Levidow, Levidow & Oberman

Attorneys at Levidow, Levidow & Oberman needed case-law precedent to oppose Avianca Airlines' motion to dismiss a personal injury suit brought by client Roberto Mata. Attorney Steven Schwartz, with 30+ years of experience but no prior AI research tools training, turned to ChatGPT to "supplement" the legal research, treating it as a search engine and assuming it would only surface real, existing cases.

Schwartz used ChatGPT to find and draft citations supporting the brief, and when he had doubts, asked ChatGPT directly whether the cases were real — it falsely confirmed they were genuine. Neither Schwartz nor co-counsel Peter LoDuca cross-checked the citations against legal databases like Westlaw or LexisNexis before LoDuca signed and filed the brief.

All six cases cited in the brief were entirely fabricated by ChatGPT, complete with fake docket numbers, judges, quotes, and legal reasoning. Judge Kevin Castel found the attorneys acted in subjective bad faith and sanctioned Schwartz and LoDuca $5,000 each, and separately dismissed Mata's underlying lawsuit as time-barred under the Montreal Convention.

• Generative AI can produce entirely fabricated, plausible-sounding legal citations ("hallucinations") that pass a casual read and mimic real citation formatting.
• AI-generated research output must be independently verified against authoritative legal databases before it is relied on or filed.
• Courts will impose direct financial sanctions on attorneys for negligent or bad-faith reliance on unverified AI-generated content.