AI Hallucinations in the Courtroom: The Latham & Watkins Incident and Attorney Liability

When Claude Hallucinates in Court: The Latham & Watkins Incident and What It Means for Attorney Liability

Latham & Watkins filed a court declaration in Concord Music Group v. Anthropic that contained fabricated citation details generated by Claude. The firm, which bills over $2,000 an hour, failed to catch the errors because the AI provided correct URLs but incorrect metadata.

Why This Matters

The technical reality is that LLMs like Claude can produce output that is convincingly and professionally wrong in ways that evade standard human review. While ideal models are expected to summarize accurately, this case demonstrates metadata hallucinations—where a real source is misdescribed with the wrong title or author—creating a structural verification gap that Rule 11 of the Federal Rules of Civil Procedure was not originally designed to handle.

Key Insights

• In May 2025, Latham & Watkins submitted a filing where Claude correctly linked to a URL but hallucinated the paper’s title and authors.
• Rule 11 of the Federal Rules of Civil Procedure requires attorneys to certify factual contentions, a duty that the Eastern District of Texas ruled does not transfer to AI in Gauthier v. Goodyear Tire & Rubber Co. (2024).
• According to the ABA’s 2024 Legal Technology Survey Report, nearly 75% of lawyers cite accuracy as their primary concern regarding AI adoption.
• Judge Michael Wilner of California fined a law firm $31,000 after discovering nearly one-third of brief citations were AI-fabricated.

Practical Applications

• Use case: Using LLMs for citation formatting. Pitfall: Relying on AI-generated metadata for real URLs, which can lead to metadata hallucinations that are harder to detect than phantom citations.
• Use case: Legal research and drafting by firms. Pitfall: Treating AI output as a clerical task and failing to perform metadata verification, such as checking authorship and titles line-by-line against the original masthead.

References:

• https://www.theregister.com/2025/05/15/anthopics_law_firm_blames_claude_hallucinations/
• https://www.nexlaw.ai/blog/ai-hallucination-legal-risk-2025/
• https://www.bakerbotts.com/thought-leadership/publications/2024/december/trust-but-verify-avoiding-the-perils-of-ai-hallucinations-in-court
• https://suprmind.ai/hub/ai-hallucination-rates-and-benchmarks/
• https://www.spellbook.legal/learn/why-lawyers-are-switching-to-claude
• https://www.cpomagazine.com/data-protection/2026-ai-legal-forecast-from-innovation-to-compliance/
• https://natlawreview.com/article/85-predictions-ai-and-law-2026