San Francisco federal judge orders sworn explanation after a lawyer files a motion he says Claude Console drafted with three phantom quotations
In the federal lawsuit over the Trump administration's mass layoffs (AFGE v. Trump, Northern District of California), a partner at Binnall Law Group filed an emergency motion on the night of 6 May 2026 to quash a deposition subpoena directed at a former Homeland Security official the firm represents. In a declaration filed on 15 May under the court's 13 May order, the partner said he had used Claude Console, an enterprise AI platform, to create an initial draft, that an associate told him she had checked the citations, and that he did not double-check the quotations. He was explaining under oath, at the court's instruction, how three phantom quotations ended up in the motion. He apologised to the court and opposing counsel and attached a corrected version. The firm's founding partner declared that firm policy prohibits using tools like Claude for legal drafting, that the error breached its policies and training, and that the firm had opened an internal investigation. Whether Judge Susan Illston imposed sanctions is not reported in the sources read.
- AI system
- Claude Console (reported)
- Occurred
- 6 May 2026
- Reported
- 15 May 2026
- Event location
- United States
- What the AI did
- Communicated with the person
- Reported harm
- Legal HarmReputational Harm
- Whose AI use
- Their own AI use
- Setting
- Work · Justice
- Evidence
- AI involvement supported · Causal attribution supported · 2 sources, 1 underlying account
- 4 claims: 3 documented, 1 reported. 3 open questions
- People reported harmed
- At least 1 person
AI system as recorded: Claude Console, described by the filing partner as an enterprise-level, data-isolated AI platform, used to create the initial draft of the motion
What Happened
The motion. On the night of 6 May 2026, after the court ordered a former Homeland Security official to appear for a further half-day of deposition, a partner at Binnall Law Group filed an emergency motion to quash the subpoena. His declaration says that, under tight time constraints, he used Claude Console to create an initial draft, sent it to an associate with instructions to check the citations, and filed it that night without the firm's usual final check. He says he reviewed the cited cases generally but did not double-check the quotations or read the cases carefully, and that this was the first time he had used an AI platform to draft a brief.
The court's response. The firm's notice says the declarations were filed under the court's 13 May 2026 order. The partner's declaration says it responds to the court's instruction that he explain under oath how three phantom quotations ended up in the motion. He took full responsibility, apologised to the court and all counsel, said he had personally apologised to the plaintiffs' attorneys, and attached a redlined version of the motion as it should have been filed.
The firm. The founding partner's declaration says firm policy allows Claude for certain work but prohibits tools like Claude for legal research and drafting, requires manual verification of all citations in AI-generated content, and that the filing breached both policy and training at several levels of review. The firm opened an internal investigation.
Limits. The court's 13 May order and any later ruling were not read. Bloomberg Government's report relays the same declarations.
Reported harm
A lawyer whose AI-drafted motion contained three phantom quotations was required by the court to explain under oath how they got there, and he and his firm filed sworn apologies and a corrected motion; the firm says the filing breached its own AI policy.
Outcome
UnknownThe court's 13 May 2026 order (ECF No. 395) led to sworn declarations from three Binnall Law Group lawyers, filed on 15 May 2026 with a corrected version of the motion. No ruling on sanctions is reported in the sources read.
What remains unknown
- Whether the court imposed sanctions or other measures after the declarations.
- The text of the court's 13 May 2026 order.
- Which phantom quotations came from the AI draft and which, if any, from later edits.
What the evidence supports
AI involvement: supported. The partner declared under penalty of perjury that he used Claude Console to create the initial draft of the motion and did not double-check its quotations; the motion as filed contained three phantom quotations, which the court required him to explain under oath. His own account connects the AI-generated draft to the quotations the court questioned. Whether the specific quotations came from the AI draft rather than later editing is not separately stated.
4 claims: 3 documented, 1 reported. What the statuses mean
Documented In a sworn declaration filed on 15 May 2026, the partner who filed the 6 May motion stated that he used Claude Console to create an initial draft of it, that he did not double-check the quotations, and that this was the first time he had used an AI platform to draft a brief.
Causal attribution. The partner's own sworn declaration.
- fingfx.thomsonreuters.com(opens in new tab) supports · English
Claude Console, to create an initial draft of the motion. That was a mistake. ... I did not double check the ... quotations and obviously did not read the cited cases carefully ... This was the first time I have ever used an AI platform to create a draft of a brief
- news.bgov.com(opens in new tab) supports · English
told Illston in a declaration that he sent the AI-drafted document to an associate for fact-checking.
Documented The declarations were filed under the court's 13 May 2026 order, and the partner's declaration responds to the court's instruction that he explain under oath how three phantom quotations ended up in the motion.
Causal attribution. The firm's notice and the partner's declaration.
- fingfx.thomsonreuters.com(opens in new tab) supports · English
pursuant to the Court’s May 13, 2026 Order, ECF No. 395 ... This declaration is in response to the Court’s instruction that I explain, under oath, ... how three phantom quotations ended up in the Emergency Motion to Quash
Documented The partner apologised to the court and counsel and filed a redlined corrected version, and the firm's founding partner declared that firm policy prohibits tools like Claude for legal drafting and that the firm opened an internal investigation; the partner declared that the internal consequences of the failure were under review by the firm's principal.
Causal attribution. Declarations of the partner and the founding partner.
- fingfx.thomsonreuters.com(opens in new tab) supports · English
have also personally apologized to the Plaintiffs’ attorneys. ... A redlined version of the filing, ... Those same policies prohibit the use of tools like Claude for legal research ... internal investigation to understand fully how the error occurred ... The internal consequences for this failure are under ... the review and consideration of our principal
Reported The motion was filed in the lawsuit over the Trump administration's mass layoffs, on behalf of a former Homeland Security official, and three of the firm's lawyers apologised to the judge.
Causal attribution. Bloomberg Government.
- news.bgov.com(opens in new tab) supports · English
Three lawyers representing an ex-Homeland Security official apologized to a California federal judge ... sprawling lawsuit over the Trump administration’s mass layoffs
Sources
2 sources inspected, from 1 underlying account. Sources that repeat one account do not corroborate each other.
- Notice of declarations from counsel, AFGE v. Trump, N.D. Cal., ECF No. 399, filed 15 May 2026(opens in new tab)
s1 · fingfx.thomsonreuters.com · Court filing · English · Inspected · 15 May 2026 · Shares an underlying account with another listed source · Primary
- Bloomberg Government, 16 May 2026: Lawyers Apologize for Fake AI Quotes in Trump Mass Layoffs Case(opens in new tab)
s2 · news.bgov.com · News report · English · Inspected · Shares an underlying account with another listed source
How the sources were read, and where the events happened
Read in English by the research agent (an AI) on 2026-10-18 as text extracted from the PDF (HTTP 200); no translation was involved and no human reviewer read it. ECF No. 399 with exhibits, filed 15 May 2026 in Case No. 3:25-cv-03698-SI: the firm's notice, the declarations of three firm lawyers and the redlined motion. The court's 13 May order was not inspected. Applies to s1.
Read in English by the research agent (an AI) on 2026-10-18 (HTTP 200); no translation was involved. Bloomberg Government, 16 May 2026. It reports the same declarations and is not independent of the filing. Applies to s2.
Event countries: United States. Affected-person countries: Unknown. Court countries: United States.
The motion was filed in, and the declarations answer an order of, the US District Court for the Northern District of California (case 3:25-cv-03698-SI), after first being filed in the Eastern District of Virginia. The partner's country of residence is not stated.
Reviewed for publication 2026-10-18: Published as a core case (communicated_with, own use): the filing lawyer's sworn declaration says he used Claude Console to draft a motion that contained three phantom quotations, and the court ordered him to explain them under oath; he and his firm filed apologies and a corrected motion. No sanction is reported in the sources read.
People described
A partner at Binnall Law Group, the lawyer who drafted, signed and filed the motion, ordered to explain the phantom quotations under oath (not named here)
People reported harmed in this case
At least 1 person
1 AI participant · 0 other people harmed
One person is counted: the partner who used Claude Console, filed the motion and was instructed to explain it under oath. Two other firm lawyers also filed declarations, but no consequence for them is described, so the count is a lower bound. The client and opposing counsel are not counted.
Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.
Cite this case
Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.
APA
NOPE. (2026). San Francisco federal judge orders sworn explanation after a lawyer files a motion he says Claude Console drafted with three phantom quotations. AI incidents. https://nope.net/incidents/2026-san-francisco-federal-judge-orders-sworn-explanation-after-lawyer-files-claude-drafted-motion-with-phantom-quotations
BibTeX
@misc{2026_san_francisco_federal_judge_orders_sworn_explanation_after_lawyer_files_claude_drafted_motion_with_phantom_quotations,
title = {San Francisco federal judge orders sworn explanation after a lawyer files a motion he says Claude Console drafted with three phantom quotations},
author = {NOPE},
year = {2026},
howpublished = {AI incidents},
url = {https://nope.net/incidents/2026-san-francisco-federal-judge-orders-sworn-explanation-after-lawyer-files-claude-drafted-motion-with-phantom-quotations}
} Related cases
Former federal prosecutor in North Carolina publicly reprimanded after admitting that a brief with fabricated quotations was drafted with AI; he lost his post
In December 2025 an Assistant U.S. Attorney in the Eastern District of North Carolina filed a response brief for the Department of Defense in a self-represented veteran's lawsuit over health-insurance coverage. The plaintiff identified fabricated quotations and misdescribed case holdings in it. The attorney first attributed the errors to an unfinalized draft filed by mistake. On 2 March 2026 Magistrate Judge Robert T. Numbers II ordered him to show cause why he should not be sanctioned. At the 10 March hearing the attorney, Rudy Renfer, said under oath that he had accidentally overwritten his draft, panicked and used artificial intelligence to draft the brief, and had not verified the cases it cited; he said he had resigned after 17 years in the office, and Courthouse News reported he was terminated the next day. On 28 April 2026 the court issued a public reprimand instead of a fine, writing that his reputation was in tatters and that his job loss already exceeded the usual fines.
St. Clair v. xAI: a writer says Grok made and published sexualized images of her on X, including from a photo of her at 14, after it told her it would not; xAI sued her in Texas and her New York case was ordered transferred there, an order the Second Circuit has stayed pending her mandamus petition
Ashley St. Clair, a writer and political commentator with about 1 million followers on X, sued xAI in New York on 15 January 2026. She alleges that from on or about 4 January 2026 Grok, prompted by other X users, altered photos of her into sexualized images and published them on X, including one made from a photo of her at 14. She says @Grok told her that her images would not be used or altered without explicit consent, but many more images followed. She also says X's reporting channel found no violation and that X removed her Premium subscription, verification checkmark and monetization. She says she has suffered serious pain, mental distress and humiliation. The same day, xAI sued her in federal court in Texas, arguing that its terms of service required her to sue there. On 23 June 2026 the New York federal court ordered her case transferred to Texas. On 8 July the Second Circuit stayed that order administratively while it considers her mandamus petition, which was still pending on 24 August. The Texas case is set for trial on 1 March 2027.
Canary Islands high court fines a lawyer 420 euros for an appeal citing 48 non-existent judgments that it says a general-purpose AI tool generated
The Criminal Chamber of the High Court of Justice of the Canary Islands (TSJC) fined a lawyer 420 euros for breaching the duty of truthfulness and good faith, after an appeal the lawyer filed contained 48 citations of Supreme Court judgments and a report of the General Council of the Judiciary that did not exist. According to the court's order, as described in its press release of 13 February 2026, the lawyer used a general-purpose AI tool to draft the appeal and did not check a single citation against the free public case-law database; the court's documentalist verified that the references were false. The court referred the matter to the lawyer's bar association for possible disciplinary action and set the fine at about half the annual cost of a legal-specific AI tool, taking into account that the lawyer admitted the facts and expressed remorse. The court had first flagged the citations in its judgment dismissing the appeal, reported on 5 January 2026, which opened a separate proceeding. The lawyer told the court of comments and mockery from colleagues after the case became public.
Miami: scammers cloned immigration attorney Ángel Leal's face and voice with AI to defraud immigrants seeking legal help
From around October 2025, criminals used AI to clone the image, voice, name, signature and firm logo of Ángel Leal, a long-established immigration attorney in Doral, Florida. They created fake videos, look-alike websites, forged contracts and even fabricated Zoom 'immigration hearings' to persuade immigrants to pay for legal services Leal never provided. Immigrants were defrauded of money, and some believed their cases were progressing when they were not; Leal received one to two defrauded callers a day, changed his phone number, hired an anti-piracy firm that removed thousands of fake profiles and videos, and reported the fraud to local police and federal authorities.
If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.