The N.J. Supreme Court's Preliminary Guidelines treat AI-output verification as an RPC duty.

The Supreme Court of New Jersey issued its Preliminary Guidelines on the Use of Artificial Intelligence by New Jersey Lawyers via Notice to the Bar on January 24, 2024. The Guidelines do not create new RPCs; they map existing Rules of Professional Conduct onto generative-AI use. Two operational additions to the ABA Opinion 512 baseline matter for New Jersey-licensed lawyers and any matter venued in a New Jersey tribunal.

First, verification of AI output framed as an RPC duty. The Guidelines state: "Because AI can generate false information, a lawyer has an ethical duty to check and verify all information generated by AI to ensure that it is accurate. A lawyer who uses AI in the preparation of legal pleadings, arguments, or evidence remains responsible to ensure the validity of those submissions." [S12] This is a stricter verification rule than ABA Op. 512's competence-and-candor framing — New Jersey treats the verification step itself as the operative ethical duty, with the lawyer of record on the hook for the validity of any submission containing AI-derived material.

Second, disclosure not required, but not an excuse either. The Guidelines hold: "While the RPCs do not require a lawyer to disclose the use of AI, such use does not provide an excuse for the submission of false, fake, or misleading content." [S12] The disclosure default is narrower than NYSBA's disclosure-forward stance — New Jersey does not impose an affirmative duty to tell the client or the tribunal that AI was used — but the verification rule is stricter than the ABA baseline. The combined posture is: documentation is the safe harbor, not disclosure.

The operational consequence for matter-management is that the matter file, not the engagement letter, carries the heavier load. The engagement-letter clause library still applies; the Preliminary Guidelines push the documentation burden into the per-filing record.