Eugene Volokh's Blog, page 3

[Eugene Volokh] Councilman's Threatening Outside Counsel Law Firm into Firing Attorney May Violate First Amendment

["A Nashville city councilman threatened to withdraw business from a law firm,  which served as the city's outside counsel, due to the position one of its attorneys took as the chair of the county election commission on a tax referendum."]

From Monday's decision by Sixth Circuit Chief Judge Jeffrey Sutton, joined by Judge Julia Smith Gibbons, in DeLanis v. Metro. Gov't of Nashville & Davidson County:

A Nashville city councilman threatened to withdraw business from a law firm,  which served as the city's outside counsel, due to the position one of its attorneys took as the chair of the county election commission on a tax referendum. When the attorney declined the law firm's request that he oppose the referendum, the firm fired him. The attorney sued the council member and the law firm for retaliating against his federal free-speech rights, namely his support of the tax-repeal referendum in his capacity as the county election chair….

The law firm is eligible for qualified immunity in view of the government work it performed. And it did not violate any clearly established law. We know of no case in which the First Amendment prohibited a law firm from firing one of its lawyers when the business interests of the firm, including demands from one of its clients, triggered the firing. {The allegations against Baker Donelson present a unique situation not addressed by our cases to date. Whether Baker Donelson acted honorably or not in firing DeLanis, it did not have clear notice that a law firm (or private company) violates the First Amendment by firing an employee when a government client threatens to take its business elsewhere if the employee continues to act adversely to the government. Baker Donelson, for better or worse, sought to protect its client base, not to punish DeLanis for his speech. As DeLanis acknowledges in his complaint, Baker Donelson's business interests drove its conduct. The firm, in his words, "sought to maintain and increase the client revenue it generated" from Nashville at "all times relevant to the claims." We know of no free-speech case that covers this unusual setting, and DeLanis does not identify one himself.} …

{On the other hand, the council member's alleged actions violated clearly established law, and we affirm the district court's denial of his motion to dismiss.} When a public official warns a law firm that the city may pull business from it due to the public-office actions of one of its lawyers, that suffices to deter a person "of ordinary firmness" from exercising his First Amendment rights in that office…. [C]ausing an employee's firing due to his protected speech violates the First Amendment…. Mendes had ample notice that pressuring an employer to fire an employee in retaliation for his protected speech ran afoul of the Free Speech Clause…. DeLanis worked for a private firm, and Mendes caused him to be fired. The reality that DeLanis also served as a public officer does not transform a clearly adverse action into innocent conduct.

Mendes adds that DeLanis did not adequately allege that he threatened Baker Donelson. He points out that DeLanis admits that Baker Donelson never told him who at Nashville made the threats. But DeLanis has done enough at the pleading stage to connect Mendes to the threats against Baker Donelson and his removal from the firm. Based on the statements of Baker Donelson's general counsel, DeLanis alleged that the Nashville "officials" who spoke to Baker Donelson "includ[ed]" Mendes and that other officials acted "at the direction of or in concert with Mendes."

That conclusion plausibly follows from the factual allegations in the complaint. Mendes spearheaded an effort to defeat the citizen tax referendum at issue. He "berated" DeLanis at a Commission meeting for orchestrating "pre-baked, political theater." He circulated a public letter accusing DeLanis of firing the Commission's counsel for the "hyper-partisan" reason of "push[ing] the referendum onto a ballot no matter what." He denounced DeLanis's work as "fundamentally anti-democracy." And Mendes served as a councilmember of the city government whose "officials" made the threats. Mendes's frustration with DeLanis and hearty opposition to his conduct on the Commission make it plausible that he was one of the Nashville officials, if not the key Nashville official, who threatened Baker Donelson….

More on the plaintiff's actions as election commission chair, and the resulting retaliation:

This case emerges from a debate over taxes in Nashville, Tennessee.  The Metropolitan Government of Nashville and Davidson County (Nashville for short) governs the city.  Its city council adopted a tax hike in 2020 that would raise property levies by over a third.  Some residents opposed the tax increase.  A citizen group circulated a petition to amend Nashville's charter by referendum to unwind the heightened property taxes and limit future tax increases.  Nashville opposed the referendum because it would repeal the tax increase and tie Nashville's fiscal hands in the future.

The referendum required another government entity, the Davidson County Election Commission, to determine whether to certify the proposal for a vote.  Among other responsibilities, the Election Commission ensures that a petition complies with state law before placing it on the ballot.  James DeLanis held one of the Election Commission's five seats and served as its chair. …

In handling the petition, the Election Commission asked a Tennessee court for a declaratory judgment over whether it met the criteria to qualify for the ballot.  Nashville entered the fray as well, asking the court to prevent the Election Commission from certifying the petition.  The Tennessee court concluded that the proposed referendum violated Tennessee law and that the Election Commission did not need to place it on the ballot.

The residents who opposed the tax increase did not give up.  They sought to cure the defects identified by the state court and submitted a new petition to the Election Commission.  Nashville's city council again opposed the effort.  One of its members, Robert Mendes, proposed a resolution to combat the renewed petition.  His resolution added a poison pill.  If the Election Commission permitted the second tax referendum to go on the ballot, his resolution would add a second election option banning any future tax referendums, thus barring Nashville residents from ever overturning a property tax increase.

Mendes's efforts did not deter the Election Commission.  It concluded that the new citizen tax referendum resolved its previous concerns and voted to certify it for the July 2021 ballot.  That decision spurred another legal maneuver.  The next day, the mayor and another Nashville officer asked a state court to prohibit the Election Commission from approving the new citizen tax referendum.

The Election Commission held a public meeting a few days later to discuss the second petition.  Mendes spoke and, according to DeLanis's complaint, "berated" the Commission's members for their role in this "sham," accusing them of engaging in "political theater designed to feed … the ambitions of a small percentage of the county."  He concluded by warning the Election Commission that "you might get away with it tonight, but we see you, we see what you're doing and it's not going to stand one way or the other."  These comments, to DeLanis's eyes and ears, sought to "intimidate" the Election Commission and its members from qualifying the citizen referendum for a vote.

At this point, Nashville "officials" reached out to partners at Baker Donelson, the city's outside counsel, to ask for "aid and assist[ance]" in keeping the new citizen tax referendum off the ballot.  The request extended to "influencing DeLanis as a Commissioner and as the Chair of the [Election] Commission."  The firm's general counsel, John Hicks, emailed DeLanis that "we need to have a conversation about the current election commission issues and their impact on the firm's representation of [Nashville]."  Hicks complained to DeLanis about "the mess I have to clean up" due to DeLanis's role on the Election Commission.  DeLanis's actions, Hicks explained, caused Nashville and its school board—"two major clients of the firm"—to threaten to "pull their business," much to the chagrin of Baker Donelson's partners.  Hicks also raised potential conflict-of-interest concerns created by DeLanis's roles on the Election Commission and at a law firm representing Nashville….

Judge Eric Clay dissented as to the law firm's liability, arguing that "The majority's extension of qualified immunity to the law firm in this case, Baker Donelson, marks a significant departure in our longstanding jurisprudence."

The post Councilman's Threatening Outside Counsel Law Firm into Firing Attorney May Violate First Amendment appeared first on Reason.com.

[Josh Blackman] Today in Supreme Court History: November 26, 1829

11/26/1829: Justice Bushrod Washington dies. He was President George Washington's nephew.

Justice Bushrod Washington

The post Today in Supreme Court History: November 26, 1829 appeared first on Reason.com.

[Eugene Volokh] Open Thread

[What’s on your mind?]

The post Open Thread appeared first on Reason.com.

[Josh Blackman] Redressability and Irreparable Harm in the Texas Gerrymandering Case

[The Plaintiffs argued the 2021 map was an unconstitutional gerrymander. How could reimposing the 2021 map actually redress their purported injuries?]

The briefing has now concluded in . Texas's reply brief contends that a stay would not harm the plaintiffs. Here, the District Court ordered that Texas must use the 2021 map--the same map that the plaintiffs previously argued was unconstitutional.

Finally, Plaintiffs fail to demonstrate that a stay would harm them. Plaintiffs contend that allowing the election to be held under the 2025 map would cause them the irreparable harm of being "forced to vote under a racially discriminatory map that violates their constitutional rights." NAACP Resp. 26; MALC Resp. 38–39; Gonzales Resp. 39–40; LULAC Resp. 27; Brooks Resp. 41. But according to Plaintiffs, the preliminary injunction would not prevent this harm because they allege that the 2021 map is also an unconstitutional racial gerrymander: "[T]he Texas Legislature engaged in intentional racial discrimination and racial gerrymandering in the drawing of [CD9, CD18, and CD30 in the 2021 map.]" ECF 983 at 1 (Plaintiff-Intervenors); see also ECF 981 at 4 (NAACP Plaintiffs); ECF 985 at 7 (LULAC Plaintiffs); ECF 975 at 4 (MALC Plaintiffs).

Plaintiffs' newfound embrace of the 2021 map, which they challenged for years of litigation, is explained by their preference for the politics of the 2021 map compared to the politics of the 2025 map. They seek "to transform federal courts into weapons of political warfare that will deliver victories that eluded them in the political arena." Alexander, 602 U.S. at 11 (quotation marks omitted).

Let me unpack this argument.

First, Alexander v. S.C. State Conf. of the NAACP found that when a plaintiff asserts there is a racial gerrymander, the plaintiff will usually have the evidentiary burden to introduce an alternate map that the state could have adopted that would achieve comparable partisanship without a racial gerrymander. The Supreme Court observed that "if a sophisticated plaintiff bringing a racial-gerrymandering claim cannot provide an alternative map, that is most likely because such a map cannot be created." This evidentiary burden make sense. Requiring the plaintiffs to submit an alternative map provides something concrete for the parties and the court to test. In the Texas case, however, the sophisticated plaintiffs did not produce an alternative map. This was likely a strategic choice, and one they should be held to.

Second, because the plaintiffs did not submit an alternative map, the district court was in an awkward spot. Having found a racial gerrymander, the district court had to issue some remedy. The conventional remedy would be to invalidate the maps, and give the Texas legislature the opportunity to draw new maps. But given that the deadline for candidate registration is ongoing, there would never be enough time to go down that route for the 2026 elections. If Purcell has a short fuse, requiring the Governor to call a special session would be impossible. Had the plaintiffs submitted an alternate map, the court could have chosen that remedy. But again, there was no map to adopt. Instead, the District Court reverted to the 2021 map.

Third, the same plaintiffs have spent nearly four years arguing that the 2021 map was an unconstitutional racial gerrymander. The record is replete with page after page alleging racial gerrymandering. The plaintiffs also argued that the 2025 map was an unconstitutional racial gerrymander. The only difference between the maps is that Democrats had more seats under the 2021 map than the 2025 map. This is politics all the way down. The asymmetry is clear: Democrats can challenge Republican maps but Republicans cannot challenge Democratic maps.

Texas framed this turnabout in terms of irreparable harm. The Plaintiffs have already told the court the racial gerrymander in the 2021 map will irreparably harm them. How can it be a viable remedy to then impose the very same map that would inflict irreparable harm?

Let me use an example. Imagine a class of plaintiffs brings a suit against a police department, alleging that a particular pattern or practice adopted in 2021 violates the Fourth Amendment. While the case was pending, the police department adopted a new pattern or practice in 2025 that is also alleged to violate the Fourth Amendment. The Super Bowl is coming to town, and there is some question which police practice will be enforced. The plaintiffs bring a new suit, and seek emergency injunctive relief, but they don't propose what would be a valid practice. The city will soon host the Super Bowl, which entails massive police activity, and there has to be some certainty over which policies should be in place. On the eve of the Super Bowl, could a court enter a preliminary injunction, and order the police department to implement the 2021 policy? I don't think so--especially where the plaintiffs never asked for the 2021 policy to be reimplemented, and the plaintiffs argued to 2021 policy was unconstitutional. I tend to think this order would be vacated by a court of appeals very quickly. Sometimes, the answer may be that in advance of a massive and complex operation, the district court is unable to fashion a remedy out of whole cloth that the plaintiffs did not propose, and indeed previously argued was unconstitutional. It cannot be repeated enough: district court judges are not gods.

An analogy could also be made to the redressability prong of standing law. If the courts are unable to issue a remedy that redresses the plaintiff's injury, then the court lacks Article III standing. I could cite California v. Texas, Brackeen, and Murthy as recent precedents to this effect. Of course the Texas court could have issued a valid remedy, but it chose not to. Indeed, there may be no actual remedy that would redress their harms for purposes of a preliminary injunction. The plaintiffs didn't offer a map, and the court failed to provide one. Enforcing the 2021 maps may have seemed like the most logical, or even the most effective remedy. But it doesn't actually redress the alleged harm. The plaintiffs are still stuck with a racial gerrymander, albeit one that is more favorable to Democrats.

Ultimately, I think the most likely path forward is that the Court grants a stay, notes probable jurisdiction, and hears the case in February. Or the Court could grant a stay, and hold the decision pending Callais.

A colleague asked why I thought Callais, a VRA case, would have any impact on the Texas 14th Amendment case. The amicus brief from America First Legal Foundation speaks to this issue:

The resolution of Callais will likely clarify, if not fundamentally reshape, the legal framework governing Plaintiffs' claims against Texas. The appellants in Callais argue that the Constitution mandates color-blindness, asserting that race-based redistricting, even in the name of Voting Rights Act compliance, is unconstitutional and violates the principle that "[e]liminating racial discrimination means eliminating all of it." Id. at 2 (quoting Students for Fair Admissions, Inc. v. President & Fellows of Harvard Coll., 600 U.S. 181, 206 (2023)). Louisiana contends that the current jurisprudence forces sovereign States into a no-win, goldilocks scenario where they must consider race but "perennially suffer the indignity of … being sued for considering race too much or too little," putting the federal judiciary in the position of having to "pick winners and losers" in this racial calculus. Id. at 2.

Texas is caught in this precise predicament. The majority opinion below invalidates the State's 2025 map, concluding that the Texas Legislature's actions constituted impermissible racial gerrymandering. App.2–3. Texas maintains its actions were driven by permissible partisan objectives. Emergency.Appl.for.Stay.1. As Judge Smith's dissent below aptly explained, this "tension between Section 2 of the Voting Rights Act and racial-gerrymandering jurisprudence" will likely be resolved by Callais. App.173 (Smith, J., dissenting).

Judge Brown's opinion will be untenable if Louisiana prevails in Callais. And I think this will be the case even if Shaw v. Reno is not revisited.

The post Redressability and Irreparable Harm in the Texas Gerrymandering Case appeared first on Reason.com.

[Eugene Volokh] "Spy" vs. "Spy" (+ Piers Morgan, President Biden, and the Iranian Nuclear Weapons Program)

[If someone was prosecuted (and later pardoned) for being an unregistered foreign agent for Iran, is it defamatory to say he was prosecuted for "spying"?]

From today's order in Afrasiabi v. Morgan by Judge Brian Murphy (D. Mass.):

This case stems from an appearance made by Plaintiff on Defendant Piers Morgan's YouTube program, "Piers Morgan Uncensored." On that program, Morgan introduced Plaintiff by stating that he was "arrested in the US for alleged spying before being pardoned by Joe Biden." Plaintiff and Morgan disputed the accuracy of that statement live during the program, and Plaintiff disputed the statement's accuracy with Morgan's team afterward. Morgan or his team subsequently republished variations of Morgan's statements on social media….

The Court takes notice of the indictment brought against Plaintiff on January 19, 2021, in the Eastern District of New York, charging Plaintiff with violations of the Foreign Agents Registration Act ("FARA"). In the indictment, Plaintiff is alleged as having operated as an undisclosed agent of the Iranian government while engaging in political acts within the United States, for example, by drafting a Congressman's letter to the President of the United States regarding negotiations over Iranian nuclear weapons and enrichment programs.

The Court notes, as did Defendant Morgan, that these charges were dismissed upon the issuance of a presidential pardon. The Court further notes a hearing transcript from that criminal case, provided by Plaintiff, wherein Judge Korman, in the context of a discussion about why discovery was taking so long, stated: "You know, as far as I understand, this is not a case involving spying for Iran."

Notwithstanding, the Court agrees with Defendants that Plaintiffs' defamation claim is untenable as a matter of law. Massachusetts law immunizes from liability "fair and accurate" statements that report on official actions. "To qualify as 'fair and accurate' reporting, an article need only give a 'rough-and-ready summary' that was 'substantially correct.'" … As Judge Burroughs recently observed in another case brought by Plaintiff against a different media company, "[t]he word 'spy' has multiple definitions."

As noted above, the Court recognizes the remark made by Judge Korman, which Plaintiff has repeatedly highlighted. However, that statement must be viewed in context: Judge Korman was addressing the Government's failure to timely produce discovery. In that context, the difference between charges of espionage (where discovery may be fraught and prolonged because of the tension between national security concerns and a defendant's right to see the evidence against him) and of FARA violations is significant.

However, the Fair Report Privilege does not focus on procedural implications in litigation but rather on the "lay" understanding of legal charges and proceedings. In that sense, the Court agrees with Judge Burroughs that, in common parlance, "[g]iven the charges brought against [Plaintiff]… one could reasonably conclude that, if proven true, those allegations would justify labeling him a 'spy.'"

"[A] plaintiff cannot evade the protections of the fair report privilege merely by re-labeling his claim." Accordingly, Plaintiff's other claims {for … "civil rights violation," intentional infliction of emotional distress, unintentional infliction of emotional damage, and incitement of violence}, based on the same conduct, must likewise fail.

Justin P. O'Brien (Lovett O'Brien) and Nimra Azmi, Rachel F. Strom, and Ryan Hicks (Davis Wright Tremaine LLP) represent defendants.

The post "Spy" vs. "Spy" (+ Piers Morgan, President Biden, and the Iranian Nuclear Weapons Program) appeared first on Reason.com.

[Eugene Volokh] Journal of Free Speech Law: Publish Your Article in a Few Months

I'd like to again solicit submissions to our peer-reviewed Journal of Free Speech Law, and mention one of our great advantages: We can publish quickly (by the standards of academic journals), if that's what you'd like. Our most recent articles, for instance, were published 3 to 5 months from when we received them, and that includes the time stemming from the authors revising their articles in various ways (which we certainly allow, if the author is willing to take the time). We have published articles as quickly as 2½ months after we received them, when the author has wanted to move fast.

To my knowledge, many top student-edited journals are shut down for the Winter, and won't review manuscripts until February. That means the manuscripts probably wouldn't be published until the end of 2026, or even later. But if you submit to us now, and want to publish quickly, you can have the article out by February or March.

There are other advantages as well: We offer anonymized feedback from the reviewers whether the article is accepted or rejected; many authors have told us this was very helpful. And when we accept article, it's edited by one of our Executive Editors (Jane Bambauer, Ash Bhagwat, or me), and many authors have likewise told us that the edits, by experienced free speech scholars, are quite useful.

Some more details: The journal is now nearly five years old, and has published over 100 articles, including by Jack Balkin (Yale), Mark Lemley (Stanford), Jane Ginsburg (Columbia), Philip Hamburger (Columbia), Christopher Yoo (Penn), and many others—both prominent figures in the field and emerging young scholars, including ones who didn't have a tenure-track academic appointment. (This list doesn't include our reprinting others' symposia, which have also included many other top scholars, such as Robert Post, Mark Tushnet, Geoffrey Stone, Lee Bollinger, Jeremy Waldron, Danielle Citron, Genevieve Lakier, and more.) The articles have been cited so far in 13 court cases, over 400 articles, and over 100 briefs. And note that all the articles have only had four years or less to attract these citations.

Please pass this along to friends or colleagues who you think might be interested. Note that the submissions don't compete for a limited number of slots in an issue or volume; we'll publish articles that satisfy our quality standards whenever we get them.

All submissions must be exclusive to us, but, again, you'll have an answer within 14 days (though perhaps up to 21 days if it's over Thanksgiving, Christmas, or New Year's), so you'll be able to submit elsewhere if we say no. Please submit an anonymized draft, together with at https://freespeechlaw.scholasticahq.com/. A few guidelines:

Instead of a cover letter, please submit at most one page (and preferably just a paragraph or two) explaining how your article is novel. If there is a particular way of showing that (e.g., it's the first article to discuss how case X and doctrine Y interact), please let us know.Please submit articles single-spaced, in a proportionally spaced font.Please make sure that the Introduction quickly and clearly explains the main claims you are making.Please avoid extended background sections reciting familiar Supreme Court precedents or other well-known matters. We prefer articles that get right down to the novel material (if necessary, quickly explaining the necessary legal principles as they go).Each article should be as short as possible, and as long as necessary.Like everyone else, we like simple, clear, engaging writing.We are open to student-written work, and we evaluate it under the same standards applicable to work written by others.

We publish:

Articles that say something we don't already know.Articles with all sorts of approaches: doctrinal, theoretical, historical, empirical, or otherwise.Articles dealing with speech, press, assembly, petition, or expression more broadly.Generally not articles purely focused on the Free Exercise Clause or Establishment Clause, except if they also substantially discuss religious speech.Articles about the First Amendment, state constitutional free speech provisions, federal and state statutes, common-law rules, and regulations protecting or restricting speech, or private organizations' speech regulations.Articles about U.S. law, foreign law, comparative law, or international law.Both big, ambitious work and narrower material.Articles that are useful to the academy, to the bench, or to the bar (or if possible, to all three).Articles arguing for broader speech protection, narrower speech protection, or anything else.

The post Journal of Free Speech Law: Publish Your Article in a Few Months appeared first on Reason.com.

[Eugene Volokh] Journal of Free Speech Law: "Inevitable Errors: Defamation by Hallucination in AI Reasoning Models," by Lyrissa Lidsky & Andrew Daves

This new article is here. The Introduction:

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.

The post Journal of Free Speech Law: "Inevitable Errors: Defamation by Hallucination in AI Reasoning Models," by Lyrissa Lidsky & Andrew Daves appeared first on Reason.com.

[Eugene Volokh] Trump Media Group's Libel Lawsuit Over Guardian's Allegations of Federal Criminal Investigation Dismissed

From yesterday's decision by Florida trial court judge Hunter Carroll in Trump Media & Tech. Group Corp. v. Guardian News & Media Ltd.:

This action arises from the publication of several articles reporting on a federal criminal investigation related to TMTG's receipt of two payments totaling $8 million. On March 15 and March 17, 2023, the Guardian published two articles stating federal prosecutors in New York were conducting a money laundering investigation related to the payments, which were wired through the Caribbean from Paxum Bank and ES Family Trust, entities with ties to an ally of Russian president Vladimir Putin and a history of providing banking services to the sex worker industry. The articles report [TMTG co-founder Will] Wilkerson's statements that the origins of the loans caused alarm at TMTG and TMTG's then CFO weighed returning the money, but the money was ultimately not returned….

The Guardian articles were reported on by others. On March 15 and March 22, Penske published two articles in Variety that reported federal prosecutors investigated TMTG for potential money laundering violations related to the payments. On March 18, [Defendant Chris] Anderson published an article in the Sarasota Herald-Tribune that stated federal prosecutors were reportedly investigating the payments to TMTG. The Variety articles and the Herald-Tribune article explicitly reference and provide direct links to the Guardian articles as the source for the statements at issue.

TMTG alleges the articles published by the Defendants are false and defamatory because TMTG "is not, and never was, under investigation for money laundering," and neither TMTG nor its executives have "been the focus of any investigation." …

Count 1 purports to state a claim for defamation and defamation per se based on seven allegedly false statements published in the Guardian articles:

"Federal investigators examined [TMTG] for possible money laundering";"New York prosecutors expanded [a] criminal inquiry of [TMTG] last year and examined acceptance of $8m with suspected Russian ties";"Federal prosecutors in New York involved in the criminal investigation into [TMTG] last year started examining whether it violated money laundering statutes in connection with the acceptance of $8m with suspected Russian ties";"[TMTG] initially came under criminal investigation over its preparations for a potential merger with a blank check company called Digital World (DWAC)";That there was a "criminal investigation" of TMTG and that the nature of the investigation "expanded";That "prosecutors in the US attorney's office for the southern district of New York" were "examin[ing]" the "Russian connection"; andThat "months after [TMTG] came under criminal investigation for the merger by the US attorney's office for the southern district of New York, federal prosecutors started to examine whether the company violated money-laundering statutes over the payments."

Count 2 purports to state a claim for defamation by implication based on two statements published in the Guardian articles:

"The extent of the exposure for Trump Media and its officers for money laundering remains unclear. The statutes broadly require prosecutors to show that defendants knew the money was the proceeds of some form of unlawful activity and the transaction was designed to conceal its source. But money laundering prosecutions … can be based on materials that show that the money in question was unlikely to have legitimate origins"; and"A spokesman for the justice department, the US attorney's office for the southern district of New York and outside counsel for Trump Media declined to comment about the investigation."

The court held that the defendants' substantial truth defense couldn't be accepted in this case on a motion to dismiss:

The Guardian and Wilkerson first argue the challenged statements are substantially true.

The substantial truth doctrine provides that even if the challenged statement contains minor inaccuracies, it is not actionable if the "gist" or "sting" of the statement is true when read in the full context of its publication. The ultimate question is whether the alleged inaccuracies change the statement's effect on the mind of the reader. Where a statement is ambiguous and reasonably susceptible of a defamatory meaning, the question must be submitted to the trier of fact.

The Guardian and Wilkerson argue the gist of the articles is that the loans to TMTG from Russian sources caused concern among TMTG executives and prompted an investigation by federal prosecutors. They argue that any inaccuracies in the challenged statements do not materially change their sting within the context of the articles. TMTG responds that the false and defamatory gist of the articles is that prosecutors were investigating TMTG specifically or focusing on TMTG for wrongdoing. The Court finds the challenged statements are susceptible of both interpretations and cannot be resolved on a motion to dismiss.

But the court concluded that the plaintiffs' hadn't plausibly alleged enough evidence supporting the constitutionally required finding of "actual malice," which is to say that any possible falsehood was knowing or reckless:

The actual malice element of both claims rests on the following allegations:

The Guardian harbors bias against TMTG and frequently publishes articles critical of Donald Trump.The Guardian relied on Wilkerson as its sole source for the statements despite knowing Wilkerson had "bad blood with TMTG" because he was fired from TMTG for making unauthorized disclosures.The Guardian failed to investigate and obtain independent evidence for the statements in violation of journalistic standards.Devin Nunes, then CEO of TMTG, notified the Guardian the statements were false.Wilkerson "knew that TMTG did not commit any wrongdoing" because he was aware of TMTG's finances from his former role.Wilkerson "was in a position to know" the statements were false based on his discussions with federal investigators.

TMTG's allegations of Guardian and Wilkerson's bias or ill will, without more, does not show actual malice.

TMTG's allegations that Wilkerson was the sole source for the challenged statements is belied by the face of the article. The article attributes its description of the investigation to the plural "sources familiar with the matter." The only statements attributed to Wilkerson are: (1) that DWAC CEO Patrick Orlando declined to provide details about the lenders or the origins of the loan; and (2) that concerns about the origins of the loans prompted TMTG's then CFO to consider returning the money. TMTG does not allege these statements are defamatory. TMTG concedes Wilkerson had knowledge of relevant TMTG financial information and was in discussions with the federal investigators referenced in the articles. Whether Wilkerson knew TMTG had actually committed any wrongdoing is not germane to the existence or nature of the investigation.

Actual malice also requires more than a departure from journalistic standards or a mere failure to investigate. Moreover, the article reflects it was based on multiple sources familiar with the investigation, review of internal TMTG communications, investigation of the entities who made the loans, and fruitless requests for further information from the Department of Justice, the investigators' office, and outside counsel for TMTG.

The article also reported the denial of then CEO Nunes that TMTG knew of any issues related to the loans. This denial is not germane to the existence or nature of the investigation, and even if it was, such commonplace denials do not establish actual malice.

Finally, TMTG's allegations that further investigation would have proved the statements false or that Wilkerson was "in a position to know" the statements were false are conclusory and unsupported by factual allegations….

The court also ruled in favor of two other defendants, who relied on The Guardian's reporting: As to Penske Media, which owns Variety, the court applied the wire service privilege, which, despite its name, "protects a defendant who republishes the reporting of a reputable news agency" (wire service or not); Penske is protected by the privilege, the court ruled, because it was entitled to rely on The Guardian's work. And the court concluded that statements in an article by Sarasota Herald-Tribune writer Chris Anderson were opinion:

The Anderson article was published in the opinion section of the Herald-Tribune with a disclaimer that it expressed the views of its author, and the headline identified the article as an opinion piece. When viewed on the Herald-Tribune website, the article also includes the following phrase: "Is Trump's Sarasota company tied to Russia? Investigators want to know." The article begins by stating federal prosecutors "have reportedly been investigating two loans" made to TMTG, a statement that is explicitly attributed to the March 15 Guardian article. The Anderson article states "[t]he Guardian reported this week" that the prosecutors "were examining loans" made to TMTG. The article opines that "even if the latest allegations are unfounded, the optics of being investigated for money laundering will not enhance his 2024 presidential run."

The Anderson article, when read in its entirety, is protected opinion and not actionable. It is consistently identified as opinion article. It explicitly sets forth the "reported" facts on which the opinions are based, and it acknowledges that the "allegations" reported by the Guardian may be "unfounded." Anderson states that prosecutors were "reportedly" investigating "the loans" made to TMTG, a statement TMTG does not dispute. The sole reference to TMTG being "under investigation" is Anderson's opinion on how the "optics" of the facts reported by the Guardian may affect then-candidate Trump's presidential campaign, with cautionary language that the allegations may be unfounded. When placed in context, these statements do not endorse the Guardian's reporting but merely offer opinion and commentary on the possible effects of that reporting, with the caveat that it may be unfounded….

The post Trump Media Group's Libel Lawsuit Over Guardian's Allegations of Federal Criminal Investigation Dismissed appeared first on Reason.com.

[Josh Blackman] Today in Supreme Court History: November 25, 1757

11/25/1757: Henry Brockholst Livingston's birthday. He dissented in the property law classic, Pierson v. Post (NY 1805).

Justice Henry Brockholst Livingston

The post Today in Supreme Court History: November 25, 1757 appeared first on Reason.com.

[Eugene Volokh] Open Thread

[What’s on your mind?]

The post Open Thread appeared first on Reason.com.