A California judge recently took a firm stance against the undisclosed use of AI platforms by lawyers in legal filings. Judge Michael Wilner discovered a supplemental brief riddled with fabricated legal citations and quotations. His response? Sanctions of $31,000 levied against the two law firms involved.
Judge Wilner expressed his concern regarding such practices. He stated that relying on AI for legal research and writing without proper oversight is a significant mistake for any competent attorney. Legal experts Eric Goldman and Blake Reid also highlighted the issue on Bluesky.
Lawyers sanctioned for undisclosed AI use in legal research
The judge recounted his experience, noting that the cited cases initially seemed persuasive, leading him to investigate further. It was then he made these decisions simply didn’t exist. He emphasized the potential danger, explaining how easily such false information could have found its way into a judicial order.
Basically, this is another case where “AI hallucinations” reveal the use of chatbots. A similar situation occurred in late 2024 in an—ironically—anti-AI case.
The situation unfolded when a plaintiff’s legal team used AI to draft an outline for a brief against State Farm. This outline, containing what the judge termed “bogus AI-generated research,” was then passed on to another firm, K&L Gates, which incorporated it into the final brief. Shockingly, neither firm apparently verified the AI-generated information before submitting it to the court.
Judge identified AI hallucinations thanks to his experience
Upon review, Judge Wilner identified at least two entirely nonexistent legal authorities. When he requested clarification from K&L Gates, the revised brief they submitted contained even more fabricated citations and quotations. This prompted the judge to issue an order requiring the lawyers to explain their actions under oath. This ultimately confirmed the use of AI. The attorney who created the initial outline admitted to using Google Gemini and the AI legal research tools within Westlaw Precision with CoCounsel.
Judge Wilner concluded that the initial, undisclosed use of AI to draft the brief was fundamentally wrong. He further stressed the risk involved in passing on AI-generated material without revealing its origin, potentially misleading other legal professionals.
2025-05-14 15:04:54