Lawyers Apologise to US Federal Judge After Using Anthropic's Claude AI to Draft Motion Containing Fabricated Quotations in Trump Mass Layoffs Case
Three lawyers representing a former Homeland Security official have apologised to a California federal judge after partner Jason Greaves of Binnall Law Group used Anthropic's AI platform Claude Console to draft a motion that contained fabricated — or "phantom" — quotations. The incident arose in a lawsuit challenging the Trump administration's mass federal government layoffs. Judge Susan Illston of the US District Court for the Northern District of California received the apology on Friday. The case adds to a growing body of US attorney AI-misuse incidents, some of which have resulted in formal sanctions against lawyers who submitted AI-generated content containing non-existent case citations or fabricated quotes without adequate verification. The incident directly implicates professional conduct standards around the duty of candour to the court and the obligation to verify the accuracy of submissions.
Why this matters
This incident is legally significant because it demonstrates that AI hallucination risk — where a large language model (LLM) generates plausible but factually false content — is not limited to citations but extends to direct quotations, a more serious form of misrepresentation. US and UK professional conduct rules both impose a duty of candour to the court; in England and Wales, this duty sits in the SRA Code of Conduct and the Bar Standards Board's equivalent. The growing list of such incidents is building pressure on regulators and courts to issue formal guidance on AI use in litigation submissions.
On the Ground
On a litigation matter where AI drafting tools are in use, a trainee would be responsible for verifying every citation and quotation against original source documents before filing, checking court filing and service deadlines, and maintaining a chronology of submissions to identify any inaccuracies early.
Interview prep
Soundbite
AI hallucination in court filings tests whether professional conduct rules written for human error are fit for machine-generated misrepresentation.
Question you might get
“Under English professional conduct rules, what obligations does a solicitor have before submitting an AI-drafted document to the court, and what are the consequences of filing a document containing fabricated content?”
Full answer
Lawyers at Binnall Law Group apologised to a federal judge after using Claude to draft a motion containing fabricated quotations in a case about the Trump administration's mass federal layoffs. The professional conduct implications are serious: the duty of candour to the court requires that lawyers verify all representations in submissions, regardless of how they were drafted. This incident joins a growing list that is pushing US and UK regulators toward formal AI-use guidance for litigators. The trend suggests that firms deploying AI in litigation drafting will need formal verification protocols — which itself creates a new category of compliance and risk management work.
My notes
saved