Answers are grounded in this paper and cite their sources.
In 2023, a lawyer filed a brief citing six cases that did not exist. They had been invented by ChatGPT. The court fined the lawyers $5,000 — not for using AI, but for failing to check. This essay asks who is lying when an AI hallucinates a citation: the machine, or the lawyer who signed the filing? Drawing on Mata v. Avianca, two Stanford studies of legal hallucinations, a Northwestern survey of 112 federal judges, ABA Formal Opinion 512, and California SB 574, it argues that an AI cannot lie, because lying requires intent. The lawyer who signs the filing is the one who answers for it. The piece closes with a practical rule: use AI to draft, never to source — and verify every citation. Previously published on Substack (Past Perfect Future Tense), November 8, 2026: https://cnewhall.substack.com/p/who-is-lying-the-lawyer-or-the-ai
No takes yet. Share an insight, caveat, or question.
Stanley Clark Newhall (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: