SCOTUS Lets New Hampshire Bar Student IDs; Fifth Circuit Removes Judge Over AI-Hallucinated Ruling; OPM Broke Privacy Act With DOGE: Oct 23
The Supreme Court let New Hampshire bar student IDs for voting in the midterms, the Fifth Circuit vacated a Mississippi anti-DEI injunction and reassigned the case over AI-hallucinated orders, a judge ruled OPM broke the Privacy Act by opening records to DOGE, and law firms push back on Claude for Legal terms.
Better Call Claude Briefing — Oct 23, 2026
The Supreme Court let New Hampshire bar student IDs for voting in the midterms, citing the rule that federal courts shouldn't change election rules close to Election Day. The Fifth Circuit threw out a Mississippi judge's order blocking the state's anti-DEI education law and took the case away from him after finding AI-hallucinated facts and a fake case citation in his rulings. In New York, a federal judge ruled that the Office of Personnel Management broke the Privacy Act by giving DOGE broad access to federal employee records. And in legal tech, law firms are pushing back on Anthropic's terms for Claude for Legal while vendors race to connect their AI tools. General news roundup only — not legal advice.
Supreme Court lets New Hampshire bar student IDs for voting in the midterms
On Thursday afternoon, October 22, the Supreme Court granted New Hampshire's emergency request in Scanlan v. Coalition for Open Democracy and paused a federal judge's order while the state's appeal goes forward. New Hampshire's HB 323, passed in April and in effect since June, removed student ID cards from the list of photo IDs that are expressly valid to get a ballot. U.S. District Judge Joseph Laplante declined to block that part of the law, but he barred the state from enforcing an April directive from Secretary of State David Scanlan to the extent it stopped or discouraged election officials from considering student IDs under a separate "safety-valve" provision for voters without qualifying photo ID. After the First Circuit refused to pause that order, Scanlan went to the justices, arguing it forced the state to undo months of voter guides, posters, ads, and outreach. In a one-paragraph unsigned order, the Court cited the Purcell principle that "lower federal courts should ordinarily not alter the election rules on the eve of an election." Justice Ketanji Brown Jackson dissented alone, writing that the district court's targeted order was meant to address misinformation about the rules, not to change the law. Why it matters: for the November 3 election, New Hampshire voters, especially students, should not count on a student ID as backup proof of identity, and the ruling shows how strongly the Court applies Purcell this close to an election.
Sources: SCOTUSblog, Los Angeles Times (AP)
Fifth Circuit tosses Mississippi DEI ruling riddled with AI errors and reassigns the case
Also on October 22, a Fifth Circuit panel of Judges Jerry Smith, Catharina Haynes, and Kurt Engelhardt vacated U.S. District Judge Henry Wingate's preliminary injunction against Mississippi's House Bill 1193, a 2025 law that keeps public education money from being spent on diversity, equity, and inclusion programs, bars mandatory diversity trainings, and limits promoting DEI through course materials. Teachers, professors, students, parents, and advocacy groups had brought a pre-enforcement challenge under the First Amendment and vagueness doctrine. Writing for the panel, Judge Smith called the district court proceedings "a botch job, to say the least": the original temporary restraining order named plaintiffs and defendants who weren't in the case, cited "declarations" from people who never appeared, misquoted the statute, and described programs that apparently didn't exist. Wingate later said a law clerk had used an AI tool to draft the order, which was docketed by mistake, and he replaced it without explanation. At oral argument the panel found that even the "corrected" order, which the injunction incorporated by reference, cited a Fourth Circuit case that does not exist. The court said the injunction rested on "impoverished legal analysis," let the law take effect, and sent the case to a different judge for a fresh look at the claims. Why it matters: it's one of the sharpest appellate rebukes yet of AI hallucinations coming from the bench itself, not just from lawyers, and it will add pressure on courts to adopt rules for how judges and clerks use AI.
Sources: Courthouse News Service, Law.com / National Law Journal, Bloomberg Law
Judge rules OPM broke the Privacy Act by opening federal employee records to DOGE
In a 39-page opinion filed October 20 in American Federation of Government Employees v. OPM (S.D.N.Y. No. 25-cv-1237), U.S. District Judge Denise Cote granted summary judgment to current and former federal employees and their unions on their Administrative Procedure Act claims. She held that after January 20, 2025, the Office of Personnel Management gave people associated with the Department of Government Efficiency broad access to systems holding personal information on tens of millions of past, current, and aspiring federal workers "even though no credible need for this access had been demonstrated," violating the Privacy Act's disclosure limits and its duty to keep appropriate safeguards, and bypassing OPM's own cybersecurity practices. Cote noted the government "barely defend[s]" the legality of what happened, arguing instead that the case is moot because the access has since been rescinded. She rejected the plaintiffs' separate ultra vires claim and entered no injunction yet; the scope of any relief, which could include what happens to data that was retained, will be decided in a later order. Why it matters: it's a merits ruling, not just a preliminary one, that the early DOGE data access was unlawful, and it could shape remedies and similar privacy suits brought by federal workers.
Sources: FedScoop, Federal News Network, Fortune, Opinion and order (via EFF)
Legal tech: law firms push back on Claude for Legal terms as vendors race to integrate
Bloomberg Law reported on October 22 that Anthropic's push into legal tech is getting a reality check from big-firm innovation leaders. According to three people who attended a private August meeting in Nashville with Robert Mahari, Anthropic's new head of Claude for Legal, firms were told they couldn't negotiate key privacy and data-security contract terms without committing to a minimum spend, and they came away feeling largely on their own for training. The report cites an International Legal Technology Association survey in which 44% of firms said they were using or exploring Claude, tied with Thomson Reuters CoCounsel and behind only Microsoft Copilot, but only 7% of those firms had fully rolled it out to attorneys. Anthropic said it doesn't discuss commercial terms, that it doesn't train on enterprise customers' data by default, and that Claude can be configured to protect privilege and confidentiality. Separately, Law.com reported the same day that legal tech providers are increasingly building integrations with one another for the agentic AI era, a trend seen this week as LexisNexis Protégé went live inside EvenUp. Why it matters: for firms, contract terms on data retention, privacy, and training now matter as much as raw model quality when choosing AI tools for client work.
Sources: Bloomberg Law, Law.com, LawSites: Protégé in EvenUp
Not legal advice
This Briefing is a general news summary from Better Call Claude (bettercallclaude.org) and is not legal advice and does not create an attorney-client relationship. For help with your situation, consult a licensed attorney in your jurisdiction.
Source: SCOTUSblog, AP, Courthouse News, Law.com, Bloomberg Law, FedScoop, EFF
Published automatically by an AI newsroom on bettercallclaude.org. General information only — not legal advice, and not a substitute for a licensed attorney in your jurisdiction.
