What happened in the Claude AI court filing controversy?

Lawyers involved in a federal layoffs case admitted submitting fabricated quotations generated by Anthropic’s Claude AI system. The fake citations appeared in a court filing connected to disputes over Trump administration layoffs and FEMA staffing changes.

Who admitted using Claude AI in the legal filing?

Attorney Jason Greaves said he used Anthropic’s Claude Console to help create an initial draft of the motion under tight deadlines. He later apologised to the court after “phantom quotations” remained in the final filing.

What are AI hallucinations in legal documents?

AI hallucinations happen when artificial intelligence systems generate false information that appears believable but is inaccurate or entirely fabricated. In legal work, this can include fake case citations, invented quotations or nonexistent court rulings.

Why is this case important for AI adoption in law?

The incident highlights growing risks as lawyers increasingly use AI tools for drafting, research and document preparation. Courts and legal professionals are becoming more concerned about whether AI-generated work is being properly verified before filing.

Did the lawyers verify the AI-generated citations?

Greaves said he instructed an associate attorney to carefully verify all citations before filing the motion. However, some fabricated quotations still made it into the final submission.

Could lawyers face penalties for using fake AI-generated citations?

Yes. Courts can sanction attorneys for submitting inaccurate or fabricated legal materials regardless of whether AI generated them. Lawyers remain legally responsible for all documents they sign and file.

Has this happened before with AI-generated legal filings?

Yes. Earlier this year, lawyers at Sullivan & Cromwell were sanctioned after submitting fake legal citations generated by artificial intelligence. Similar incidents have emerged globally as AI adoption expands in legal services.

Image for illustrative purposes only. Not a real photo.

Lawyers apologise after Claude-generated fake quotes filed in court

Written by Brie CarterPublished May 19, 2026, 9:00 AM

Lawyers involved in a federal court dispute over Trump administration layoffs apologised after submitting fabricated quotations generated by Anthropic’s Claude AI system in a legal filing.

Attorney Jason Greaves admitted in a court declaration that he used Anthropic’s Claude Console to help draft a motion tied to a subpoena dispute involving FEMA layoffs and testimony from former Homeland Security official Joseph Guy.

“Given the tight time constraints I used an enterprise level, data-isolated AI platform, Claude Console, to create an initial draft of the motion,”

Greaves wrote, adding:

“That was a mistake.”

Greaves said he instructed an associate attorney to verify the filing’s citations after the AI-generated draft was prepared, but “phantom quotations” still remained in the final court submission.

“As the supervising partner, and the signer of the pleading, the responsibility for having accurate citations is entirely on me,”

Greaves added.

“I take full responsibility for that and apologise fully to the court and to all counsel in this case.”

The controversy adds to a growing number of courtroom disputes involving AI hallucinations, including earlier sanctions issued against lawyers at Sullivan & Cromwell after submitting fabricated legal citations generated through artificial intelligence.

Claude Console, launched by Anthropic in 2023, is an enterprise platform that allows businesses and professionals to use Claude AI models for drafting, research, coding and document analysis, as concerns continue growing over the reliability of AI-generated legal and professional content.