US courts have sanctioned lawyers for AI-fabricated citations in a growing number of cases. In the Charlotin AI Hallucination Cases database (page last updated 10 October 2026), the US decisions that involve a lawyer rose from 14 in 2024 to 207 in 2025 and 361 so far in 2026, by our count of its CSV download. This page tracks the US matters where a court fined, removed, referred, suspended or reprimanded a lawyer over fake or misrepresented authority. Each row gives the docket, the order date, what the court said was decisive, and a link to the order.
Reviewed and corrected on 12 October 2026. Every row was re-read against the court's order, with one flagged exception. This is general information, not legal advice; pull the order before relying on any row.
TL;DR
- This tracker holds 19 US matters. It is a curated list. The latest row is a June 8, 2026 order. We did not systematically check for later orders in these matters.
- The money range is wide. It runs from $2,000 in Gauthier v. Goodyear and $2,500 in Fletcher v. Experian, through $10,000 in Noland and Doiban, $15,000 per lawyer in Whiting, and $31,100 in Lacey, up to $110,204.38 in Couvrette v. Wisnovsky, which includes fee-shifting.
- Penalties go beyond fines. Colorado suspended a lawyer for one year and one day, with 90 days to be served and the rest stayed on probation. The Ninth Circuit suspended two lawyers from practice before it for six months. A Mississippi federal court barred two lawyers from appearing for two years and disqualified two others. The Sixth Circuit denied appointed counsel his fees and referred him for possible discipline. Nebraska's Supreme Court referred a lawyer to its disciplinary counsel, and local television reported a temporary suspension.
- Tools vary. ChatGPT is named in five rows (Mata, Park, Crabill, Johnson, Noland). Claude is named in Gauthier and Noland. Tools sold to lawyers appear too: counsel in Farris admitted using Westlaw's CoCounsel, and a lawyer's declaration in Lacey named CoCounsel, Westlaw Precision and Google Gemini. Many orders name no tool.
- Several orders point to the response, not only the fake. Courts criticized denying AI use, explanations they found not credible, cosmetic fixes and repeat errors.
- This is general information, not legal advice. Pull the underlying order before you rely on any row.
Your associate's brief cites a case you cannot find. She says she asked the chatbot to confirm it, and it said the case is real. What does the record in Mata v. Avianca suggest about that kind of check?
Part of our legal AI verification and hallucination guide series. For the rule-by-rule lens, see AI hallucinations in legal research and how to avoid sanctions. For the biggest single total, see our Couvrette breakdown.
Methodology
This is a curated list and it does not cover every matter. A matter earns a row when a US court did one of five things over AI-fabricated or misrepresented authority: fined a lawyer, referred one to bar discipline or a grievance panel, removed counsel from a case, suspended a license or the right to practice before the court, or issued a formal public reprimand or admonishment.
For this version, we read the court's own order or opinion for every row, and we link it (for Crabill, the disciplinary judge's approved summary). The one exception is a reported suspension in Nebraska, which rests on a local television report. Where the order does not say a tool was used, we say so. In three rows (Whiting, Nwaubani and Prososki) the court did not decide whether AI produced the fakes. We kept them because the orders discuss AI and the problem, fabricated authority, is the same.
Counting rules: every row is a licensed attorney, and pro se filings are not listed. When a court added the other side's fees to a flat penalty, we show both. When a court set a per-item price, we note it. Dates are the order date, not the filing date.
For the broadest running count, including pro se and non-US filings, see the Charlotin database and our read of it.
AI hallucination sanctions tracker: US cases
Ordered newest first.
| Case (docket) | Court | Order date | What the order describes | What the court stressed | Outcome | Order |
|---|---|---|---|---|---|---|
| Withers v. City of Aberdeen (1:24-cv-00218) | N.D. Miss. | Jun 8, 2026 | Filings from both sides in a contract dispute over legal fees cited hallucinated cases. Wilson (plaintiff's pro hac counsel) used generative AI to draft; Williams (the City's pro hac counsel) used an AI tool for research. Neither verified. | The court wrote that "blindly relying on technology" produced the citations. Local counsel did not draft the filings and had not reviewed them. | Wilson and Williams: pro hac vice revoked, barred from appearing in the district for two years, fined $2,500 and $3,500 (Wilson also ordered to take an AI ethics CLE). Local counsel Ridgeway and McClinton: disqualified from the case, $1,000 each. Order sent to the Mississippi, Louisiana and Texas bars. | Order (ECF 123) |
| Lnu v. Blanche (24-4790) | 9th Cir. | Jun 3, 2026 | Immigration briefs with multiple nonexistent cases, misattributed quotations and gross misrepresentations of real cases. At the firm, unlicensed law-school graduates ("Brief Writers") wrote the legal arguments and attorneys did not vet their citations. | The attorneys called the errors typographical, and one told the court three times at argument that AI "was not used," then said it was "possible." The court said the discipline "is owed to this repeated failure of candor." | $2,500 each, six-month suspension from practice before the Ninth Circuit, and orders to give the order to their clients, opposing counsel, the presiding judges and the firm's lawyers. A two-year requirement to certify AI use in future filings. Order served on the State Bar of California. | Order |
| Coomer v. Lindell (1:22-cv-01129) | D. Colo. | Jul 7, 2025 and May 7, 2026 | A response to a motion in limine with nearly thirty defective citations, including cases that do not exist. Counsel admitted AI use only after questioning from the bench. In 2026, a mis-cited case (a district court decision described as Tenth Circuit authority) in a post-trial brief. | The court said the record undercut counsel's account that a wrong version was filed. In 2026 it noted counsel had already been sanctioned for "this exact type of error," said it was skeptical the new error was only human error, and wrote that a seasoned attorney should not make it whether or not AI was used. | 2025: $3,000 on the lead lawyer and his firm, jointly and severally, and $3,000 on co-counsel. 2026: a further $5,000 on the lead lawyer and firm. No bar referral, because he said he had stepped back from trial work for health reasons. | 2025 order (ECF 383); 2026 order (ECF 424) |
| Fivehouse v. U.S. Department of Defense (2:25-cv-00041) | E.D.N.C. | Apr 27, 2026 (filed Apr 28) | A response brief by a government attorney with fabricated quotations and misstated holdings. The court found he used a generative AI tool to draft it and did not verify the contents. | The court found he knowingly submitted the brief, was not candid about using AI, and held a position of trust. It did not find his account of how the brief came to be filed credible. | Public reprimand. No fine, because of the professional consequences he had already suffered. | Order (ECF 129); the court's May 7, 2026 amended order (ECF 131) corrects a typographical error and makes no substantive change |
| United States v. Farris (25-5623) | 6th Cir. | Apr 3, 2026 | Court-appointed counsel admitted using Westlaw CoCounsel to draft both briefs. They had three invented quotations and misstated the holdings of two real cases. No case was fake. | He relied on "staff" to supervise the AI output. The court credited his prompt, candid response. | No Criminal Justice Act payment for the appeal; counsel removed by separate order; opinion sent to the chief judge to consider discipline and to the Kentucky Bar's disciplinary clerk. | Opinion |
| Prososki v. Regan (S-25-295), 321 Neb. 38 | Neb. Sup. Ct. | Mar 20, 2026 | Divorce-appeal brief. The court's own chart lists 20 problematic authorities, four labeled "Fictitious case," plus real cases with fictitious quotations and holdings. Counsel said AI was not used; the court did not decide the point. | The court called the claim that he filed the wrong draft not credible, and counsel conceded the authority was not grounded in Nebraska law only after the bench pressed him. | Brief stricken, appeal dismissed, counsel referred to the Counsel for Discipline. Local TV reported an April 2026 temporary suspension; we did not read that order. | Opinion (blocks automated checks; open in a browser); court case page; our breakdown; WOWT |
| Doiban v. Oregon Liquor and Cannabis Commission (A181982), 347 Or App 742 | Or. Ct. App. | Mar 18, 2026 | Opening brief by William Ghiorso with at least 15 fabricated case citations, at least nine fabricated quotations, and inaccurate descriptions of real cases. Counsel told the court AI was used for an outline and staff then copied from search results. | Opposing counsel flagged the problem in an April 2025 answering brief. The reply did not respond, and counsel addressed it when questioned at argument in November 2025. | $10,000 payable to the court. The Ringo formula would have been at least $16,500; the court capped it because of counsel's explanation, new procedures and a response filed before Ringo. A redacted replacement brief with certification was allowed. | Opinion |
| Whiting v. City of Athens (24-5918, 24-5919, 25-5424) | 6th Cir. | Mar 13, 2026 | Over two dozen fake citations and misrepresentations of fact across three consolidated appeals. The court asked counsel whether AI was used; they did not answer, and it names no tool. | Counsel refused to answer the show-cause questions and called the order void. Egli had a 2017 public censure for lack of candor; Irion was suspended from the Eastern District of Tennessee for five years in August 2025 for lying to that court. | $15,000 each in punitive sanctions, plus the appellees' full fees on appeal and double costs (Fed. R. App. P. 38 and inherent authority). Opinion sent to the chief judge to consider discipline. | Opinion |
| In re Nwaubani (25-9517) | 4th Cir. | Mar 11, 2026 | Three nonexistent cases across an opening and reply brief in Bolden v. Baltimore Gas and Electric. Counsel denied using AI. The court did not decide whether AI was used. | The court held that citing nonexistent cases violates D.C. Rule 8.4(d) however the cases were produced. | Public admonishment (unpublished opinion). | Opinion |
| Fletcher v. Experian (25-20086) | 5th Cir. | Feb 18, 2026 | A reply brief with 16 fabricated quotations and 5 other serious misstatements. The court found counsel used AI to draft a substantial part, if not all, and did not verify. She named vLex and Casetext among her sources; the court did not find which tool produced the errors. | Her initial show-cause response did not mention AI. The court said that had she accepted responsibility, "it is likely that the court would have imposed lesser sanctions." | $2,500. The order came after the court's January 6, 2026 opinion in the same appeal. | Order |
| Couvrette v. Wisnovsky (1:21-cv-00157-CL) | D. Or. | Dec 12, 2025 (fees set Mar 23, 2026) | Three briefs with 15 nonexistent cases and fabricated quotations attributed to eight real authorities. | The amended briefs deleted the fake cites but left the statements of law; the court found no appropriate corrective action and called it a "coverup." | $110,204.38 across two lawyers: a $15,500 penalty at $500 per fake case and $1,000 per fake quotation, plus $94,704.38 in fees. Briefs struck, claims dismissed with prejudice. Details. | Dec. 2025 order; Mar. 2026 order |
| Noland v. Land of the Free (B331918), 114 Cal.App.5th 426 | Cal. Ct. App., 2d Dist. | Sep 12, 2025 | Counsel said he wrote initial drafts, "enhanced" them with ChatGPT, then ran them through other AI platforms to check for errors. In the opening brief, 21 of 23 case quotations were fabricated, and the reply had more. | He admitted at argument that he did not read the "enhanced" briefs before filing. | $10,000 to the court, copy of the opinion to his client, and the clerk was told to send it to the State Bar. | Opinion |
| Johnson v. Dunn (2:21-cv-01701) | N.D. Ala. | Jul 23, 2025 | Five problematic citations across a motion for leave and a motion to compel. A partner at the defendant's firm used ChatGPT and did not verify; two other lawyers signed. | The attorneys admitted ChatGPT use in a written response filed the day the supplemental show-cause order issued, and again at the hearing. The court said reprimands and modest fines had proven insufficient deterrents. | Three lawyers publicly reprimanded, disqualified from the case and referred to the Alabama State Bar. No fine. Two other lawyers and the firm released without sanction. | Order |
| Lacey v. State Farm (2:24-cv-05205) | C.D. Cal. | May 5, 2025 (special master; filed May 6) | A supplemental brief where about nine of 27 citations were incorrect and at least two authorities did not exist. A lawyer at one firm used CoCounsel, Westlaw Precision and Gemini for an outline; lawyers at the second firm built it into the brief without checking. | The special master found the lawyers collectively acted in a manner "tantamount to bad faith," and that a revised brief still contained a half-dozen AI-related errors. | Briefs struck and discovery relief denied. The two firms owe $31,100 jointly and severally: $26,100 in special master fees plus $5,000 to the defense. No sanction on individual lawyers. | Order (ECF 119) |
| Wadsworth v. Walmart (2:23-cv-00118) | D. Wyo. | Feb 24, 2025 | Motions in limine cited nine cases; eight did not exist. The drafter used his firm's in-house platform, then included the output without verifying. | The court listed the drafter's "honesty and candor" as a mitigating fact. The two lawyers who signed had not reviewed the motions. | Drafter: pro hac vice revoked, removed from the case, $3,000. The two signers: $1,000 each. | Order |
| Gauthier v. Goodyear Tire & Rubber (1:23-cv-281) | E.D. Tex. | Nov 25, 2024 | A summary-judgment response drafted with Claude cited two nonexistent Fifth Circuit cases and quoted language not found in six real ones. | A Lexis AI check did not flag the problems, and counsel took no corrective step after the defendant's reply pointed them out, until the show-cause order. | $2,000 penalty (Rule 11(b)(2)), at least one hour of Texas CLE on generative AI, and a copy of the order to his client. | Order |
| Park v. Kim (22-2057), 91 F.4th 610 | 2d Cir. | Jan 30, 2024 | A reply brief cited a nonexistent state decision that counsel said ChatGPT suggested. | Counsel could not furnish the case when ordered; her response called it "this non-existent case." | Referred to the court's Grievance Panel and ordered to give the decision to her client. | Opinion |
| People v. Crabill (23PDJ067) | Colo. Presiding Disciplinary Judge | Nov 22, 2023 | A motion to set aside a judgment cited cases found through ChatGPT that were incorrect or fictitious. He discovered the problem before the hearing but did not tell the court or withdraw the motion. | When the judge raised concerns, he falsely blamed a legal intern; six days later he filed an affidavit saying he used ChatGPT. | Suspension of one year and one day, 90 days to be served and the rest stayed on completion of a two-year probation, by approved stipulation. | Summary |
| Mata v. Avianca (1:22-cv-01461) | S.D.N.Y. | Jun 22, 2023 | An opposition brief cited fake opinions, including Varghese v. China Southern Airlines. Avianca's counsel listed seven decisions it could not find. | The court found bad faith from "conscious avoidance and false and misleading statements," including shifting accounts of how the cases were found. | $5,000 penalty, jointly and severally on two lawyers and their firm (Rule 11 or inherent authority), plus letters to the plaintiff and to the six judges falsely named as authors. | Order (ECF 54) |
Note: the Fletcher sanction order followed the court's January 6, 2026 opinion in the same appeal, which vacated an earlier fee sanction against the trial lawyer. Coomer v. Lindell produced two sanction orders, which is why two dates appear. For Couvrette, the later order set the fee split, so the row carries both dates.
By the numbers
A quick read of the 19 rows above, with the Charlotin database used where noted.
- By order year: 2 matters in 2023, 2 in 2024, 5 in 2025 and 10 in 2026 (through the June 8 Withers order), counting Coomer at its later order.
- Largest total: $110,204.38 in Couvrette, of which $15,500 is a penalty paid to the court and the rest is fee-shifting. In the Charlotin CSV we downloaded on 12 October 2026, that is the largest US monetary entry. The next is ByoPlanet v. Johansson (S.D. Fla., 2025), where fee awards after sanctions came to $85,567.75 (sanctions order and fee order). ByoPlanet is not a row in this table.
- Penalties paid to the court: from $2,000 in Gauthier to $15,500 for one lawyer in Couvrette and $15,000 for each lawyer in Whiting.
- Beyond money: 13 of the 19 rows involve a step other than a fine, such as a suspension, a bar, a removal from the case, a revoked pro hac vice admission, a reprimand or admonishment, or a bar or grievance referral. The other six are money, with or without a struck brief, a CLE order or a certification requirement (Coomer, Doiban, Fletcher, Lacey, Gauthier and Mata).
- Tools: the orders and the lawyers name ChatGPT in five rows, Claude in two (Gauthier, Noland), CoCounsel in two (Farris, Lacey), and a firm's in-house platform in Wadsworth. The rest name no tool, or the court did not decide whether AI was used.
Patterns: what gets lawyers sanctioned
Read the rows together and a few arcs repeat. A lawyer reaches for AI to find authority or draft a section. The output includes citation-shaped text that is not real, or real cases with invented quotations. Nobody checks it against a primary source before filing.
Several orders say the lawyer's response after the problem surfaced mattered. Read the examples below as what the orders say. They do not add up to a formula, because totals also depend on how many fakes there were and whether fees were shifted.
Asking the chatbot to confirm. In Mata, Judge Castel wrote that "existing rules impose a gatekeeping role on attorneys to ensure the accuracy of their filings." The record there includes ChatGPT telling Steven Schwartz that Varghese was a real case. The bad-faith finding rested on conscious avoidance and on false and misleading statements, including shifting explanations about how the cases were found.
Denying AI use. In Lnu, one lawyer told the Ninth Circuit three times that AI was not used and then said it was possible. The court said denying AI use gains nothing and tied the gravity of its discipline to the lack of candor. In Fletcher, counsel's first response did not mention AI, and the court said lesser sanctions were likely if she had been more forthcoming. In Fivehouse, the lawyer admitted AI use after questioning, and the court found a lack of candor.
An explanation the court did not credit. The Nebraska Supreme Court said the claim that counsel simply filed the wrong draft "lacks credibility." The federal court in Coomer found that the record undercut counsel's account of filing the wrong version. The Fivehouse court found the lawyer's account of the circumstances not credible.
A cosmetic fix. In Couvrette, Brigandi deleted the fake cases and fabricated quotations but left the statements of law standing. The court said that was no correction and called it a "coverup." Compare Wadsworth, where the court appreciated prompt withdrawal of the motions and candor about AI use, though the drafter still lost his pro hac vice status and paid $3,000.
A repeat. Coomer shows a second sanction for a new mis-citation after the first order.
Blaming someone else. Zachariah Crabill blamed a legal intern when the judge raised concerns, and the stipulated discipline was a suspension. In Farris, the Sixth Circuit said relying on "staff" instead of an attorney to supervise AI output fell short of counsel's obligations. In Wadsworth and Withers, lawyers who signed without reviewing were sanctioned too.
Two broader points. First, the problem is not limited to consumer chatbots. In Farris, the Sixth Circuit said the risk persists "even when new tools are sponsored by trusted legal technology providers." Second, the language in the orders is converging. The California court in Noland wrote that no filing should contain citations that the responsible attorney has not "personally read and verified," and the Sixth Circuit in Whiting quoted it. On how much to fine, the Whiting court wrote that "smaller fines have plainly been inadequate."
What the rules require now
The orders rest on existing rules. The duty to verify your own citations predates ChatGPT.
ABA Formal Opinion 512, issued July 29, 2024, applies the Model Rules to generative AI. It says lawyers should either gain a reasonable understanding of a tool's benefits and risks or draw on others who have it (Rule 1.1), that output used in court must be "carefully reviewed to ensure that the assertions made to the court are not false" (Rules 3.1, 3.3 and 8.4(c)), and that managerial and supervisory lawyers must set policies and supervise (Rules 5.1 and 5.3).
Federal Rule of Civil Procedure 11(b) is the other backbone in district courts: by presenting a paper, a lawyer certifies that, after a reasonable inquiry, the legal contentions are warranted by existing law. Mata, Wadsworth, Gauthier and Couvrette are Rule 11 orders. Appellate courts used other tools: the Sixth Circuit in Whiting relied on Appellate Rule 38 and inherent authority, the Fifth Circuit in Fletcher on Appellate Rule 46(c) and inherent authority, and the Ninth Circuit in Lnu on Appellate Rule 46(b). In Oregon's state appellate court, the basis in Ringo and Doiban was ORCP 17 C.
Courts and individual rules add layers. The Eastern District of Texas has a local rule (AT-3(m)) telling lawyers who use generative AI to verify the content. The Fifth Circuit considered and declined to adopt a circuit-wide AI certification rule, and said existing rules are sufficient. The Ninth Circuit in Lnu required the sanctioned firm to certify AI use and cite-checking in future filings for two years. The rules differ by court, so check the court where you appear, and see the state-by-state AI rules guide.
How to not end up on this list
Treat this as a practice checklist drawn from the orders, not a legal rule.
- Treat every AI citation as a lead. Mata shows a chatbot assuring a lawyer a fake case was real.
- Check every cite and every quotation in a primary source. Park says lawyers must "read, and thereby confirm the existence and validity of, the legal authorities on which they rely." Farris involved invented quotations and misstated holdings from real cases, and in Noland most of the cited cases were real while many of the quotations were fabricated (the court also noted a few of the cited cases did not exist), so an existence check alone would have missed most of the problems.
- Read the cases, not just confirm they exist. The Ninth Circuit in Lnu said a competent lawyer must also "read and reason."
- Supervise nonlawyers and juniors. Farris faulted reliance on staff, and Wadsworth and Withers sanctioned lawyers who signed without reviewing.
- Have a response plan. If a cite is flagged, verify in a primary database, say so to the court and opposing counsel quickly, and fix the propositions as well as the citations. Lnu gives the instruction: "immediately alert the court and opposing counsel of the error and disclose its source."
FAQ
How many lawyers have been sanctioned for AI hallucinations? This tracker lists 19 US matters. That is a curated set, not a count of every incident. The Charlotin database, last updated 10 October 2026, lists 2,184 decisions worldwide and 1,492 in US courts. In the CSV we downloaded on 12 October 2026, about 587 US entries involve a lawyer and 885 involve a self-represented litigant. The database counts decisions that address AI-hallucinated content, including warnings, not only sanctions.
What is the largest AI hallucination sanction in the US? The largest total we found is $110,204.38 in Couvrette v. Wisnovsky (No. 1:21-cv-00157-CL, D. Or.), made up of a $15,500 penalty on San Diego lawyer Stephen Brigandi plus $94,704.38 in the defendants' fees split between him and Portland local counsel Timothy Murphy. Charlotin's database shows it as the largest US monetary entry as of its 10 October 2026 update.
Can a lawyer lose their license for using AI? Not for using AI itself. In Colorado, the disciplinary judge approved a suspension of one year and one day for Zachariah Crabill (No. 23PDJ067), with 90 days served, after he blamed an intern and filed fake cases. The Nebraska Supreme Court referred Greg Lake to its disciplinary counsel, and local television reported a temporary suspension in April 2026. We did not read that order. The Ninth Circuit's six-month suspension of two lawyers in Lnu is a suspension from practice before that court, not a license suspension.
Was Mata v. Avianca the first AI hallucination sanction? It was the first high-profile one. The Fifth Circuit's Fletcher order describes it as "the first high-profile incident of AI-fabricated case citations in the federal courts." In Mata (No. 1:22-cv-01461, S.D.N.Y., June 22, 2023), Judge P. Kevin Castel ordered Steven Schwartz, Peter LoDuca and their firm to pay $5,000 jointly and severally.
Do tools sold to lawyers show up in these cases? Yes. In Farris, counsel admitted using Westlaw's CoCounsel to draft both briefs. In Lacey, a lawyer's declaration named CoCounsel, Westlaw Precision and Gemini. In Fletcher, counsel named vLex and Casetext among her sources, and the court did not find which tool produced the errors. In each, the duty to read the cited authority stayed with the lawyer.
What rule do these sanctions rely on? It varies by court. District courts commonly use Federal Rule of Civil Procedure 11 and local rules. Appellate courts used Appellate Rules 38 and 46 and inherent authority. Oregon's state appellate court applied ORCP 17 C. Bar discipline rests on the professional conduct rules, such as competence, candor to the tribunal and supervision, which ABA Formal Opinion 512 ties to generative AI.
Why do some lawyers get a small fine and others get suspended? The orders point to conduct after the problem surfaced as well as to the fakes. The Ninth Circuit tied its discipline to "this repeated failure of candor," and the Fifth Circuit said lesser sanctions were likely if counsel had been forthcoming. The size of the penalty also depends on the number of fakes, the court's rules and whether fees are shifted, so there is no formula.
Is this tracker complete? No. It is a curated list of US matters, each tied to a court order you can read. For the broadest running count, see the Charlotin database.
How do I keep my firm off a list like this? Treat every AI-produced citation as a lead and check each one, and each quotation, against a primary source before filing. Read the cases. Put one person in charge of verification and have a response plan for when a cite is flagged. The verification checklist is a working starting point.
What this tracker excludes and why
We leave some matters out on purpose. Pro se filings are excluded, although they make up most US entries in the Charlotin database. Warning-only orders are excluded, where a judge flagged a bad cite but imposed no sanction. Fabricated factual statements with no fake or misquoted legal authority do not qualify. Non-US matters are out of scope.
A living document
This tracker is updated as new US sanctions and disciplinary orders are confirmed against the order itself. It is general information, not legal advice.
Update log: reviewed and corrected on 12 October 2026. Every row was re-read against the court's order, the amounts and dates were corrected where the order differed, and several claims that the orders do not support were removed. The latest order in the table is June 8, 2026.
For the US statutes, regulations and court rules these laws sit in, see the US primary law API.
New legal AI guides, weekly.
Further Reading
Judges Use AI, Lawyers Get Sanctioned: Is It a Double Standard?
Read postHow to Build a Litigation Chronology From Documents With AI
Read postCan AI Give Legal Advice?
Read postIs Claude Private for Legal Work?
Read postDoes ChatGPT Have Attorney-Client Privilege?
Read postCan ChatGPT Be Used Against You in Litigation?
Read post
Co-Founder & CEO · Attorney
Arshita leads product and strategy at Vaquill AI, the US primary law data API. As an attorney, she decides which law we carry, how each section is sourced and cited, and what legal AI and compliance teams need from the data.