Lawyers have been sanctioned for filing AI-invented citations — the cautionary tale of the profession. Used with verification discipline, though, AI is transforming review, research, and drafting economics.
Why: Models fabricate convincing case law — right reporter format, plausible parties, invented holdings. Multiple attorneys have been sanctioned; every incident started with skipping this step.
Why: Confidentiality duties don't bend for convenience, and privilege waiver arguments over third-party AI processing are now live litigation issues.
Why: The economics are real — review hours drop dramatically — but the license, the malpractice exposure, and the duty of candor all attach to the human signature.
Why: Courts diverge — some require certification, some ban generative drafting in filings, most say nothing yet. The register costs minutes; being wrong in front of a judge costs credibility.
Why: Ethics rules make supervising lawyers responsible for nonlawyer assistance — which now includes software. A written, trained workflow is the defensible position.
The fake-citation sanctions cases did the profession a service: they made the failure mode vivid early. The lesson was never “don’t use AI” — it was that generative output is a draft from an unlicensed assistant with a known fabrication rate, and everything downstream follows from treating it that way.
Meanwhile the quiet transformation is in volume work. Document review, due-diligence extraction, and first-draft generation from firm templates now run at a fraction of their former hours at firms that built the verification workflow. The competitive gap isn’t between firms that use AI and firms that don’t — it’s between firms with a supervised process and firms with improvisation.