed in some way by the new Privileges or Immunities Clause. The question is in what sense. One possibility, as the above suggests, is that the clause requires nondiscrimination among a state’s own citizens with respect to that right. Consider how one treatise writer in 1871 explained that Article IV guaranteed the freedom of speech in a similar manner by prohibiting discrimination in the provision of this right against out-of-state citizens:

[The clause’s] intention is to secure the like privileges and immunities to all those American citizens …. What like privileges and immunities? Those specified and enumerated in the federal constitution; the enjoyment of life, liberty, property, and the pursuit to happiness, no matter where located or domiciled…. The states without [this clause], by their local legislation, might, and perhaps would, impose different restrictions on the residents of each other in their necessary trade and intercourse, or upon American citizens migrating from one state to another, with the view of becoming residents, militating against those unalienable rights ….

It is possible, of course, that the author believed Article IV incorporated the Bill of Rights against the states, contrary to Barron v. Baltimore. Yet it seems evident enough that the author supposed those rights would be protected in the same way that ordinary rights to life, liberty, and property would be. He does not appear to have been suggesting that such rights were identically defined and secured in all the states. His reading appears to have been conventional: however the state defines and secures its citizens’ fundamental rights—to contract, property, bear arms, freedom of speech, and so on—it must guarantee those same rights to out-of-state citizens.

Part II of this essay, relying on the prior work identified above, canvasses the relevant evidence relating to the freedom of speech. It demonstrates that the historical players generally assumed that the freedom of speech was protected in a variety of ways, but none compels the conclusion that any of them thought the freedom of speech as defined in the First Amendment was “incorporated” against the states. They suggest an equality reading: whatever speech regulations existed must extend equally to all citizens without arbitrary discrimination. At most, they suggest that the framers of the Fourteenth Amendment expected the Privileges or Immunities Clause to secure only those rights that all free governments had to. Whether, on either reading, a state would be prohibited from banning the sale of violent video games to minors, from regulating student speech, or from punishing flag burning, viewing of animal crush videos, protesting at a dead soldier’s funeral, or stealing valor, may well be doubted.

First Amendment disputes frequently require courts to assess governmental assertions that contested expression is unacceptably dangerous. This assessment requires courts to choose when to defer to the government’s assertions of danger—and when instead to distrust those assertions. The centenary of the Supreme Court’s decision in Gitlow v. New York invites us to revisit the role that deference has played, and could play, in Free Speech Clause theory and doctrine. And because a great deal of the First Amendment law developed since Gitlow is at least as much about suspicion of the government as it is about deference to the government, Gitlow‘s centenary also invites us to consider the role that suspicion has played, and could play, in First Amendment law.

Benjamin Gitlow was charged not “with the commission of any overt illegal act,” nor “with conspiracy to commit an illegal act,” nor “with advocating that anyone else go out and commit an overt illegal act.” Instead, he and his co-defendants “advocated ideas that, if enough people agreed with them, might lead to illegalities at some point in the future.” In defense, Gitlow argued that the First Amendment did not permit the government to punish this advocacy absent evidence of its causal connection to “some substantive evil, consummated, attempted or likely.” A majority of the Supreme Court disagreed, deferring to the legislature’s determination that this advocacy, without more, carried sufficient potential for harm to justify its regulation. But the Court failed to explain why it chose to defer to the legislature’s statutory conclusion that certain speech was dangerous by its very nature.

As we’ll see, a variety of principled reasons can support a court’s choice to defer to the government’s assessment of expression’s danger. So too can multiple principled reasons support a court’s choice to treat the government’s assessments instead with suspicion. As we’ll also see, courts sometimes explain their choice to be deferential, and sometimes they don’t—and courts sometimes explain their choice to be suspicious, and sometimes they don’t. But exposing and justifying these choices helps courts improve the quality of their threshold decision to defer or distrust, as well as the quality of their subsequent operationalization of that decision through the creation and application of free speech doctrine.

The centennial of Gitlow v. New York is upon us. Gitlow is typically praised as an essential step in the development of modern First Amendment doctrine, so that it is said that “[f]ew individual stars shine as brightly in the constellation of American civil liberties cases.” Yet, closely examined, Gitlow seems a puzzling choice for constitutional canonization.

Decided at a time when there were virtually no First Amendment protections for speech, Gitlow held that government could punish mere abstract advocacy of violent revolution. Over the dissenting votes of Oliver Wendell Holmes, Jr. and Louis Brandeis, Gitlow stood for the proposition that “a State in the exercise of its police power may punish those who abuse” freedom of speech “by utterances inimical to the public welfare, tending to incite to crime, disturb the public peace, or endanger the foundations of organized government and threaten its overthrow by unlawful means.” It would not be too much of an exaggeration to characterize Gitlow as “the clearest expression of the Supreme Court’s acceptance of seditious libel.”

Modern First Amendment doctrine, by contrast, is founded on the fundamental axiom that the state may not punish seditious libel. If we now celebrate Gitlow, therefore, it is certainly not for its articulation of the substance of First Amendment protections. It must rather be because Gitlow construed the liberty interests protected by the Due Process Clause of the Fourteenth Amendment, which apply as against the States, to include free speech protections analogous to those protected by the First Amendment, which apply as against the federal government. “The precise question presented, and the only question which we can consider under this writ of error,” Gitlow states, is “whether the statute, as construed and applied in this case, by the State courts, deprived the defendant of his liberty of expression in violation of the due process clause of the Fourteenth Amendment.”

We are now apt to interpret this language through the lens of incorporation doctrine, which conceives the Due Process Clause of the Fourteenth Amendment as transparently reproducing the exact doctrines of the First Amendment “jot-for-jot and case-for-case.” But any such conception of incorporation developed well after Gitlow, which understood itself instead to be explicating the specific nature of the liberty protected by the Fourteenth Amendment. The question in Gitlow was not what the First Amendment required; it was instead what was required by the liberty interests safeguarded by the Due Process Clause.

We should be clear that these same liberty interests also underlay Lochner v. New York. These same liberty interests fueled the Court’s use of the Due Process Clause to strangle social and economic legislation. In the decade before Gitlow, for example, the Court had invoked these same liberty interests to strike down state efforts to protect union members and to regulate the fees of employment agencies. Determined to resurrect and expand the reach of Lochner, the Taft Court during the 1920s was vilified as “the zenith of reaction.” Enlarging Fourteenth Amendment liberty to include freedom of speech threatened to enlarge the scope of judicial control over all state regulation.

Seen from this perspective, Gitlow can suddenly come to seem deeply enigmatic. We might ask why the Court bothered to extend free speech protections to states when its understanding of the substance of these protections was so stunted and feeble. What was actually at stake for the Court majority when it decided to interpret the liberty interests protected by the Due Process Clause of the Fourteenth Amendment to include freedom of speech?

Conversely, why did Holmes and Brandeis, who were strongly opposed to the Court’s resurgent Lochnerism, nevertheless join the Court in holding that the “general principle of free speech … must be taken to be included in the Fourteenth Amendment, in view of the scope that has been given to the word ‘liberty’ as there used.” How were Holmes and Brandeis able to interpret the word “liberty” in the Due Process Clause so that it might entail strict and independent judicial review of state censorship of speech, but nevertheless require deference to ordinary state social and economic regulations?

To unravel these questions, we must make a strenuous effort of historical imagination. We must put Gitlow back into the context of its time. The effort will prove rewarding. We shall learn, for example, that no Justice in the 1920s interpreted the word “liberty” in the Due Process Clause in anything like the manner of the contemporary Court in a case like Dobbs v. Jackson Women’s Health Organization. So far from viewing the scope of constitutionally protected liberty as a “fact” determined by the historical data of history and tradition, all Justices in the 1920s understood liberty interests to be defined by values they perceived as immanent in the Constitution.

For the majority of the Court, these values centered around protecting the constitutional “right of the citizen to be free in the enjoyment of all his faculties,” a libertarian constitutional vision first articulated at the start of the Lochner era in Allgeyer v. Louisiana. Both Holmes and Brandeis strongly objected to this constitutional vision, which they condemned as inconsistent with the proper role of Article III courts. In Gitlow, Holmes and Brandeis instead interpreted Fourteenth Amendment liberty in light of their commitment to the specific constitutional value of freedom of speech.

Because they so often joined each other’s opinions on this subject, we often imagine that Holmes and Brandeis supported freedom of speech for the same reasons. But the history of Gitlow illustrates that this was not the case. Whereas Holmes believed that freedom of speech was necessary to sustain the authority of positive law, Brandeis believed that freedom of speech was the lifeblood of a democracy whose purpose was to empower citizens to develop their own autonomous faculties. Freedom of speech was for Holmes a jurisprudential necessity, whereas for Brandeis it was “both … an end and … a means.” In contrast to Holmes, Brandeis was prepared to read democratically required liberties other than freedom of speech into the Due Process Clause.

Modern First Amendment doctrine has descended from Brandeis, not from Holmes. But Holmes’s brief dissent in Gitlow nevertheless contains an important lesson for contemporary constitutional theorists. Holmes was the founder and most analytically powerful practitioner of positivism in the history of American jurisprudence. Holmesian positivism has recently been revived by originalists and by those committed to interpreting the Constitution in light of a factually based account of history and tradition. Yet Holmes’s short opinion in Gitlow illustrates that even rigorous positivists cannot ultimately interpret the Due Process Clause solely in terms of facts, whether the facts of original public meaning or of history and tradition. An internally consistent positivism must instead find its ultimate ground in fundamental constitutional values.

Artificial general intelligence is “probably the greatest threat to the continued existence of humanity.” Or so claims OpenAI’s Chief Executive Officer Sam Altman. In a seeming paradox, OpenAI defines its mission as ensuring “that artificial general intelligence—AI systems that are generally smarter than humans—benefits all of humanity.”

Whether artificial general intelligence becomes a universal boon or an existential threat—or both—there is general agreement concerning its revolutionary potential. Indeed, Microsoft founder Bill Gates has called it “as fundamental an advance as the personal computer or the internet,” and Google CEO Sundar Pichai has predicted that it will “bring about a more profound shift in human life than electricity or Promethean fire.”

Thus far, AI systems are not generally smarter than humans. Not yet. Large Language Models (LLMs), however, are advancing at a startling pace. LLMs use artificial intelligence to synthesize massive amounts of textual data and then predict text and generate responses to users in “natural” human language. On a scale measuring the progress of LLMs toward general intelligence, OpenAI’s flagship model, GPT-3 scored zero percent in 2020. Just five years later, OpenAI’s o3-preview scored between 75% and 88%. Meanwhile, OpenAI competitors such as Anthropic, Google, and DeepSeek are likewise racing to deliver on the promise of “systems that can think and act rationally in ways that mirror human behavior and intelligence.”

Even as LLM models make progress toward general intelligence, there are already AI systems that have exceeded human performance on narrow, clearly scoped tasks. For example, chess engines have been performing at superhuman levels for years, and AI models can now help detect breast cancer far earlier than human experts—and the models continue to improve. Meanwhile, OpenAI’s o1 reasoning model has an LSAT score higher than the median student admitted to the law schools at Harvard, Yale, and Stanford.

As AI systems begin to mirror human thought, it pays to remember the words of Seneca: “[N]ot all mirrors stick to the truth.” LLMs now regularly create outputs that appear to be the product of independent thought, but LLMs are essentially prediction engines that “answer” prompts (or inputs) by calculating which words are most likely to come next and then assembling them into an output. LLMs, as such, do not predict truth but instead predict probabilities. In doing so, they sometimes replicate false information common in their training data.

They also inevitably produce “plausible yet false outputs,” commonly referred to as hallucinations. An LLM may produce fake legal documents, non-existent academic citations, or false biographical data. Although LLM producers and users can employ various tactics to reduce hallucinations, these errors cannot be eliminated. And they are quite prevalent. In fact, data gathered from multiple chatbots suggests that hallucinations occur in 3% to 10% of outputs, leading one legal scholar to dub LLMs “Large Libel Models.”

Radio host Mark Walters found out about hallucinations the hard way. Walters is the first person in the United States to sue an LLM producer for defamation. Walters alleged that OpenAI’s Large Language Model iteration operating at the time, GPT-3.5, falsely accused him of fraud and embezzlement. The accusation, which was unequivocally false and potentially reputation-damaging, arose in response to a prompt from a journalist named Frederick Riehl. Fortunately, Riehl had more than an inkling that the accusation was false, and he contacted Walters shortly after receiving it. No one else received the hallucination. Walters nonetheless sued OpenAI for its defamatory output.

The rise of defamation-by-hallucination cases presents significant challenges for defamation law. In most states and in most cases, defamation liability hinges on the defendant’s scienter, or mental state. As a matter of tort law, plaintiffs generally must establish that the defendant was at least negligent in publishing a defamatory, false, factual statement about them. As a matter of constitutional law, plaintiffs who are public officials or public figures must prove actual malice on the part of the defendant: In other words, such plaintiffs must prove that the defendant knew at the time of publishing that the defamatory statement was false or recklessly disregarded its falsity. To state the obvious, it is not possible to judge whether an LLM “knew” of or was reckless as to the falsity of an output prior to publishing it in response to a prompt. Nor is it possible to determine whether the LLM exercised reasonable care. Only human actors have scienter as to falsity, or even as to the act of publishing.

Generative AI is not the first technology to which defamation law has had to adapt. Laws addressing infliction of reputational harm predate the printing press, and each new technology of communication has forced the common law to reexamine how to protect reputation without unduly deterring valuable speech. In the United States, courts have long adapted tort doctrines to achieve what they perceive to be sound communications policy. This was true even before the Supreme Court intervened to balance state interests in protecting reputation against free speech concerns. Adapting defamation law to generative AI is simply the latest challenge in defamation law’s long and ongoing evolution.

Nonetheless, the stakes are high. If defamation law imposes liability on AI companies for every defamatory hallucination, the potential of LLMs as a tool for discovery, data analysis, information gathering, and creativity enhancement may be squandered. In fact, it is easy to imagine a new form of barratry emerging, in which lawyers and others search name after name to generate endless numbers of new clients. The costs of defending against claims and the threat of massive, unpredictable monetary awards may cause developers of Large Language Models to implement excessive filtering and content moderation, shift costs to end users, and reduce accessibility. In addressing defamation by hallucination, therefore, legal decision-makers must balance the desire to impose accountability for reputational harm with the need to foster the development of an important and perhaps revolutionary medium of information gathering and generation.

This article proposes a two-pronged approach to address harms caused by the newest AI reasoning models. First, this article introduces a framework for applying centuries-old tort doctrines to these models. Rooted in tort law rather than the First Amendment, this framework takes as a given that AI “speech” is valuable to recipients. and that public access to powerful, new information-gathering tools is essential to sound communications policy. Second, recognizing that tort law alone cannot adequately address the reputational harms the newest models pose, this article argues for statutory reforms focused on reducing such harms and making them easier to prove when they do occur.

Part I examines the application of existing defamation doctrines—both common law and constitutional—to “foundation-model” generative AI speech, highlighting the points at which existing doctrines may be deployed to accommodate communications policy concerns. This Part examines these doctrines through the prism of the trial court’s order in Walters v. OpenAI, which addressed (1) whether a hallucination generated by an LLM in response to a prompt is a factual assertion when the prompter has serious reason to doubt the veracity of the output; (2) how to apply defamation’s scienter requirements when the speech in question is generated by an LLM; and (3) whether a plaintiff can recover presumed damages when the evidence indicates no damage actually occurred. This Part also explores a fourth doctrine unexplored in the Walters order: namely, whether an LLM is a publisher or a distributor of the content it produces.

Part II highlights the challenges and opportunities posed by the newest LLM reasoning models, particularly “chain of thought” models. Reasoning models can “understand” concepts, contexts, and ambiguity in ways that their predecessors could not, which makes them more capable than their predecessors of solving “complex reasoning and problem-solving tasks.” Chain-of-thought models, in particular, promise to deliver greater accuracy (i.e., more correct results) as their final outputs. Even as these models produce more reliable final outputs, however, new avenues for defamatory outputs arise in their intermediate “chains of thought.” Moreover, even those reasoning models that “show their sources” continue to produce hallucinated outputs. Relying on the latest computer-science research on hallucinations, and particularly on a hallucination “taxonomy” developed by Yale researchers, this Part demonstrates that hallucinations are inevitable and, in some instances, surprisingly valuable. This fact has significant implications for those seeking to “regulate” hallucinations by way of defamation law.

Based on insights from Part II, Part III identifies the chief policy considerations that should underpin the adaptation of defamation law to this new technology. This Part begins to sketch how defamation law can balance the protection of individual reputation with the need to accommodate those hallucinations that are inextricably linked to LLMs’ generative capacities. Specially, this Part contends that legal rules in this area must attempt to incentivize compensation for provable reputational harm, correction of the digital record, model transparency, innovations for safety and accuracy, the exercise of reasonable care at all stages of AI development, and the exercise of reasonable care by AI users.

Part IV proposes a legal framework for addressing defamatory hallucinations. This framework rests on the proposition that the law should treat defamatory hallucinations—both in an LLM’s final output and in its chains of thought—as “inevitable errors,” in much the same way that the Supreme Court in New York Times v. Sullivan recognized that some journalistic errors must be tolerated in order to produce “uninhibited, robust and wide-open” public discourse.

Part IV first explores some of the practical and technological limitations of prior scholarly proposals addressing defamation by hallucination in “foundation model” LLMs. Next, it draws lessons from defamation law’s adaptation to past technologies, such as the telegraph, to argue that LLMs bear more similarities to information distributors than to publishers, and AI developers’ liability for defamation should be governed accordingly. However, any tort law “privilege” accorded to AI developers must be balanced by the imposition of statutory duties designed to protect victims of defamation by hallucination. Specifically, LLM producers should have a duty to warn (and educate) their users that it is unreasonable to rely on AI outputs without verification. Furthermore, LLM producers should have a duty to preserve search records for a limited period of time in order to enable defamation plaintiffs to prove the existence and scope of their harm. Finally, this Part contends that users who repeat defamatory hallucinations without taking reasonable steps to verify them should often face liability, because such users will often be the cheapest cost avoiders in the realm of reputational harm. We conclude with caveats about the limits of defamation law as a regulatory tool for dealing with hallucination-based harms.

Over the past decade, the Roberts Court has sought to disrupt two major domains of the First Amendment: the Religion Clauses and free speech. These interests have recently merged to yield a flurry of cases raising complex questions at the intersection of free speech and religious liberty. This Article argues that the Court’s emerging approach to such cases threatens to unravel longstanding free-speech doctrine and the core values underlying it.

These dangers are on full display in the Court’s analysis of a recent case addressing the constitutional quandary posed by the religious speech of public employees. Kennedy v. Bremerton School District involved Joseph Kennedy, a high-school football coach and devout Christian who, after each game, knelt in prayer at midfield, joined by players, adult community members, and the media. After repeatedly requesting that Kennedy refrain from this so-called “demonstrative prayer,” Bremerton School District placed Kennedy on administrative leave due to its concerns about the consequences of his behavior, including the difficulty of ensuring security at the games and the risk that the District would be violating the Establishment Clause by allowing Kennedy to continue. Kennedy refused to reapply for his coaching job and alleged that the District had violated his free-speech and free-exercise rights.

The Supreme Court has long recognized that public employees like Kennedy enjoy some degree of free-speech protection. In recognizing this qualified protection, the Court seeks to strike a careful balance. On one hand, employee-speech doctrine vindicates public employees’ free-speech rights. On the other, it aspires to vest in school districts, government agencies, and other public institutions the leeway to manage themselves—and their workforces—effectively. To negotiate this fundamental tension, for public-employee speech, the Court has eschewed the stringent review typical of other areas of free-speech doctrine in favor of a more flexible balancing test: When a public employee speaks as a citizen on a matter of public concern, the Court balances the “interests of the [employee] … in commenting upon matters of public concern” against “the interests of the State … in promoting the efficiency of the public services it performs through its employees.” However, when an employee speaks as part of her public employment, the employee is owed no free-speech protections at all because it is, in effect, the government—not the employee—speaking.

Kennedy appreciated little of this fragile détente. Taking up both Kennedy’s free-speech and free-exercise claims, the Court granted certiorari on the questions of “whether a public-school employee who says a brief, quiet prayer by himself while at school and visible to students is engaged in government speech that lacks any First Amendment protection” and “whether, assuming that such religious expression is private and protected by the Free Speech and Free Exercise Clauses, the Establishment Clause nevertheless compels public schools to prohibit it.” Justice Gorsuch authored an opinion for a six-Justice majority holding that the District’s actions violated Kennedy’s free-speech and free-exercise rights, and that the District’s Establishment Clause interest failed to save its otherwise unconstitutional action.

Rather than evaluate Coach Kennedy’s claims on their own merit and according to the doctrine applicable to each, Justice Gorsuch flattened the claims into a zero-sum, culture-war battle over religious liberty. For Gorsuch, Kennedy was no tough case. It was, boiled down, a “government entity [seeking] to punish an individual for engaging in a brief, quiet, personal religious observance doubly protected by the Free Exercise and Free Speech Clauses of the First Amendment.” As Gorsuch saw it, in disciplining Coach Kennedy, the school district had flouted the principle that “[r]espect for religious expressions is indispensable to life in a free and diverse Republic—whether those expressions take place in a sanctuary or on a field, and whether they manifest through the spoken word or a bowed head.” Religious expression is religious expression, Gorsuch told us—”form” and “context” be damned.

And yet, Gorsuch’s creative deviations in Kennedy notwithstanding, the Court’s well-established precedents provide a relatively tidy doctrinal framework for each of Kennedy’s claims. Neither of those frameworks prescribes the analysis Gorsuch performs in Kennedy.

On the free-speech front, the Court has held that public employees receive free-speech protections only when speaking “as citizens.” Public employees who instead speak “pursuant to their official duties” speak not as citizens but as employees and are not “insulate[d] … from employer discipline” at all. Further, a public employee may receive protection only for “speech on a matter of public concern”—not for speech on “private matters.” This distinction “must be determined by the content, form, and context of a given statement, as revealed by the whole record.” Fifty years of precedent can thus be synthesized into a (deceptively) straightforward rule: A public employer’s disciplinary actions trigger the First Amendment only when an employee speaks (1) as a citizen (2) on a matter of public concern. If both conditions are satisfied, the Court then conducts “particularized balancing,” weighing the employee’s particular speech act against the government’s particular interests in regulating it.

The free-exercise framework is similarly streamlined. If a rule or action that burdens free exercise is neutral and generally applicable, it is subject only to rational-basis review. However, if the rule or action is either not neutral or not generally applicable, it is subject to strict scrutiny and likely fails. A government policy is not neutral if it is “specifically directed at … religious practice,” is “discriminat[ory] on its face,” or otherwise has “religious exercise” as its “object.” And a government policy fails the “generally applicable” requirement if the state allows for individualized exemptions from the policy but denies a religious exemption, or exempts comparable secular conduct.

Perhaps blinded by a reflexive desire to defend religious speech, the Kennedy majority glossed over the many tensions latent in the First Amendment’s clauses, opting instead to focus only on their supposed synergies. Justice Gorsuch declared that free speech and free exercise “work in tandem” to “provide[] overlapping protection for expressive religious activities.” But from that alluringly simple proposition flow two unavoidable problems. The first bears on the scope of this supposed “overlapping protection” in the realm of public employment: How far does the public employee’s bundle of First Amendment protections extend before it crashes into the constitutional bar on government establishment of religion? Justice Sotomayor’s dissent and a growing body of literature respond forcefully to this question. Largely missing from that discussion, though, is an acknowledgment and exploration of Kennedy‘s incompatibility with, and devastating consequences for, free-speech doctrine.

This Article seeks to tease out this second conflict. For all its high-soaring rhetoric about the Free Speech and Religion Clauses operating in tandem, Kennedy simply does not fit within the Court’s existing free-speech doctrine. This Article illustrates how Justice Gorsuch’s blurry conception of Coach Kennedy’s expression runs afoul of the Court’s own employee-speech doctrine and threatens to distort it into an inactionable muddle.

In the wake of the mega-million-dollar settlement of U.S. Dominion’s defamation action against Fox News over the network’s broadcast of false election fraud claims after the 2020 U.S. presidential election, shareholder derivative actions were brought in Delaware against the parent company Fox Corporation’s board of directors for breach of fiduciary oversight duties under state corporate law. The shareholder plaintiffs claimed that the Fox Corporation board breached its fiduciary duties by allowing Fox News knowingly to air false programming that put the company at risk of massive defamation liability. The Delaware Chancery Court denied Fox Corp.’s motion to dismiss the action for lack of standing, so the derivative action is currently pending.

But should corporate fiduciary duty law be interpreted to impose liability on the boards of companies that own news outlets for failing to control defamation and other speech tort risks associated with the editorial judgments made by their news subsidiaries? What makes the In re Fox Corporation Derivative Litigation (hereinafter “In re Fox“) significant beyond its specific facts is that the plaintiffs’ rationales seek to expand and supercharge the traditional oversight requirements of corporate law. If accepted, this turn to strengthening the disciplinary power of corporate governance in the news media context is likely to undermine press functions and the public interest in a free and independent press.

The expansive interpretations of corporate governance principles advanced in In re Fox could attract support on the basis that corporate oversight duties can serve to minimize misinformation in political discourse. Surveys reveal that many Americans see political misinformation as a social threat. If using corporate law to combat misinformation could lead to robust censorship effects on falsity, then many could consider this a significant public benefit. This could incentivize additional lawsuits against the press.

At the same time, such a development is likely to undermine press activity in ways harmful to public discourse. If these kinds of corporate governance claims are successful, they promise to generate a regulatory regime of editorial control by risk-averse corporate boards with much broader business interests than the protection of press freedom. The possibility of multi-million-dollar personal liability for parent company board members—or at least corporate insurers—is likely to generate excessive board-level micromanagement.

It is reasonable to expect that this would lead directly to journalistic self-censorship by news subsidiaries, deter journalism discouraged by a press-hostile government, and worsen journalistic timidity in covering the powerful and litigious. The self-regulatory compliance and oversight systems likely to be implemented in media companies as a response to heightened governance liability will inevitably extend to coverage of matters beyond clearly false information.

Enhanced board obligations may also lead to uneven effects. If the most likely plaintiffs in defamation actions continue to be the politically powerful, wealthy, or socially notable, parent company boards worried about follow-on oversight lawsuits might feel disproportionate pressure to reduce critical coverage of such elites. Society loses when the powerful are not held to account. Moreover, heightened compliance requirements could provide cover for targeted and politicized efforts by board members to influence the content of their news units. Such results would all be dangerous for the press function and, ironically, for the same public discourse that anti-misinformation initiatives seek to improve.

Proponents of expanded oversight doctrine may attempt to dispute these predictions of a chilling effect on journalism by noting that damages payouts in successful shareholder derivative actions go to the corporate treasury. So if a derivative action based on the company’s prior payments to defamation plaintiffs is successful, the recovery may in fact offset the company’s defamation payouts by recouping the money from the culpable directors themselves.

But such theoretically reallocated liability cannot in fact be expected to temper either the corporate costs of expanded oversight litigations or the expected chilling effect on news companies’ journalist functions. If the Fox plaintiffs’ arguments to change corporate oversight doctrine are successful, the true costs are likely to be extensive. When oversight compliance requirements are effectively dictated by corporate insurers with little or no commitment to journalism, intrusive oversight into and second-guessing of the editorial process is practically guaranteed. Even if this would lead to desirable results for the most extreme cases, the consequences of overzealous compliance are likely to be overbroad and troubling for the public interest.

The functions of an independent press are democratically necessary and already subject to excessive economic, social, and governmental pressure (including legally aggressive lawsuits against FCC-regulated broadcast outlets by a sitting President). Adding even more pressure is bad policy. In light of the sustained recent attacks on constitutional press protections in defamation cases, the limits to other newsgathering protections, and press-skeptical courts and juries, the press is already in a particularly vulnerable spot legally. Recent settlements of lawsuits against CBS and ABC brought by President Trump trigger suspicions that the executive branch is not only demanding but also obtaining exceptional capitulation from conglomerate-owned press entities.

The anti-misinformation frame implicit in In re Fox thus offers an opportunity to address key questions about what types of trade-offs we should accept between two of our foundational social commitments—to the democratic value of the independent press and the democratic value of truthful political discourse. Because the deterrent effects on misinformation of expanding corporate oversight duties to this context are unclear and the negative consequences for the press are predictable, the likely effects of expanding corporate fiduciary liability to parent corporations vis-à-vis the coverage decisions of their news media organizations should be resisted—even by those who deplore Fox News’ 2020 election coverage. Ultimately, the Essay argues that courts should be reluctant to impose oversight liability in the news company context where executives or boards of directors did not actively direct clearly illegal conduct.

The Essay does not advance a doctrinal First Amendment argument. Nor does it request special and disproportionate exceptions or advantages for the press. It is, rather, a plea that before courts decide to advance anti-misinformation efforts by expanding ordinary corporate law principles to reach oversight of defamation risk in journalistic contexts, as proposed in In re Fox, they consider the potential impact of such an expansion on the ability of press organizations to perform their critical democratic functions.

To be sure, media owners are free to engage in intrusive oversight voluntarily. Nevertheless, the Essay argues that the effects of adopting a legal requirement are likely to lead to accelerated and industry-wide owner oversight over editorial decisions than is reported today. This poses a clear threat to journalistic independence. And since such intrusions are also unlikely to be open and transparent to those outside the organization in many instances, they could well obscure independent assessment of the degree of owner constraint on the outlet’s reporting.

The Essay proceeds as follows: Part I.A describes In re Fox, the Delaware Chancery Court’s denial of the defense’s motion to dismiss the suit for demand futility, and subsequent developments. In so doing, it provides a “mini-overview” on shareholder derivative suits to set the context and clarify the procedural posture of the case for the unfamiliar. Part I.B examines the In re Fox litigation through an anti-misinformation lens. Part II.A sketches board oversight duties under current Delaware corporate law. Part II.B unpacks the expanded board monitoring duties sought by the plaintiffs in In re Fox. Part III explores our dual—and here conflicting—social commitments to press editorial freedom and truthful political dialogue. Part III.A takes the first step by showing how the plaintiffs’ theories of liability in In re Fox do not justify expansion of current doctrine. Part III.B then addresses the dangers of expanded monitoring obligations to press functions—particularly since many news outlets are owned by other entities and since the current politico-legal environment amplifies the vulnerability of the press. Part III.C argues that the anti-misinformation benefits of the doctrinal expansion sought in In re Fox are at best uncertain and likely outweighed by the predictable chilling effects of expanded corporate law oversight duties on press functions. While recognizing the limits of its suggestions, Part III.D ends with some thoughts on other ways to promote press accountability.

Free speech absolutism has seemingly had its day, inasmuch as scholarly and public (if not judicial) opinion are increasingly hostile to the idea that more speech is invariably better. Yet, within academia, a close cousin of free speech absolutism—a principle of university management that this Article calls extramural absolutism—remains alive and well. Many academics and their supporters still champion the idea behind this principle, which is that speech undertaken by professors who are acting in a personal capacity should invite few if any adverse employment consequences. Academic disagreement with the principle, when it arises, lies more in its application to specific incidents than with the principle itself.

But extramural absolutism is deeply unpopular outside academia as well as among many stakeholders inside the university community. This Article explains extramural absolutism’s unpopularity and proposes a way forward. The problem, I argue, is not with extramural absolutism itself but with how it is presented and justified. The answer, moreover, is not to insist on distinguishing between speech that warrants protection and speech that does not: It is to acknowledge that academic labor and the way it is managed make a policy of extramural absolutism the only feasible approach.

Supporters of extramural absolutism have failed to be convincing for two reasons. First, they often imply that speech by academics merits an exceptional level of protection because academic speech is uniquely valuable to society. In this respect, they are like supporters of academic freedom writ large. “[T]he pursuit of truth without interference,” argue William Tierney and Vicente Lechuga in their defense of academic freedom, “is in the best interest of society.” Similarly, Matthew Finkin and Robert Post affirm the AAUP belief that universities are “instruments of the common good” and that the “roots of academic freedom” have “internal connections to emerging needs for knowledge and intellectual mastery.”

These lines of commentary presume a shared belief that academic speech is uniquely valuable because expert knowledge and expert pedagogy are better than their opposites. That is, academic speech deserves special protection because it contributes more to societal well-being than do other types of speech or speakers: Academic speech informs, instructs, challenges, and clarifies in addition to expressing opinion.

But, as Keith Whittington observes, extramural academic speech rarely contributes so much to intellectual progress or societal well-being—yet “failing to protect such speech might well hamper the kind of advancements in human knowledge that we most care about.” This Article offers an argument grounded in labor realities that complements Whittington’s argument, which is keyed to academic freedom. By drilling down into the realities of academic training and job performance, I show that identifying the boundaries of individual expertise is a surprisingly difficult task. Consequently, a connection to expertise cannot be used to distinguish between deserving and undeserving extramural speech in a way that is conceptually coherent.

Supporters of extramural absolutism also underestimate the magnitude of the workplace exception they are claiming. Extramural absolutism asks us to remove employment consequences from all speech by some speakers, with only very few exceptions sounding in generally applicable law. (Extramural absolutism could not immunize professors from any legal consequences because it is a principle of organizational management rather than a legal claim.)

But even if a policy of extramural absolutism would not immunize faculty from civil claims or criminal prosecution, it would protect them from negative employment consequences arising from all other types of expressive activity. This makes extramural absolutism a momentous deviation from general employment practices and the ultimate claim to academic exceptionalism, but supporters rarely acknowledge it as such. Part II draws on existing case law regarding extramural expression by employees working in public and private contexts outside academia to show why a policy of extramural absolutism vastly exceeds the legal rights and organizational practices experienced by workers elsewhere.

Despite the magnitude of the exception it represents and many shortcomings in how it has been articulated and defended, this Article argues that extramural absolutism is both reasonable and necessary as a principle of organizational management. And, somewhat ironically, the best defense of extramural absolutism also rests on an appeal to academic exceptionalism. As I show, academia is unique—but in its labor dynamics, not in the intrinsic value of its practitioners’ speech. What academics do (and are expected to do) and how academics work (and are expected to work) makes it impossible to engage in any principled boundary-drawing regarding extramural speech, including the boundaries of the category itself. Reframing the debate in terms of employment rather than expression allows supporters of extramural absolutism to justify unrivaled speech protections without resorting to unappealing and counterproductive elitism.

Part I of this Article surveys selected recent incidents where the extramural speech of tenured professors has triggered adverse employment consequences. I show that both the professors in question and their supporters consistently appeal to a principle best understood as “extramural absolutism.”

Part II contextualizes extramural absolutism within the broader landscape of employee speech rights to show that extramural absolutism is indeed a stark deviation from standard employment practice. Neither nonacademic employees in the private workforce nor public employees (including in academics) enjoy legal protections approaching the institutional protection afforded under a policy of extramural absolutism.

Part III shows why the exceptional protection required by extramural absolutism is necessary through a granular analysis of academic labor practices and constraints. I draw on social science and higher education scholarship to show that academic training and working conditions render an absolutist approach necessary. It is impossible to distinguish between types of extramural speech such that we can protect some extramural remarks but not all.

Finally, Part IV considers and refutes a few of the most common objections to extramural absolutism. Most important among these are the worries that extramural absolutism will open the floodgates for bad-acting professors by allowing them to easily evade employer discipline (IV.C) despite those professors’ articulation of opinions that exhibit manifest unfitness (IV.B). To preview the argument: These instances are neither so straightforward nor so common as the news cycle suggests and, consequently, they should not drive our approach to university management.

Ultimately, this Article finds common ground with both critics and supporters of extramural absolutism. Alongside critics, I argue that extramural absolutism is indeed a singular exception from standard practices and laws regarding employee speech rights. I further agree with critics that the inherent value of speech by academics cannot, by itself, justify this exception.

Nevertheless, alongside supporters, I argue that extramural absolutism is an exception that is made unavoidable by dynamics, constraints, and expectations that are, in fact, peculiar to academia. For better and for worse, the structure and purpose of academia depends on an absolutist approach to extramural speech.

In the United States, as is widely known, "hate speech" is generally protected by the First Amendment. Hate speech is considered "free speech" unless it provokes imminent violence or constitutes a "true threat" or "fighting words." No other nation protects the right to express hate so vigorously. Hate speech laws exist in most other countries, where the principles of free speech are said to have no bearing on the expression of racial, ethnic, or religious hatred.

Why are there no hate speech laws in America? There are many possible explanations. Some have suggested that the United States diverged from the rest of the world on hate speech regulation because of deeply ingrained national traits and tendencies, such as Americans' historic fear of government regulation and our individualistic culture. In a book manuscript in progress, I argue that the course that America took on hate speech was not foreordained but was rather the result of contingency and circumstance. Hate speech laws existed in many jurisdictions before the 1950s, and there was a good deal of popular support for hate speech laws.

The reasons why hate speech laws ultimately failed to take root in America are complex. Timing was an important factor. The onset of McCarthyism in the 1950s undercut campaigns during the previous decade to advocate for hate speech laws. Efforts to enact hate speech laws arose during the period of the development of the modern First Amendment, between 1930 and 1960. Hate speech regulations clashed with emerging civil libertarian free speech principles, and free speech ultimately prevailed.

This essay focuses on another significant reason why hate speech laws may have failed to take root in America—the vigorous opposition of minority civil rights organizations to hate speech laws at the time when those laws had their greatest potential for adoption. Minority civil rights groups such as the National Association for the Advancement of Colored People (NAACP) and the American Jewish Committee opposed hate speech laws on the theory that such laws were ineffective in curtailing hate speech, and that any limitations on freedom of speech would hinder minorities' efforts to achieve racial and religious equality.

At a time when lynchings and cross-burnings were rampant, and when American fascist demagogues and neo-Nazis routinely terrorized minorities, the leaders of these civil rights groups thought hard about the hate speech problem. They concluded that the most effective way to reduce racial and religious hatred was to combat hate speech through counter-speech and education, rather than through legal restrictions on hate speech. The opposition of these civil rights groups to hate speech laws changed American law and public policy. Notably, their arguments influenced the Supreme Court in the 1960s, which created an expansive, civil libertarian free speech jurisprudence that was intended, in significant part, to protect the civil rights movement.

This essay tells the story of how and why American civil rights organizations opposed hate speech laws for much of the twentieth century. Civil rights groups like the NAACP could have sought laws banning hate speech, just as they fought for the desegregation of public facilities. Instead, eminent civil rights leaders—including Thurgood Marshall, W.E.B. Du Bois, and Louis Marshall, among others—rejected hate speech laws as incompatible with the pursuit of equality and civil rights.

This essay discusses significant episodes in the twentieth century when civil rights groups opposed hate speech laws, with important consequences for free speech law and public policy. Part I narrates the NAACP's campaign to have the film The Birth of a Nation censored, and how the failure of that campaign convinced the national NAACP of the ineffectiveness of legal regulations on hate speech. Part II describes Jewish civil rights organizations' rejection of hate speech laws in the 1920s and '30s, and how those groups developed methods to stop antisemitic attacks through means other than legal restriction. Part III explores black and Jewish civil rights groups' opposition to proposed hate speech laws in the 1940s, and their creation of a "quarantine" or "silent treatment" policy as an alternative to legal suppression of hate speech.

Part IV explains how the civil rights movement joined with the American Civil Liberties Union in the 1960s to defend the free speech of white supremacists, resulting in landmark First Amendment precedents. Part V details how and why some civil rights groups changed their positions on hate speech starting in the 1970s, a shift that was most visibly demonstrated in litigation surrounding neo-Nazi attempts to march in Skokie, Illinois. The conclusion emphasizes the wisdom of the civil rights groups' earlier stance opposing restrictions on hate speech.

Much American copyright rhetoric vaunts technological progress and econom­ic incentives. One reading of the constitutional copyright clause characterizes copyright as a necessary (if unappealing) encouragement to the advancement of innovation. These emphases tend to obscure the centrality of human creativity to copyright law and theory.

In this article, provocatively titled "Humanist Copyright," I develop a counter-narrative. I seek to highlight the role of human authorship in the copyright scheme. The title references not only current debates over AI-generated outputs but also the proposition that authors' rights embody and advance human achievement. Copyright celebrates human creativity, for multiple reasons, economic and social, but also grounded in the person of the author. I trace these concepts to Italian Renaissance humanism and the emergence of the author as entrepreneur.

My exploration of the role of authorship proceeds in three parts: historical, doctrinal, and predictive. First, I will review the development of author-focused property rights in the pre-copyright regimes of printing privileges and early Anglo-American copyright law through the 1909 U.S. Copyright Act. Second, I will analyze the extent to which the present U.S. copyright law does (and does not) honor human authorship. Finally, I will consider the potential responses of copyright law to the claims of proprietary rights in AI-generated outputs. I will explain why the humanist orientation of U.S. copyright law validates the position of the Copyright Office and the courts that the output of an AI system will not be a "work of authorship" unless human participation has determinatively caused the creation of the output.

The phrase "humanist copyright" nods to Italian Renaissance philosophers such as Giovanni Pico della Mirandola, whose 1485 Oration on the Dignity of Man emphasized human autonomy in a human-centered universe. Pico declared that "we have been born into this condition of being what we choose to be"; man stands "at the very center of the world … as the free and proud shaper of [his] own being, to fashion [him]self in the form [he] may prefer." While Pico emphasized self-determination in shaping individual lives, the kinship between the authorship of one's being and the authorship of works of art and literature is apparent. Concepts of creative autonomy took root and flowered in 16th-century Italy, as Giorgio Vasari's Lives of the Artists attests. My counter-account of copyright thus begins in 16th-century Rome and focuses on one protagonist in the development of authorial rights.

And this is also a reminder that we publish material related to free speech generally, not just the First Amendment. This includes many kinds of articles related to copyright law, which "the Framers intended … to be the engine of free expression" (and which is at the same time a speech restriction), as well as articles about trade secret law, trademark law, common-law rules related to libel, state statutory protections for speech, and a wide range of other speech restrictions and protections.

The Press Clause: The Forgotten First Amendment, a Report from the Floyd Abrams Institute for Freedom of Expression, is a powerful argument for a broader understanding of the Free Press Clause. Much of its analysis will, I expect, prove important and useful to judges, lawyers, legal academics, and citizens. But one of its core premises—that the Free Press Clause should be read as conferring extra rights on the institutional press, beyond those possessed by others who speak to the public—strikes me as mistaken.

The Court's current precedents take the view that the First Amendment secures an equal right of everyone to use mass communications technology. These precedents generally do not offer special First Amendment rights to "the press" in the sense of a particular set of businesses or institutions. Rather, they protect the freedom of all to use "the press" in the sense of the printing press and its modern technological descendants. And this is also the approach taken by the great bulk of authorities from before the Framing through the 1800s and 1900s to today.

Under this model, the Press Clause is far from "forgotten" or stripped of "independent meaning or impact": It secures the critically important right of all people to use the means of mass communications. By itself, the Speech Clause could easily have been understood as just protecting "speech" in the longstanding historical sense of face-to-face oral expression. Indeed, in the 1600s and 1700s many governments deliberately tried to constrain printing presses on the theory that mass communication via the printing press was more dangerous than face-to-face oral communication and thus needed to be specially suppressed.

The Press Clause made clear that the use of mass communication technology (originally just the printing press) should be as protected as the use of one's voice. This understanding has ensured that all mass communicators—institutional media as well as others—are constitutionally protected. To the extent that today courts often use "speech" as shorthand for speech and press (and petition), that is a product of the vigor of the Press Clause, not a sign that the Clause has been forgotten.

And, I argue below, the sources cited in the Report's originalist, traditionalist, precedential, and structural arguments do not support special First Amendment treatment for the institutional media. Instead, many of the sources the Report cites actually support the thesis that the right belongs to all who sought to communicate to the public.

You can also see the earlier Freedom for the Press as an Industry, or for the Press as a Technology?—From the Framing to Today, 160 U. Pa. L. Rev. 459 (2012); the Abrams et al. Report in some measure responds to that, and Prof. Schafer's article, true to its name, does as well.

Over the past decade, nondisclosure agreements intended to protect the reputation of one or both parties from embarrassing disclosures have failed, sometimes spectacularly. Among those humiliated by revelations of their past behaviors are Donald Trump, Vince McMahon, Neil Gaiman, and Harvey Weinstein; among those companies and institutions affected are the Catholic Church, the Miss USA Pageant, and some of the largest tech companies. These agreements' failures have occurred when one party to the agreement gave an interview, passed along leaked documents, or filed a public lawsuit that became the basis of public reporting about the secrets that the contract concerned, as well as about the contract itself. Commentators and critics have offered various prescriptions to stem the enforcement of the most egregious contracts intended to prevent a party from reporting sexual assault and harassment that violates criminal or civil law. In a recent article, I described the specific doctrinal, social, and informational dynamics that both render such contracts vulnerable to breach and frustrate their enforcement.

This Article considers the potential liability of third parties that assist or spur the breach of such contracts. Journalists have most frequently played key roles in these disclosures by contacting a party and encouraging them to reveal their secrets. Friends and family members have also initiated or encouraged breach, as have attorneys and activists who hope to reveal a party's wrongdoing. Such non-parties serve as "breach agents," outsiders to the contract who advise, encourage, or help one of the parties disclose the information they had agreed to keep secret. They play an integral role in breaking the dams that keep information private—whether, viewed sympathetically, to advance the public good by bringing egregious behavior to light or, viewed critically from the perspective of the nonbreaching party, to violate a legally enforceable promise and undermine contractual stability.

As non-parties to the contract, breach agents are vulnerable to suit under the common law tort of interference with performance of a contract. The relationships between that tort and contract law—and between the tort and the First Amendment rights of journalists and others to gather news, to speak, and to publish—last inspired widespread consideration and commentary in 1996, when the tobacco company Brown & Williamson attempted to use a non-disclosure clause in an employment contract to silence a former executive who had been interviewed by the CBS television news show 60 Minutes.

Two developments since that time have made this issue newly relevant for attorneys and legal academics. The first is the Supreme Court's decision in Bartnicki v. Vopper (2001), which concerned a suit against third parties who disclosed confidential information that another party had illegally obtained. Bartnicki simultaneously established a test that in most factual scenarios will protect the press or a source which did not itself obtain the information illegally, and also refused to grant a blanket constitutional protection for reporting truthful information that would apply no matter how a source had obtained the information.

Second, individuals and institutions have increasingly relied upon NDAs outside of the employment context to protect personal and corporate reputations, even after the recent wave of breached contracts and the resulting embarrassing disclosures. Occasionally, the parties litigate in the aftermath of a breach; for example, one state court recently extended constitutional protection to The New York Times, which had used its successful persuasion of a party to an NDA to breach as the basis for a front-page story about President Donald Trump's finances. Given both the supply of NDAs and the high demand for their breach among the press and the public alike, more such litigation is likely to arise.

This Article describes the robust protection that the First Amendment offers third parties from tort claims, as well as the situations that mark such protection's likely limits. Parts I and II describe, in turn, reputational NDAs and the breach agents who play key roles in encouraging disclosure of the secrets bound by such agreements. Part III explains the tortious interference with contract doctrine on which parties to the contract can rely to seek recovery against breach agents. Part IV discusses the broader First Amendment protections that breach agents can use to defend themselves from tort suits, how courts have resolved the few such lawsuits which have reached them, and the general parameters of those protections.