Attorney Jason Greaves of Binnall Law Group PLLC said Claude Console was used while drafting a filing in a Trump-era federal firings case, after fabricated citations surfaced in the record and prompted an apology to the court.
Bloomberg Law reported that Greaves told U.S. District Judge Susan Illston he used Anthropic’s Claude to help draft a motion to quash a subpoena under tight time constraints. The matter adds another example to the growing list of professional filings affected by unchecked generative AI output.
The issue matters because the mistake reached a live federal court record. In legal, financial and regulatory work, even one unverified citation can weaken the credibility of a document and create follow-on risk for the people who signed it.
The Error Acknowledged in Court Records
According to statements submitted in the case, the attorneys used Anthropic’s Claude Console while preparing the filing. Some AI-generated language was incorporated without sufficient verification, and those passages included citations and quotations that did not match the underlying authorities.
The disputed material appears in a filing on the American Federation of Government Employees, AFL-CIO v. Trump case docket. That is a stronger problem than a private drafting error because the content became part of a court-facing document.
Greaves apologized to the court and took responsibility for the accuracy of the filing. The incident underlines a basic point for professional AI use: drafting assistance does not remove the obligation to verify every source, quotation and citation before submission.
Why Claude’s Use Attracted Attention
Anthropic’s Claude Console can be used for drafting and research support, but this case shows the risk of treating model output as verified legal authority. The tool was not the central issue; the failure was allowing unverified material into a formal court filing.
For crypto, fintech and financial research teams, the lesson is broader than the lawsuit itself. AI-assisted summaries, regulatory notes and market reports still need source checks, especially when they reference filings, enforcement actions, court records or official statements.
Trust Issues Grow in Legal Proceedings
The discovery of fabricated citations could lead the court to scrutinize the reliability of the submitted documents more closely. Even where the attorneys acknowledge the mistake, the episode raises questions about supervision standards for AI tools in professional settings.
The core message is simple: AI can help draft, but the final check has to remain with humans. In law, finance and regulated markets, unsupported citations can quickly turn a workflow shortcut into a credibility problem.
This article is for informational purposes only and does not constitute legal, financial or investment advice.
