Law firm associate fired over AI-generated fake case cites (news.bloomberglaw.com)

Cozen O’Connor terminated associate Daniel Mann after he relied on artificial intelligence to generate legal citations during a Nevada state-court trial that turned out to reference cases that do not exist, the firm said in a Sept. 3 filing. The fabricated citations came while defending Reno internet provider Uprise in a contract dispute. Cozen framed the firing as a breach of professional conduct and its own policies, underscoring that lawyers must verify authorities before presenting them in court. The incident highlights a concrete, high‑stakes example of “AI hallucination” — where generative models produce plausible but false information — and its ethical and practical risks for the legal profession. For AI/ML practitioners and vendors, it underscores the need for retrieval‑augmented systems, robust source provenance, and automated citation verification to prevent fabricated outputs. For law firms and courts, it signals likely tighter controls: mandatory human verification, audit trails for AI use, updated bar guidance, and potential malpractice or sanctions exposure if AI-generated content isn’t properly vetted. The case is a cautionary data point for any domain that relies on authoritative sourcing and illustrates why model transparency and tooling to validate outputs are critical as AI tools proliferate in professional workflows.