Headline: AI “Hallucination” in Trump Layoffs Case — Lawyers Apologize After Claude-Generated Fake Quotes Slip Into Brief
A federal lawsuit tied to the Trump administration’s mass federal layoffs has turned into the latest courtroom example of AI-generated errors creeping into legal filings. In a declaration filed Friday, attorney Jason Greaves acknowledged that a motion he filed included fabricated quotations produced by Anthropic’s Claude Console.
Greaves said he used the enterprise version of Anthropic’s AI to draft the motion under tight time pressure. He told the court he sent that AI draft to an associate the morning of May 6 with explicit oral instructions that the text had originated from AI and that citations needed careful verification. The associate later reported she had reviewed and “verified” each citation, though she noted two cases cited were incorrect and said she substituted a different case in their place.
“As the supervising partner, and the signer of the pleading, the responsibility for having accurate citations is entirely on me,” Greaves wrote, taking full responsibility and apologizing “to the court and to all counsel in this case, for the phantom quotations that appeared in the motion.” He said he has personally apologized to the plaintiffs’ attorneys.
Anthropic’s Claude Console—originally launched in March 2023 as the Anthropic Developer Console—is an enterprise platform that businesses use to draft, summarize and research documents with Claude AI models. The incident highlights the risks of relying on large language models for legal work without exhaustive human verification.
This episode follows a similar April incident in which Sullivan & Cromwell lawyers were sanctioned after submitting fake legal citations generated by AI. The current litigation involves suits challenging the Trump administration’s staffing cuts and proposed changes to civil service protections, including disputes over FEMA staffing and testimony from former Homeland Security official Joseph Guy; Guy’s lawyers filed the motion at issue.
For crypto and other regulated industries that increasingly use AI tools for research, drafting, and compliance, the case is a reminder: AI can speed work but also introduce hazardous errors into high-stakes documents. Legal teams and compliance officers will need tighter verification workflows to avoid similar “hallucinations” making their way into court filings or regulatory submissions.
Read more AI-generated news on: undefined/news
