In a per curiam opinion joined by eight judges, the Fifth Circuit held in Tesla v. NLRB that an NLRB decision about unfair labor practices by Tesla would be vacated and remanded for further proceedings:

We hold that Musk’s tweets are constitutionally protected speech and do not fall into the categories of unprotected communication like obscenity and perjury. And the Board does not dispute the general rule that it (like every other part of the Government) is powerless to delete protected speech.

But nine other judges didn’t join that opinion. As detailed below, Judge Haynes concurred in the judgment only, and eight judges joined a dissent. So what the NLRB is supposed to do on remand is not entirely clear.

 

Hon. Jennifer Walker Elrod has taken office as the new Chief Judge of the Fifth Circuit, succeeding Hon. Priscilla Richman. The Texas Lawbook has a good story on this “changing of the guard” for this critical leadership role. The new Chief tells the Lawbook:

“I believe that we are all in this endeavor together to uphold our Constitution and try to follow the rule of law and this enterprise works best when everyone gets to participate and have their say … And we can learn from each other if we’re all at liberty to engage. … I also believe iron sharpens iron … by learning what others who think differently believe, and what the basis for that belief is, that can help you to better reinforce your view, or perhaps, to change your mind.”

Unsurprisingly, given all three judges’ discomfort with the Fifth Circuit precedent that dictated the panel holding in Abraham Watkins v. Festeryga, that case will be considered by the en banc court. The issue, as summarized by the panel majority, is this:

Edward Festeryga, an attorney embroiled in a dispute with his former law firm, wants this case heard in federal court and contends we have appellate jurisdiction over the district court’s remand order because waiver is neither an issue of subject-matter jurisdiction nor a defect in removal procedure under 28 U.S.C. § 1447(c). We agree, but our 40-plus-year-old precedent provides otherwise, holding that a waiver-based remand order is jurisdictional under § 1447(c) and thus unappealable under § 1447(d).

While no longer in the academy, the capable Rory Ryan offered this insightful analysis of this case on X.

Dickson v. Janvey, a dispute about the scope of an anti-litigation injunction, offers two basic reminders about the process of separating holding from dicta:

  • Not discussed, likely not holding. “[W]e have previously held that a district court’s power over a receivership enables it to enjoin third parties or non-parties from pursuing certain claims involving the res of the receivership estate. But our statements in those cases implicated the equitable remedies available to the district court and not its jurisdiction. No one objected to personal jurisdiction in those cases, likely because any such objection would have been frivolous. But even if there were a latent in personam defect in those cases, our silence could never be construed as an implicit holding.” (citations and footnotes omitted).
  • Not necessary, likely not holding.Hall suggested in dicta that federal courts may enter in rem injunctions in aid of a previous in rem judgment.  It is unclear to us what the Hall court meant by this dicta.  But … the Hall court’s reference to an ‘inrem injunction’ was unnecessary to its decision. Federal jurisdiction in that
    case was not in rem. So the court’s in rem discussion was nonbinding dicta.” (citations omitted).

No. 23-10726 (Aug. 9, 2024) (Enthusiasts of dicta-holding distinctions will recall that courts have discretion whether to give effect to obiter dicta–an unnecessary but thoughtful statement–under the circumstances of a particular case.)

After another Supreme Court term featuring reversals of the Fifth Circuit on standing grounds in high-profile cases, I wrote this op-ed in today’s Dallas Morning News about that recurring issue.

Late last week, Judge Ada Brown from the Northern District of Texas held that the FTC exceeded its authority by its new rule about noncompetition agreements, granted a preliminary injunction, and set the matter for trial in late August. Notably, as of now, the relief granted does not include a nationwide injunction about the rule.

Much ink will be spilled over the Supreme Court’s use of history in resolving cases this term about the Appropriations Clause, possession of firearms by dangerous people, and the SEC’s internal courts. One key point in Jarkesy is the extent to which a word, used in the 1790s, has the same general meaning today. As the below quote illustrates, “fraud is fraud” oversimplifies the Supreme Court’s holding – but not by much. “Fraud” in today’s commercial law means pretty much what it did in the 1790s. Other words, such as “economy,” are not so easily transported through time, and that reality calls for caution in seeing Jarkesy as providing broad support for some other “originalist” ideas.

On remand from the Supreme Court after that court’s rejection of a challenge to the CFPB’s funding based on the Appropriations Clause, the Fifth Circuit issued a short judgment reflecting that ruling. Interestingly, the judgment expressly identifies the rehearing deadline while striking an earlier 28j filing by the plaintiff:

That filing is no longer available online, but the CFPB’s response suggests that the parties dispute the scope and effect of the Supreme Court’s mandate–and what that may mean for the other challenges to the CFPB presented in this case.

After receiving considerable public comment at the start of 2024, the Fifth Circuit chose not to implement a rule about the use of generative AI, stating:

 

As noted previously, the Fifth Circuit denied en banc review by an 8-8 vote in a contentious forum dispute. The breakdown of the votes is follows (the entire panel majority opinion appears in the chart, and the panel dissent is reproduced as an exhibit to the dissent from the denial of en banc review):

I hope you find this cross-post from 600 Commerce to be informative!

The most recent Advocate (the quarterly publication of the State Bar of Texas Litigation Section) has several articles about how the new Fifteenth Court of Appeals will get off the ground. I have a short piece on where the new court is likely to look for precedent, since it will have none of its own to start. I hope you find it useful in thinking about this important new appellate forum.

Several disputes about inter-circuit venue transfers are ongoing (I was recently interviewed by Bloomberg about this phenomenon):

  • SpaceX. In a dispute between SpaceX and the NLRB, the Fifth Circuit is considering whether to review a transfer order en banc. The NLRB recently filed its response to an unusual order from the panel asking the NLRB to explain several actions taken earlier in the proceedings. The gist of the NLRB’s response was:

Only one court may have jurisdiction at a time. The transferee court was notobliged to follow the February 26 order, and zealous advocacy required the NLRB to present its legal arguments as to why it should not be followed to the Central District of California. Thus, the NLRB urged that court, not to ignore this Court’sorder, but to acknowledge it and respectfully decline retransfer. 

  • CFTC. In a dispute involving the Commodities Futures Trading Commission, the District of the District of Columbia has received the district court’s request to return the case, along with briefing and argument from the parties about the appropriate next step, and as of April 6 continued to have that request under consideration.
  • CFPB. In a dispute involving the CFPB and a new rule about credit-card late fees, a 2-1 panel decision granted mandamus relief on April 5–after a case had been transferred to the District of the District of Columbia, concluding:

Because the Chamber had a short window of time to either (1) comply with the Final Rule, or (2) seek a preliminary injunction, the district court’s inaction amounted to an effective denial of the Chamber’s motion for a preliminary injunction. That effective denial is properly before us on appeal. The district court lacked jurisdiction to transfer the case after this appeal was docketed because doing so would alter its status. … The district court is ORDERED to reopen the case and to give notice to D.D.C. that its transfer was without jurisdiction and should be disregarded.

  • CFPB dissent. The dissent in the CFPB case concluded: “For the foregoing reasons, I believe that the new proposition of law created by the majority is incompatible with district court discretion over docket management and prudent policing of forum shopping. Finally, I am confident the District Court for the District of Columbia will give the suggestion that it should disregard a case docketed by it its closest attention.”
  • HHS. The Court has expedited argument (to May 1) of National Infusion Center v. Becerra , a challenge to the dismissal of a case about 2022 drug-reimbursement regulations on venue grounds (after the dismissal of a party for jurisdictional reasons).

A recent policy statement from the Judicial Conference of the United States recommended changes to judge-assignment practices in district courts. The statement has drawn considerable attention both pro and con; this Volokh Conspiracy post is a good summary of the “con” side. A recent letter from the Chief Judge of the Northern District of Texas says that its judges have declined to materially change that district’s judge-assignment policies.

After an earler (unexplained) grant of an administrative stay touched off weeks of fast-paced appellate litigation about Texas’s “SB4” immigration law, a majority of the Fifth Circuit’s merits panel denied any further stay of the trial court’s injunction against enforcement of that law. USA v. Texas, No. 24-50149 (March 26, 2024). Argument is scheduled next week; barring Supreme Court intervention, merits opinions similar to these are likely.

All eyes will be on New Orleans this morning, for the (videoconferenced) arguments in United States v. Texas, where last night’s order suggests a 2-1 decision will be forthcoming that continues to bar enforcement of Texas’s SB4 during the pendency of its appeal. The Supreme Court will likely be asked about the resulting order, whatever it may be.

You can tell your argument isn’t working when the Fifth Circuit summarizes it as follows:

“SCW’s last remaining counterargument is that it should be able to pick and choose different clauses from the LSAA and the Grant Agreement and then mush them together to demand money from debtors.”

and when the Court begins its opinion:

“Widowed octogenarians Iris Calogero and Margie Nell Randolph received dunning letters from a Louisiana law firm … .”

Calogero v. Shows, Cali & Walsh, LLP,  No. 22-30487 (March 15, 2024). More substantive review to follow in the week ahead!

The Fifth Court ordered a rare reversal for a new trial because of improper closing argument in Clapper v. American Realty Investors. The Court summarized the improper statements as “employ[ing] nearly every type of improper argument identified by our court, including highly improper and personal attacks against opposing counsel, remarks about Clapper’s wealth, a discussion of matters not in the record, insinuations that Clapper had lower moral standards because he was from Michigan, and suggestions of Clapper’s bad motives through counsels’ opinion.”

The Court concluded: “We remind all practitioners in our court that zealous advocacy must not be obtained at the expense of incivility. As Judge Reavley aptly explained, ‘Although earnest, forceful, and devoted representation is both zealous and proper, Rambo and kamikaze lawyers lead themselves and their clients to zealous extinction.'” No. 21-10805 (March 8, 2024).

Notably, footnote two dismisses several arguments about preservation, concluding that “[t]he serious nature of the argument in this trial … indicates that substantial justice requires a new trial ….”

February 2024 has been a busy month for en banc votes. The recent vote by the full court about a stay in U.S. v. Abbott, as well as votes to deny en banc review of Baker (a takings case about police destruction of a home) and Solis (holding that a preferential-transfer claim was stated as to a Stowers-related payment) provided an unusual snapshot of the full court’s views on multiple issues at the same time.

Those votes are now supplemented by an 8-9 vote to deny review in McNeal v. LeBlanc, a panel opinion that denied qualified immunity in an “overdetention” case. The below chart summarizes those votes (a “yes” vote is for en banc review or issuance of a stay, as appropriate):

Judges Jones, Ho, and Oldham voted “yes” for review of each of these four cases. Judges Southwick, Haynes, Higginson, Douglas, and Ramirez voted “no” for review of each of these four cases.

The National Court Reporters Association recently published a fascinating “white paper” about “ethical and legal issues related to the use of artificial intelligence … and digital audio recording of legal proceedings.” It’s succinct, thoughtful, and raises questions relevant to just about any area of law practice or court administration that’s touched by the influence of generative AI and related technologies.

A Fifth Circuit panel applied circuit precedent to reject a liablity claim involving Snapchat in Doe v. Snap, Inc., No. 22-20543 (June 26, 2023), stating: “Parties complaining that they were harmed by a Web site’s publication of user-generated content . . . may sue the third-party user who generated the content, but not the interactive computer service that enabled them to publish the content online.” By a one-vote margin, the full court denied en banc review, as follows (notably, Edith Jones voted with the court’s Democrats to not review the panel opinion): 

 

Whatever your views of the remarkable civil-rights issue presented by Wilson v. Midland County (the intersection between some highly technical immunity rules and the bizarre injustice of a county employee working simultaneously for the prosecution and the courts), one can admire the deft prose of Jude Willett’s opinion:

In a dissent from a recent, close en banc vote in a criminal case, Judge Smith reiterated an earlier warning to again caution: “Highly consequential opinions should not be designated as unpublished in an obvious effort to discourage judges from voting in favor of en banc rehearing.” United States v. Ramirez, No. 22-50042 (Sept. 19, 2023).

I had an op-ed in today’s Dallas Morning News about recent friction between the Supreme Court and Fifth Circuit on standing in some high-profile constitutional/administrative-law cases.

A few years ago, I examined en banc opinons in the Dallas Court of Appeals, and concluded that they tended to be either: (1) “error correction” of panel opinions that had become out of step with the rest of the state and/or the supreme court; (2) “successful failure” cases where en banc review became moot when the supreme court took the case; and (3) “Goldilocks” cases that involve significant issues, but not of such importance that supreme-court review is guaranteed.

The Fifth Circuit’s en banc cases fit that general taxonomy (Brackeen and the FHFA case qualifying as “successful failures”), with the recent opinion in Hamilton v. Dallas County an example of error-correction. The majority opinion summarized:

[T]he panel concluded that it was “bound by this circuit’s precedent, which requires a Title VII plaintiff” to have “suffered some adverse employment action by the employer” and which says that “adverse employment actions include only ultimate employment decisions such as hiring, granting leave, discharging, promoting, or compensating.” Because “the denial of weekends off is not an ultimate employment decision,” the panel affirmed the district court’s dismissal. The panel concluded by urging the full court to “reexamine our ultimate-employment-decision requirement” in light of our deviation from Title VII’s plain text. We granted rehearing en banc to do so.

No. 21-10133 (Aug. 18, 2023) (en banc) (footnotes omitted).

After a well-publicized debacle in New York involving “research” in which ChatGPT invented nonexistent cases, the phenomenon of “hallucinatory” generative AI has arrived in Texas. The Waco Court of Appeals recently faulted a lawyer for providing a brief with three nonexistent cases (and, like the New York counsel, not responding when the problem was pointed out by the appellee). While there is nothing wrong with using ChatGPT to help with legal analysis and writing, uncritical acceptance of purported caselaw found “on the Internet” is always risky.

In May 2022, a Fifth Circuit panel held in Jarkesy v. SEC that the Seventh Amendment’s right to civil jury trial extends to an SEC enforcement action. The full Fifth Circuit later denied en banc review of the matter.

Critics of the administrative state celebrated the ruling as an important limit on agency power; others questioned whether “originalism” was fairly applied to an agency and a set of statutes that did not exist in 1792.

On June 30 of this year, the Supreme Court granted review of Jarkesy, which will be a fascinating addition to the next term of that court:

In United States v. Texas, last Friday, the Supreme Court reversed a Fifth Circuit judgment because Texas had no standing to bring a particular claim about immigration policy. The case echoes the proceedings in California v. Texas, a 2021 matter in which the Supreme Court also reversed a Fifth Circuit judgment for lack of standing — in that case, an issue about the enforceability of the Affordable Care Act.

As Texas’s Attorney General, Greg Abbott famously quipped: “I go into the office in the morning. I sue Barack Obama, and then I go home.” In a recent interview, I suggest that these opinions are a yellow light for that approach to public-law litigation; Mark Stern makes a similar point in more colorful language for Slate. Time will tell whether that traffic signal is heeded.

Two New York lawyers filed a brief with “fake law” in it, having accepted at face value a “hallucination” by ChatGPT abouot several cases that did not in fact exist. The district court has now sanctioned them. The opinion is of broad general interest, given the great influence of generative AI in its few short months of widespread public availability. And it’s also of broad general interest in avoiding sanctions–the lawyers had at least one good chance, and maybe even two, of coming clean about their error, but instead “doubled down” with purported screenshots of the nonexistent cases.

Raskin v. Dallas ISD addresses whether a mother can sue, pro se, in federal court on behalf of her children under the Genetic Information Nondiscrimination Act. Acknowledging both the practical reality that many such actions must be brought pro se, and that the federal courts ordinarily require the professionalism of trained counsel, the majority opinion deftly read the relevant statutes to let the district court apply discretion to this issue on remand.

A dissent favored an absolute rule allowing such pro se proceedings, noting, inter alia, customs about pro se representation when the Judiciary Act was adopted in 1789. (to be fair, though, women such as Ms. Raskin did not have the ability to sue about much of anything in 1789, pro se or otherwise).

A recent Texas Lawbook article considers a new standing order by Hon. Brantley Starr about reliance on language-model AI programs when writing briefs. Judge Starr’s point is well-taken and Rule 11 should probably be revised to include a specific reference to this issue, given how pervasive these programs will likely become in the near future.

I asked ChatGPT what it thought about the matter (specifically: “What are the pros and cons of requiring attorneys to certify to the accuracy of court fiings made with the assistance of language-model AI programs, such as yourself?”) and here is what it had to say:

______

Introducing a requirement for attorneys to certify the accuracy of court filings made with the assistance of AI language models like GPT-4 could have several pros and cons. Here are some:

Pros:

  1. Responsibility and Accountability: Requiring certification ensures that attorneys review the generated content thoroughly before submitting it, increasing the reliability and accuracy of the information presented. It also maintains responsibility with the attorney, discourishing the misuse of AI and ensuring ethical conduct.
  2. Increased Accuracy: By requiring attorneys to certify accuracy, any incorrect or misleading information generated by AI can be corrected, minimizing the risk of procedural or factual errors that could potentially affect a case’s outcome.
  3. Improved Understanding: Since an attorney must certify the accuracy of the filing, they’re incentivized to understand the nuances of the AI-generated content better, making sure they can defend it in court.

Cons:

  1. Additional Burden: This requirement could add to the workload of attorneys, who are already often busy. While AI systems can increase efficiency, the certification process may negate some of this advantage.
  2. Potential Liability Issues: There might be concerns about the increased legal liability for attorneys. If an error from an AI system slips through, the attorney could be held responsible, which could lead to professional consequences or disciplinary actions.
  3. Complexity of AI Outputs: AI language models can generate vast amounts of complex content. While they can be extremely useful, it might be challenging for attorneys to verify all details accurately, especially given the complexity of the legal language and the high stakes involved.
  4. Possible Overreliance: There is a risk that attorneys might rely too heavily on AI-generated content and miss important aspects of a case that the AI, which lacks a deep understanding or context, may not catch. While certification could theoretically mitigate this, the added responsibility may be insufficient.

Ultimately, the success of such a certification requirement will depend on the balance between leveraging the benefits of AI assistance and ensuring the maintained quality, accuracy, and ethical use of legal filings. As AI evolves, it’s crucial to continually reassess such policies to ensure they meet the needs and challenges of the legal profession.

Literature is filled with nonexistent books, such as the “Necronomicon” in the works of H.P. Lovecraft, the “Treatise on the Binomial Theorem” written by Sherlock Holmes’s menace, Professor James Moriarty, etc.

To that collection, we can now add Varghese v. China South Airlines, Ltd., 925 F.3d 1339 (11th Cir. 2019).

ChatGPT helpfully offered that opinion to a New York lawyer researching a technical limitations issue. Unfortunately for that lawyer, the opinion did not actually exist and neither did any of the cases that it purported to cite. An understandably vexed federal judge is now considering sanctions about the matter.

Language-model AI tools are extraordinarily powerful and making advances every day. But they aren’t designed to do legal research, and anything they say about a specific precedent needs to be verified (and not by asking ChatGPT to verify itself, as happened in the federal case cited above).

In response to an April 27 letter from Senator Charles Schumer expressing concern about case-assignment practices in the Northern District of Texas (which, while well-intentioned, also notably misspells “Abilene” as “Abeline”), Chief Judge David Godbey wrote back yesterday to defend the District’s practices: “In addition to the fair administration of justice, we must also consider: the number and type of civil and criminal cases filed in a division, which varies significantly from division to division; the convenience of the jurors, witnesses, parties, and attorneys; the desire of communities to have local judges; the burden of travel on court personnel; and the need to provide judicial support for divisions without a resident district judge.”

After the Supreme Court’s stay ruling yesterday, a Fifth Circuit panel will proceed wth oral argument on May 17 in New Orleans. (Information about the audio livestream may be found in that link.)

While the Supreme Court’s order said very little, the votes of seven Justices were consistent with the position of Judge Haynes when the stay issue was before the Fifth Circuit, and no Justice indicated agreement with the analysis of the per curiam panel majority. (I recently observed in Slate that the “Dobbs-era Supreme Court is well aware of the judicial strand of conservative thought, as well as the political.”).