after noting that many of plaintiff's filings contained seeming AI hallucinations, despite assertions in plaintiff's filings that all the citations were valid.
From Judge Aileen Cannon (S.D. Fla.) in Paulk v. Larva Labs, LLC, decided in June but just posted on Westlaw:
This case presents a dispute over Defendant's alleged fraudulent interference with Plaintiff's Non-Fungible Tokens ("NFTs"). Plaintiff's Complaint, initially filed in Florida state court on January 20, 2026, alleges that in June 2017, Plaintiff used a publicly available software tool to generate 141 NFTs on the Ethereum blockchain. Each of these NFTs represents ownership of a unique CryptoPunk that was part of the first version of CryptoPunks ("V1"), a pioneering generative art project consisting of 24x24 pixel portraits. To create the NFTs, Plaintiff paid only nominal "gas fees" (i.e., fees paid to third parties who validate transactions on the Ethereum blockchain), and the NFTs "have remained in [a] dormant wallet" controlled to this day by Plaintiff.
Plaintiff alleges that, on June 23, 2017, Defendant "deployed" a second version of CryptoPunks ("V2"), "claiming V1 contained a 'bug.'" The Complaint then asserts that Defendant "suppressed V1 marketplace activity while promoting V2 as the sole authoritative collection, rendering Plaintiff's assets commercially impaired". Based on these allegations, the Complaint asserts causes of action for fraudulent concealment (Count I), unjust enrichment (Count II), declaratory relief (Count III), and injunctive relief (Count IV), seeking unspecified monetary damages and an injunction.
The court discussed various procedural issues, and then added:
As a separate matter, the Court is concerned more generally about the plethora of legal filings seemingly prepared by artificial intelligence and submitted by Plaintiff that make frivolous and unsupported arguments to the Court [see, e.g., ECF Nos. 15, 16, 17, 18, 19, 26, 26-1, 31, 32, 33]. As an example, Plaintiff makes the obviously incorrect submission that the citizenship of Defendant's counsel can be considered in determining complete diversity.
And many of Plaintiff's filings contain disclaimers to the Court, expressly certifying "that every legal authority, statutory provision, rule of procedure, and historical case citation referenced within [the filings] has been independently read, analyzed, verified, and cross-referenced against official federal and state legal repositories." In each of these filings, "Plaintiff guarantees that every cited case stands as valid, binding law within the United States Supreme Court and the Eleventh Circuit Court of Appeals, possesses an active and unchecked standing, and accurately represents the procedural postures asserted." And he expressly warrants that he has not included any "fictitious or unverified citations," and that he has "manually double-checked" each case to ensure that it "support[s] Plaintiff's cause perfectly, ensuring that no unverified or 'hallucinated' data is presented to the Court."
No doubt Plaintiff's artificial intelligence legal counsel told him to say all this. And the AI certainly had the right idea—litigants should not make misrepresentations to the Court. Unfortunately, Plaintiff did not practice what his artificial counsel preached. Instead, he cites to "Snapper v. Redan, 523 U.S. 83 (1998)"—a case which does not exist—for the proposition that "when a company's chief legal representative has been sent formal notification at a publicly advertised professional office location, the entity cannot assert an administrative breakdown within its secondary commercial mail hubs as an equitable basis to vacate an otherwise valid default". Notwithstanding the made-up citation, the case Plaintiff means to cite, Snapper, Inc. v. Redan (11th Cir. 1999), is an Eleventh Circuit case that says nothing of the sort. Instead, Snapper established that federal appellate courts have jurisdiction to review remand orders based on the enforcement of contractual forum selection clauses. It does not mention vacatur, or default, or even mail—and it certainly does not stand for the extremely narrow proposition for which Plaintiff cites it.
In light of all this, Plaintiff has violated his duty of candor to the Court and has committed a Rule 11 violation…. [T]he Court [therefore] imposes filing sanctions as follows:
Until Defendant's forthcoming Motion to Dismiss is resolved, Plaintiff shall not file any motion or submit any item on the docket unless Plaintiff first files a motion for leave to file the designated motion/filing and obtains the Court's approval to file the requested material. Except as stated in this Order, any motion or item on the docket that is filed by Plaintiff without this mandatory judicial pre-approval will be stricken without further notice.
Notwithstanding the above requirement, Plaintiff may file an opposition to Defendant's forthcoming Motion to Dismiss in accordance with the local rules of this district. See S.D. Fla. L. R. 7.1(c). Due to Plaintiff's prior artificial intelligence offenses, any such opposition shall be handwritten, and shall refrain from containing any inaccurate citations or hallucinated authority. The failure to comply with any aspect of this Order may cause the Court to grant the forthcoming Motion to Dismiss by default.
I have stayed away from discussing any of the many important electoral contests that will be (hopefully) decided today. But given that he is at the center of one of the most hotly-contested Senate seats, I can't help being reminded of his conduct during the campaign to overturn the 2020 election results, and what may have been the lowest of many low points in Ken Paxton's career to date.
In December 2020, when he was Attorney General of Texas, Paxton was the attorney of record in a lawsuit against the states of Georgia, Michigan, Pennsylvania, and Wisconsin, seeking to overturn the 2020 election results in those states. He petitioned the Supreme Court to exercise its original jurisdiction to hear the case, and to do so on an expedited basis. [See my blog posting at the time here; Paxton's papers filed at SCOTUS are posted here].
In his filing, he wrote:
Expert analysis using a commonly accepted statistical test raises serious
questions as to the integrity of this election. The probability of former Vice President Biden winning the popular vote in the four Defendant States—Georgia, Michigan, Pennsylvania, and Wisconsin—independently given President Trump's early lead in those States as of 3 a.m. on November 4, 2020, is less than one in a quadrillion, or 1 in 1,000,000,000,000,000. For former Vice President Biden to win these four States collectively, the odds decrease to less than one in a quadrillion to the fourth power (i.e., 1 in 1,000,000,000,000,000^4).
The same less than one in a quadrillion statistical improbability [applies to] Mr. Biden winning the popular vote in the four Defendant States. . . . Again, the statistical improbability of Mr. Biden winning the popular vote in these four States collectively is 1 in 1,000,000,000,000,000.
Put simply, there is substantial reason to doubt the voting results in the Defendant States.
Indeed there is! Prima facie proof of fraud in the counting of ballots, I'd say.
But only, of course, it it's correct, or at least close to correct. But it's not correct. It is, as anyone with a knowledge of elementary statistics would have recognized immediately, simply and laughably wrong. The technical term is, I believe, "utter bullshit." I explain why here. The probability that Joe Biden would end up with a majority in any of those is not one in a quadrillion - let alone one in a quadrillion to the 4th power (!), which is approximately equal to the probability that the sun will rise in the West tomorrow, and will have developed the capacity for human speech - nor anything close to one in a quadrillion.
Hey, we all make mistakes, right? What is most disgraceful about this episode is not that Paxton screwed up in his brief to the Court; it is that he never acknowledges his error, either publicly or, more importantly perhaps, to the Court. His assertion that the probability that Biden actually carried these States was equal to the probability that you would pick the single blue marble from a collection of 1,000,000,000,000,000 marbles, is a falsehood, and it is one that is at the core of his argument, in a case that was of grave importance to the future of the country. If Paxton didn't catch that falsehood before he filed his papers with the Court, he's merely a lousy lawyer, which I suppose is nothing to get too worked up about, given that there always will be plenty of lousy lawyers out there. But if he doesn't acknowledge that he has misled the Court after those misstatements are revealed - if he doesn't tell the Court that it has to ignore the part of his argument that relied on this incorrect understanding of statistics - then he's in a much more sinister category.
That this is the candidate whom the Texas Republican party chose as its candidate for U.S. Senate speaks volumes about the current state of the Rule of Law in this Administration. I'm sure that Mr. Paxton is looking forward, assuming the results go his way, to forging a close working relationship with Attorney General Todd Blanche, who is similarly scrupulous in his adherence to fundamental Rule of Law principles (as in the Slush Fund fraud-on-the-court perpetrated by Blanche and the DOJ].
After taking a hiatus for awhile here working on other things, I'm returning to more regular posting. I look forward to engaging with my fellow Conspirators and readers. I have been largely working on financial regulatory issues. But many of these issues are increasingly intersecting with legal challenges that readers might find interesting.
Needless to say, I've been reading the blog regularly even though I haven't been posting and I hope I have something valuable to add.
You can see the verdict form, from Thursday, here. The jury had asked whether they could suggest additional remedies besides the money:
But the court referred the jury to Instruction No. 10, which focused solely on monetary damages.
The backstory, from the opinion this April Judge Rebecca Pennell (E.D. Wash.) in Riera v. Central Wash. Univ., which let the case go forward:
Mr. Riera was employed in a fixed term, non-tenure track faculty at Central Washington University (CWU)…. On the morning of April 1, [2024,] Mr. Riera called the CWU police to report an older, "apparently homeless," woman wandering around Samuelson Hall. He said he wanted to make a report "before things get … out of control." An officer reported to Samuelson Hall and confirmed the identity of the woman as a CWU professor. No further action was taken by campus police or Mr. Riera.
The CWU professor shared her experience with two colleagues. The colleagues immediately filed bias complaints with CWU, alleging the target of Mr. Riera's call—a Black woman—had been the victim of racial profiling.
This led to a great deal of institutional response, including a discussion at a faculty senate meeting. Defendant Erdman, "a lecturer at CWU and member of the faculty senate, emailed unofficial minutes [of the meeting] to non-tenured faculty," and her notes included this:
Welcome to the Deep South, Circa 1935 ☹️
Without naming names or describing the incident, [CWU President] Wohlpart expressed his outrage at a recent unconscionable incident on campus.
The City of Poway … maintains a public Facebook page where it "posts updates about various items that may be of interest to [City] residents and the general public." The Facebook page is open to public comment, and members of the public routinely comment on posts on the page. Plaintiff is a resident of Poway who has criticized aspects of the City's decisions for several years, including by speaking during city council meetings, corresponding with City officials, and posting critical comments on the City's Facebook page. City officials and staff, including those responsible for the Facebook page, appear to recognize Plaintiff as a critic.
Several times during the past year, Plaintiff's comments on the City's Facebook page have been "hidden" or "shadow banned;" that is, his comments are not deleted and are still visible when Plaintiff looks at the posts, but are not visible when other Facebook users look at the posts. The City's Social Media Policy provides that "[t]he City's social media presence serves as a limited public forum and all content shall be monitored and moderated." The Policy indicates that user-generated posts will be deleted if those posts, among other things: "use[ ] abusive, vulgar, offensive, profane, sexual, threatening, deliberately untruthful or defamatory language or content." Additionally, "[p]ostings from a public user account which contain inappropriate or disruptive content, … or believed to be spam, may result in the public user account being blocked from any further activity on the City's social media account … at the discretion of the City Manager's Office."
Plaintiff believes that the City is hiding his comments because it wants to suppress his critical viewpoints…. The City represents that none of its employees have ever hidden or removed any comments made by Plaintiff; instead, Meta's Moderation Assist Program has hidden these comments because it has flagged Plaintiff's Facebook profile as a "non-human 'bot'" account. Meta Moderation Assist "is necessary because the City does not have the resources nor ability to actively monitor each and every comment made in real time in order to manually hide bot comments, particularly considering these comments can be submitted 24 hours a day seven days a week." The City believes that Plaintiff's comments were hidden because Moderation Assist flagged him as a bot: "Moderation Assist will automatically hide a post if (1) the commenter does not have any friends or followers, (2) the comment has a link, or (3) the commenter does not have a profile picture." Plaintiff does not have a profile picture, has only one Facebook friend, and has no posts on his own profile, which "are typically strong indicators that the account is a non-human bot."
The city agreed that viewpoint-based moderation of Jalisi's posts would be unconstitutional, and agreed to "… manually unhiding Plaintiff's prior Facebook posts." But it opposed Plaintiff's request to "remove any restriction that causes Plaintiff's comments on the City Page to be hidden from the public"; it argued that it needed to keep using the Moderation Assist program, and couldn't turn it off only for Jalisi. And the court held in favor of the city as to this:
From D.M. v. A.M., decided by Massachusetts Appellate Court Justices Ariane Vuono, William Meade & Zachary Hillman on Oct. 21:
The defendant, A.M., appeals from the extension of a harassment prevention order …..
To review a civil harassment prevention order, "we consider whether the judge could find, by a preponderance of the evidence, together with all permissible inferences, that the defendant committed acts that constituted one of the enumerated forms of harassment." As relevant here, "[h]arassment" is defined as "[three] or more acts of willful and malicious conduct aimed at a specific person committed with the intent to cause fear, intimidation, [or] abuse." To qualify, the conduct must be intended to cause and actually cause fear of physical harm or damage to property….
Here, as the District Court judge properly found, there were three or more incidents constituting harassment. On June 27, 2025, the defendant called the Target store that D.M. managed, looking for D.M., and became vulgar with the employees when she was told that D.M. was not there.
On June 28, 2025, the defendant showed up at the Target store and demanded to speak to D.M. The defendant then called Target's customer service line and made multiple threats, including that she would "go Luigi" and that she was "at the point of feeling like Luigi" and that D.M. was "gonna be sorry though."
Sauron - an example of a greater evil you can vote for a lesser evil to avoid. (NA)
Tomorrow is election day. And - to put it mildly many of the options before voters are far from wonderful. In all too many races, the candidates with a realistic chance of winning are all bad. This situation, like other recent elections, raises difficult questions about such issues as whether people have a duty to vote, whether it is permissible to vote for a badly flawed candidate if the alternative is even worse, and how to balance competing issues in making voting decisions.
This post doesn't argue for any particular voting choice, but rather compiles pieces about how we should make such decisions in general - and whether we should make them at all. Much of the post is adapted from a similar one I put up in 2024. Unless otherwise noted, these pieces were all published right here at the VC blog:
This post was inspired by the 2016 election, but almost all of it remains relevant today. In it, I explain why it is both permissible and desirable to vote for the lesser evil in an election where all of the viable options are bad. I address a variety of counterarguments, including claims that voting for a lesser evil makes you morally complicit in that candidate's wrongdoing, arguments that it's better to cast a "protest vote" for a candidate that has no chance of winning, arguments based on the very low probability that your vote will be decisive, and more. Here's a brief excerpt:
Imagine an election where the only options are Queen Cersei from Game of Thrones, and Sauron, the Dark Lord from Tolkien's Lord of the Rings. If Cersei wins, she will kill many innocent people, and oppress others. But she will leave much of the population more or less alone (as long as they don't openly oppose her or threaten her family in any way). If Sauron wins, he will kill far more innocent people, and make the survivors his slaves.
You can instead cast a protest vote for a vastly better alternative, such as Gandalf….. But, by assumption, these are purely symbolic options, because they have zero chance of prevailing. If the protest voter would otherwise have backed Cersei, the net effect of his decision to protest is to increase the likelihood of the worst possible outcome: the triumph of Sauron.
Under those circumstances, it seems clear that a person who ensures a Cersei victory has done a good deed. He or she will have saved large numbers of people from slavery or death, even though the Cersei regime would be a deeply unjust one.
The Cersei-Sauron hypothetical was obviously inspired by the contest between Trump and Hillary Clinton, and is an exaggerated version of that tradeoff. But I think it works for many other elections, as well - including many of the choices voters face this year.
Notice how my logic is different from that of those who say you can only vote for a candidate if he or she meets some minimal threshold of character or good policy. On my approach, it is morally justifiable to vote for almost any candidate - even a cruel despot like Cersei - so long as the only feasible alternatives are even worse.
Canadian columnist John Robson responded to my piece in the National Post, and I, in turn, posted a rejoinder.
Many people, including some political theorists, believe we have a moral duty to vote. I disagree. Indeed, I think - in many cases - it's better if you don't vote, especially if ignorance and bias make it likely you will make a bad decision. Casting a ballot motivated by ignorance and bias is often worse than not voting at all.
Even if there is no general duty to vote, perhaps we are morally required to do so in situations where the stakes of an election are unusually high. In this post, I explain why I disagree - at least with respect to most voters. There is, however, a limited kernel of truth in this argument. An excerpt:
There is a kernel of truth to the claim that you have a duty to vote if the stakes are high enough. But the resulting moral duty applies far less often than advocates of the argument tend to assume. And the same reasoning actually implies many people have a moral duty not to vote.
Let's start with the kernel of truth. Imagine there's an election for a powerful political office that pits Gandalf (the benevolent wizard in J.R.R. Tolkien's Lord of the Rings) against Sauron, the despotic dark lord from the same story. If Sauron prevails, millions of people will die or be enslaved, while Gandalf would rule justly if he manages to win. And all you have to do to ensure Gandalf's victory is check his name on a ballot. If you do so, Gandalf wins; if not, Sauron does.
In this scenario, it seems like you have a moral duty to vote for Gandalf, at least barring some kind of extraordinary exigent circumstance. In a real election, of course, the odds that your vote will make a difference are far smaller than in this stylized example….
However, a large enough difference between the two candidates could potentially justify a duty to vote for the "right" candidate, even if the odds of casting a decisive ballot are very low….
But notice that the duty in question is not an obligation to participate in the process for its own sake. It's a duty to help good triumph over evil in a situation where you can do so at little or no cost. If you have a moral duty to vote for Gandalf in these types of scenarios, it follows that you also have a moral duty not to vote for Sauron. Indeed, the person who votes for Sauron is more worthy of condemnation than the one who merely abstains. The former is actively helping evil win, while the latter "merely" chooses not to help stop it.
While Gandalf supporters may have a duty to vote, Sauron supporters actually have a duty to abstain from doing so. Ideally, they should stop supporting Sauron entirely. But they at least should not take any actions that increase the likelihood of his victory.
All of the above analysis assumes that the voter knows which candidate is superior and to what degree. But, in reality, we have widespread political ignorance, and most voters often don't even know very basic facts about how government and politics work. Most are also highly biased in their evaluation of the information they do know….
Unless and until a voter becomes well-informed about the issues and at least reasonably objective in his or her evaluation of political information, she has good reason to question her judgment about which candidate is superior, much less by how much. Thus, she cannot conclude she has a duty to vote to help the "right" side win. She may instead have a presumptive duty to abstain from voting until she meets at least some minimal threshold of political knowledge….
Later in this piece, I note some possible exceptions to situations to the presumptive duty to abstain; there are unusual situations where ignorant voting really is better than no voting, and even unusual situations where ignorance leads to better decisions than knowledge (I expand on the latter possibility in greater detail in Chapter 2 of my book Democracy and Political Ignorance). I also address some other caveats and counterarguments.
Some steps you can take to be a better voter, building on a helpful article in Scientific American. I think a wide range of people can do these things. You don't have to be an academic or policy expert to undertake the task. But I am skeptical that more than a small fraction of the electorate will ever be willing to put in the time and effort.
In this post, I make a tentative argument for letting children vote, so long as they show they have attained a level of political knowledge as high as that of the average adult voter -which is not that high a standard! I expanded on the idea here. I do note the crucial caveat that it may not be feasible to do this, because it's possible government can't be trusted to come up with an objective knowledge test for the children (though we do in fact impose such a test on immigrants seeking to become US citizens). Letting knowledgeable children vote is one of my least popular ideas (which is saying no little, given how many other unpopular views I hold!). But I remain unrepentant about it. My 11-year-old daughter is one of the relatively few fans of this policy.
My proposal should be differentiated from the idea of letting parents cast votes on behalf of their children, endorsed by Republican VP candidate J.D. Vance, among others.
Many of the most serious flaws of voting arise from the fact that most voters tend to be ignorant about government public policy, and highly biased in their evaluation of the political information they do know. In this article, I go over a range of possible strategies for mitigating this problem. Here's the abstract:
There is broad, though not universal, agreement that widespread voter ignorance and irrational evaluation of evidence are serious threats to democracy. But there is deep disagreement over strategies for mitigating the danger. "Top-down" approaches, such as epistocracy and lodging more authority in the hands of experts, seek to mitigate ignorance by concentrating more political power in the hands of the more knowledgeable segments of the population. By contrast, "bottom-up" approaches seek to either raise the political competence of the general public or empower ordinary people in ways that give them better incentives to make good decisions than conventional ballot-box voting does. Examples of bottom-up strategies include increasing voter knowledge through education, various "sortition" proposals, and also shifting more decisions to institutions where citizens can "vote with their feet."
This chapter surveys and critiques a range of both top-down and bottom-up strategies. I conclude that top-down strategies have systematic flaws that severely limit their potential. While they should not be categorically rejected, we should be wary of adopting them on a large scale. Bottom-up strategies have significant limitations of their own. But expanding foot voting opportunities holds more promise than any other currently available option. The idea of paying voters to increase their knowledge also deserves serious consideration.
Feeling down after considering the above? This piece might make you slightly happier; even if you don't vote, you are justified in criticizing the policies of the winners! I would add that the same logic shows that, if you vote for the winner on the grounds that she is a lesser evil, you are still justified in complaining about her policies. Vote for Cersei over Sauron, when necessary. But then condemn the evil she does. Just make sure not to be at the wrong place at the wrong time when she decides to purge her critics.
I think we have seen a new record for speed on the emergency docket. Today, shortly before noon, the Solicitor General filed an emergency application in the California election law case. The brief was about 20 pages long. Circuit Justice never even bothered calling for a response. Shortly before 7:00, the petition was referred to the full Court, and was denied.
Rick Hasen calls it a mercy killing. I think shadow execution is more appropriate.
Indeed, this ruling was likely orchestrated by Circuit Justice Kagan before the petition was even filed. The government's case was dead on arrival. And Kagan spared California the fate that befell Missouri, which had to file a pointless brief that no one would read.
Will Justice Thomas and/or Alito retire if Democrats take the Senate in tomorrow's election, hoping that the lame-duck Senate will confirm a nominee chosen by President Trump? In recent conversations, I've heard increasing speculation and support for this possibility among some judicial conservatives. And I hear corresponding worry among progressives.
Both sides think the election may create a brief, very consequential window for a Republican Senate to replace two aging justices. The GOP Senate would have to confirm their replacements before January 3, 2027. There would be political costs and procedural obstacles, but it could be done. Justice Barrett was confirmed by a Republican Senate within a month of her nomination on the eve of the 2020 election.
The rationale for retirement is straightforward: Don't get trapped in the actuarial straitjacket Justice Ginsburg found herself in when she failed to retire in 2014 (when she was 81). That was her last clear chance of retiring with a Democratic president and Senate.
Thomas is 78. Alito is 76. Neither has disclosed a serious health risk, and both have indicated they have no plans to retire this year.
But if Democrats take the Senate (prediction markets currently give them a 75% chance to do so) and if they play the sort of constitutional hardball over Supreme Court confirmations that Republicans have played in the past ten years, there's very little chance the new Senate would confirm any likely Trump nominee. Democrats needn't reject every possible Trump nominee--just those who would be likely to satisfy the conservative legal movement. So the president might not get another Supreme Court appointment before he leaves in January 2029. Calls for the eldest conservative justices to retire would surely intensify.
It's worse than that for judicial conservatives. If a Democrat wins the White House in 2028, the next opportunity for a Republican president to make an appointment might not come until January 2033. There are some important "ifs" packed in there, but they are well within the range of reasonable possibilities.
In January 2033, Thomas would be 84, Alito 82. Even assuming both remain healthy and active, that's a substantial amount of succession risk for conservatives to accept when they could potentially lock in their seats for another generation this winter.
A rational justice who prioritizes preserving his seat for something close to his judicial philosophy might well retire. But a rational justice might prioritize (or at least see himself as prioritizing) other things: continued service, jurisprudential influence, institutional reputation, and personal legacy. Justices do not see themselves simply as strategic agents of the political coalitions that produced their appointments.
For purposes of predicting what Alito and Thomas might do, the Ginsburg example cuts both ways. It illustrates the risks of refusing to retire when one's political allies control the presidency and Senate. At the same time, it illustrates the limits of assuming justices will subordinate their personal preferences regarding tenure, legacy, and judicial responsibility to the interests of that coalition.
Like Ginsburg, Thomas and Alito would probably resist renewed and intense calls for their immediate retirement. They might figure that:
(1) one of those additional "ifs" won't come to pass—that Republicans will retake the Senate in '28 or retain the White House;
(2) in any event, they'll outlast whatever period of Democratic control ensues (i.e., the mirror of the Ginsburg gamble);
(4) a lame-duck retirement immediately after an electoral defeat would be seen as an unseemly concession to the sort of overtly political behavior by judges they've long publicly decried.
Progressives may dismiss this last consideration. Many critics see Thomas and Alito as essentially unscrupulous political actors who wouldn't hesitate to time their retirements in order to preserve another generation of conservative control. But that assumes they understand their own judicial behavior the way their harshest critics understand it. I doubt they do.
They may regard voting in accordance with their understanding of the Constitution—even when the results consistently advance conservative policy objectives—as fundamentally different from announcing their retirements right after voters have elected a Democratic Senate. On this view, they might think the former is acceptable judicial behavior while the latter is not.
On the other hand, they could see strategic retirements, which are hardly unprecedented among justices, as distinguishable from the politicized judging they've denounced. Choosing the circumstances and timing of retirement is compatible with their judicial philosophies. And many conservatives might well applaud their retirement decisions as responsible constitutional stewardship.
But a retirement announced immediately after the election to permit the outgoing Republican majority to confirm a successor, would be unusually conspicuous. Whatever their views about strategic retirement as a general matter, I'm guessing Thomas and Alito would regard the reputation cost to the Court—and to themselves—under these circumstances as too high.
The irony is that Ginsburg's cautionary example supplies the strongest argument for Thomas and Alito to retire, but also a good reason to doubt they actually will. Justices may be political actors in some respects, but they are not necessarily the political instruments their allies wish them to be.
International human rights law protects freedom of thought absolutely and freedom of expression subject to limitations. The relevant established doctrine calls the two sides of that line the forum internum and the forum externum. However, the line was drawn when thinking left no trace, and with current technological developments, it no longer holds.
Within the span of six months in 2026, a United States court ordered that 20 million ChatGPT conversations be provided as evidence under ordinary discovery rules, two federal courts reached opposite conclusions on whether a litigant's exchanges with a chatbot were protected at all, and the Supreme Court held that location records kept by a third party remain constitutionally protected. Records of where a person's body had been were protected, while records of how their mind worked were not.
The law currently has no category for the search query, the exploratory prompt, the unfinished draft, or the confession made to a machine. The U.S. answer has been a quarrel about privilege: Courts refuse to grant AI exchanges between users and chatbots protection because the machine is not a professional, like a doctor or a lawyer, and the industry asks for a new privilege because it claims the model behaves like one. Both the courts and the industry assume that protection depends on what the person is using, i.e., what the machine is. This Article argues that it depends on what the person who uses the machine is doing.
"In 2021, allegations were brought against [Gino] claiming that she committed research misconduct, which triggered an approximately 18-month long inquiry and investigation by HBS administrators and faculty," and eventually to her tenure being revoked and her being fired.
Plaintiff Francesca Gino is a former tenured professor at Harvard Business School ("HBS") and brings this action against Defendants President and Fellows of Harvard College ("Harvard") and Defendant Srikant Datar ("Dean Datar"). In 2021, allegations were brought against Plaintiff claiming that she committed research misconduct, which triggered an approximately 18-month long inquiry and investigation by HBS administrators and faculty. In March 2023, the investigation committee found, by a preponderance of the evidence, that Plaintiff committed research misconduct…. In May 2025, Harvard revoked Plaintiff's tenure and terminated her appointment with HBS.
Plaintiff claims that Harvard breached its obligations under various policies and procedures governing the investigation of the research misconduct allegations and the tenure revocation proceedings. Specifically, Plaintiff argues that a jury could find that Harvard deviated from its policies and procedures such that Harvard could be found in breach of those agreements, and that Harvard did not carry its burden of proving research misconduct or grave misconduct. Plaintiff also alleges that the proceedings and related findings, including the adverse employment actions taken against her, were pretext for gender discrimination….
In approximately June of 2021, three professors—Leif Nelson, Joseph Simmons, and Uri Simonsohn, who publish under the name "Data Colada"—presented their allegations against Gino to HBS Professor Max Bazerman during a 90-minute Zoom call…. [Eventually, a Harvard] Investigation Committee issued its 41-page Final Report with nearly 1,200 pages of supporting documents, finding by a preponderance of the evidence that Plaintiff committed research misconduct intentionally, knowingly, or recklessly with regard to all five allegations.
The Investigation Committee's Final Report considered and rejected Plaintiff's contentions that unnamed "bad actors," research assistants, or other co-authors might have instead been responsible for the data anomalies. The Investigation Committee concluded that Plaintiff's "repeated and strenuous argument for a scenario of data falsification by bad actors across four different studies," involving "different co-authors, different journals, different lab personnel, and different home institutions" and spanning eight years, was "highly implausible." The Investigation Committee recommended corrections of the scientific record, putting Plaintiff on unpaid leave, and initiation of steps leading to termination of employment….
The court granted summary judgment on all these issues, in Harvard's favor. Here's an excerpt of the discussion of her breach of contract claims:
The Fifth Circuit ordered en banc rehearing Wednesday; here's an excerpt from the long panel opinion (Computer & Comm. Industry Ass'n v. Paxton), decided in July by Fifth Circuit Judges Dana Douglas and Patrick Higginbotham (I posted about the case then):
[1.] [Texas] H.B. 18 … imposes requirements on DSPs [Digital Service Providers]…. With the exception of its age-verification requirement, H.B. 18 applies only to DSPs whose digital service "connects users in a manner that allows users to socially interact with other users on the digital service," "allows a user to create a public or semi-public profile for purposes of signing into and using the digital service," and "allows a user to create or post content that can be viewed by other users of the digital service," including via a message board, chat room, or other "landing page, video channel, or main feed that presents to a user content created and posted by other users."
{H.B. 18 exempts from its coverage state and local government websites, financial institutions, medical websites, small businesses, higher education websites, employee management software, school education software, and e-mail and direct messaging services that provide only those services. H.B. 18 also exempts digital services that "primarily function[ ] to provide a user with access to news, sports, commerce, or content primarily generated or selected by the [DSP]" and provide only incidental social functions.}
H.B. 18's monitoring and filtering requirement mandates covered DSPs to develop and implement a strategy to prevent known minors' exposure to "harmful material" and other content "that promotes, glorifies, or facilitates" "suicide, self-harm, or eating disorders"; "substance abuse"; "stalking, bullying, or harassment"; and "grooming, trafficking, child pornography, or other sexual exploitation or abuse." "Harmful material" is defined with reference to its definition in the Texas Penal Code [to refer to obscene-as-to-minors sexually themed material]….
[2.] Consistent with other courts of appeal, we have interpreted Section 230 to provide DSPs with "broad immunity" from "all claims stemming from their publication of information created by third parties." … Building on Doe v. MySpace, Inc. (5th Cir. 2018), where we had previously held that Section 230 "shielded MySpace from negligence liability for publishing communications between a minor and an adult who later sexually assaulted her," in Free Speech Coalition, Inc. v. Paxton (5th Cir. 2024), we held that an age-verification requirement for pornographic websites was different, and not preempted, because it did not pertain to "immuniz[ing] web service providers from harm caused by unremoved speech on their website," which is "the point of Section 230."
We distinguished between liability "reliant on the harm done by third-party content," which is preempted by Section 230, and "liability purely based on whether plaintiffs comply with the statute," which is not. "[T]he nature of Section 230's protections," we observed, is "to protect a provider from speaker-liability stemming from the speech it hosts." …
Today, The Dispatch published my new article on the case challenging Trump's Section 338 tariffs against Canada. Here is an excerpt:
Courts have already ruled against two avenues President Donald Trump used to impose his tariff regime. That hasn't stopped his efforts to attempt to deploy obscure or outdated trade law provisions to justify new import levies.
His latest tool is the Smoot-Hawley Tariff Act of 1930, the notorious legislation that gravely exacerbated the Great Depression.
Section 338 of that law is the basis of Trump's massive 50 percent tariffs on a variety of Canadian imports, including motor vehicles, furniture, dairy products, and more. On October 20, Gellert Global Group Corp., a major food importer and producer based in New Jersey, filed the first lawsuit challenging those tariffs before the U.S. Court of International Trade, the federal court with jurisdiction over tariff cases.
Gellert, which imports Canadian cheese subject to the tariffs, is represented by the Liberty Justice Center, a public interest law firm I worked with on the case challenging the massive tariffs Trump imposed under the International Emergency Economic Powers Act (IEEPA). It resulted in the tariffs' invalidation in February by the Supreme Court, which not only ruled that IEEPA doesn't authorize tariffs but also emphasized that tariffs are a congressional power, and that the president does not have the power to "impose tariffs on imports from any country, of any product, at any rate, for any amount of time." The Section 338 tariffs are a similar power grab.
They are illegal for multiple reasons, part of an ongoing effort by this administration to usurp Congress' constitutional authority over tariffs. In addition, they harm the U.S. economy and poison relations with a key ally. Section 338 of the 1930 Tariff Act has never been used by a president to enact tariffs before; Philip Zelikow, a conservative trade policy expert and former Bush administration official, has shown that Section 338 was superseded by legislation enacted in 1962 and 1974.
Even if Section 338 is still operative, Trump's tariffs on Canadian goods violate its requirements. As the complaint explains, the Canadian trade measures cited by Trump as justification for the new tariffs do not discriminate against U.S. goods; Section 338 can only be used against discriminatory policies. Some of the Canadian measures retaliate against previous tariffs imposed by Trump. Most do not actually treat U.S. goods less favorably than imports from many other countries. In addition, the Section 338 tariffs go well beyond "offsetting" Canadian measures, violating another requirement of the law….
The administration's interpretation of Section 338 also runs afoul of the major questions doctrine, which requires Congress to "speak clearly" when authorizing the executive to make decisions of "vast economic and political significance." Under Trump's approach, Section 338 can be used to impose 50 percent tariffs whenever a trading partner imposes any restrictions on U.S. goods, or simply chooses not to buy them. And the resulting tariffs need not be calibrated to those they are supposed to "offset." This is a blank check to impose massive tariffs against almost any imports at any time—exactly the kind of power the Supreme Court, in its ruling on the IEEPA case, determined the executive cannot have…
If Section 338 is still valid and does grant the kind of nearly unlimited authority Trump claims, that would violate the constitutional "nondelegation" doctrine, which restricts delegation of legislative power to the executive…
The Section 338 tariffs are part of an ongoing series of import taxes Trump has instituted in the wake of the IEEPA decision, seeking to again usurp the unlimited tariff authority the Supreme Court denied him. He previously tried to use Section 122 and Section 301 of the Trade Act of 1974 to essentially replicate the IEEPA tariffs. In both cases, he sought to transform a narrowly limited tariff authority into a blank check to impose massive tariffs at will…
The tariffs against Canada—combined with Trump's repeated threats to make Canada the "51st state"—have also poisoned U.S. relations with a key ally. The percentage of Canadians with a favorable view of the U.S. has fallen from 54 percent in 2024 to only 33 percent this year—a worse rating than even China's authoritarian regime…. Alienating a key ally and trading partner doesn't make America "great again." It makes us weaker, poorer, and more isolated.
From Sanchez v. Almond, decided Wednesday by Judge John McConnell (D.R.I.):
Mr. Sanchez began working at the office of Rhode Island Governor Daniel J. McKee in May 2021 as a "constituent services" staffer. The office is "the first point of contact with the public for the Governor's office" where employees handled calls, emails, letters, walk-ins, constituent problems, requests, and questions about the Governor's positions.
Mr. Sanchez was not a member of senior staff, and did not formulate policy, or policy developments. While the scope of his employment is shaded differently by the parties, they all agree that his duties involved, among other things, helping Rhode Islanders access public benefits and triaging their concerns by referring them to the appropriate departments. The office received roughly 500 calls and 400 emails per month.
Mr. Sanchez's direct supervisor was David Barricelli, who reported to Joseph Almond, Senior Deputy Chief of Staff, who in turn reported to Antonio Afonso, the Chief of Staff in charge of hiring and firing in the Governor's office, including Mr. Sanchez. Mr. Sanchez's performance was never called into question, and he was never disciplined before his termination….
Sanchez was also a Providence City Council member (and his superiors had not objected to his running for local office).
On October 7, 2023, Hamas attacked Israel killing roughly 1,200 people and Israel responded with fierce military actions in Gaza, whose population is overwhelmingly Palestinian. That day, the Governor's office issued a statement condemning "the abhorrent acts of terrorism by Hamas against innocent civilians in Israel" and expressed solidarity with Israel.
The same day, Mr. Sanchez posted on his personal social media a call for "respect of all human rights, especially the Palestinian people who faced atrocities for decades[,] and at the same time condemn all attacks on civilians." Ten days later, Mr. Sanchez posted that the United States was "actively aiding in a genocide" and called for "immediate deescalation and cease-fire."
In Johnston v. County of Gila(D. Ariz.), Judge Dominic Lanza rejected, on procedural grounds, plaintiff's challenge to a Gila County Personnel Commission decision upholding his firing from his job as "a legal secretary with the Gila County Attorney's Office"; here's the factual and legal analysis from the Commission decision (see exhibit 10):
On September 10, 2025, at around 2:00 p.m. Arizona Time Appellant posted "Rot in Hell Charlie Kirk" on/via his Facebook account.
Multiple GCAO employees, including some who are not Appellant's Facebook friends and/or who do not personally use Facebook, saw and otherwise learned of Appellant's post on September 10 and in the days that followed. Multiple GCAO employees discussed the post with each other and complained about the post to GCAO's Chief Deputy Joseph Collins as well as to Gila County Attorney Bradley Beauchamp.
Beauchamp believed that the post appeared to condone murder based on the victim's political viewpoints and statements and became concerned that this post and perception would reflect on GCAO in the public's eyes and thereby undermine the public's trust in GCAO's official duty to prosecute homicides without bias and without regard to a victim's views, statements, or affiliations. Several other GCAO employees shared the perception that Appellant's post appeared to condone the murder of someone based on their statements and views and shared the concern that this would reflect poorly on GCAO and undermine public trust.
Multiple GCAO employees reported being offended or otherwise bothered by the post. Appellant's post resulted in immediate meetings and changes to office practices due to concerns of disharmony within the office and the effects it would have on GCAO operations.
Appellant's Facebook account was reportedly set to 'private' which Appellant believed would limit his post to only being seen by his existing Facebook friends and that all his friends had to do if they were upset by his post was 'unfollow' him. Notwithstanding Appellant's Facebook account setting and his belief as to the setting's effect on post-exposures, his post apparently reached multiple witnesses within a short timeframe, including some who either were no t his Facebook friends or did not use Facebook at all….
From Commonwealth v. Outlaw, decided Thursday by Pennsylvania Supreme Court Justice David Wecht, writing for a unanimous court:
In June 2022, Philadelphia County Court of Common Pleas Judge Roxanne E. Covington was holding a hearing on a motion to suppress evidence in Marquis McClain's criminal case. Seated in the gallery of the courtroom was McClain's acquaintance, Basile Outlaw. After Judge Covington announced that she would be denying McClain's suppression motion, Outlaw stood up and said "the judge is a cocksucker" loud enough for Judge Covington to hear. Judge Covington responded, "What did you say?" Outlaw replied: "Never mind. I'm leaving." Outlaw then exited the courtroom.
A deputy sheriff and a court employee began searching the building for Outlaw. The trial court recessed the suppression matter to allow McClain's defense attorney to confer with his client in the wake of the court's suppression ruling. The court subsequently adjourned, but Judge Covington asked McClain's attorney to "stick around for another moment … just in case." Around this time, Outlaw came back to the courtroom on his own, at which time he was taken into custody for direct criminal contempt.
Outlaw was tried and sentenced "to fifteen to thirty days' incarceration"; he served thirty days. But the Pennsylvania Supreme Court reversed the conviction:
Sign up here to receive all of the new Volokh Conspiracy posts in a daily e-mail.
The All-Argentina Issue is Here
New from Reason: a look into what Javier Milei
has accomplished and what it will take to make the most of Argentina’s window for real change. Out now
for subscribers.
