D. Nev.: Second Amended Standing Order and Chambers Practices, Section III (Artificial In…

What the rule requires

Magistrate Judge Couvillier’s Second Amended Standing Order, effective August 17, 2026, keeps the four-part AI section that distinguishes it from the binary use/no-use chambers rules common on the federal trial bench. The order permits AI use for drafting and discovery processing, then layers on four obligations: (1) a certification, above the signature line, on any filing that includes AI-generated or AI-drafted content, attesting both to the use and to the filer’s review and verification of that content under Rule 11; (2) corroboration of all citations to authority or assertions of fact, under Fed. R. Civ. P. 11 and LR IA 7-3, with failure to check AI-sourced citations or assertions of fact grounds for potential sanctions; (3) a strong recommendation that protective orders address AI protocols (closed versus open systems, discovery-processing limits for personal-data identifiers and trade secrets); and (4) a prohibition on uploading filed documents that are sealed or contain personal-data identifiers to open AI systems.

The fourth provision is the doctrinally distinctive part. Most federal chambers AI rules focus on hallucination risk in citations; Couvillier’s rule extends to data-leakage risk in confidential filings. Counsel may not upload to an open AI system (one “where the source code, learning materials, and underling [sic] models may be publicly accessible”) any filed document that is sealed or that contains the LR IC 6-1(a)(1)-(6) personal-data identifiers. The prohibition reaches only those filed documents; other confidential client material is addressed through the protective-order recommendation rather than a flat ban.

The protective-order recommendation (Provision 3) is the practitioner-facing innovation. By signaling that the court expects parties to address AI in protective orders, Couvillier is shifting the default expectation from “AI is silently allowed” to “AI use is governed by the protective order, including system-type and discovery-processing limits.”

Revision history

The First Amended order (effective January 5, 2026) required every filing to carry a binary certification, “I certify that Artificial Intelligence was [used / not used] to prepare the foregoing document,” immediately above the signature line. The Second Amended order narrows the certification to filings that actually include AI-generated content and folds the verification attestation into the certification itself, so a filing that used AI now certifies both the use and the review in one statement, while a filing that did not use AI carries no AI statement at all. Two smaller changes: the permission clause now reads “counsel or party” rather than “counsel or pro se party”, and the citation provision now reaches authorities “obtained from an AI tool” as well as those created by one.

R&G data corrections

Ropes & Gray’s tracker dates this entry 2026-01-05, the effective date of the First Amended order. The court’s current document is the Second Amended order effective 2026-08-17, linked from the judge’s page on nvd.uscourts.gov; the January 2026 version is no longer hosted there.

Quotable language

“Any counsel or party may use AI to draft documents or process discovery consistent with Rule 11 of the Federal Rules of Civil Procedure (Fed. R. Civ. P.), this Standing Order, and any other applicable legal or ethical obligations.”

“Any party that files (manually or e-filing) a brief, motion, request, pleading, or other similar document with the Court which includes content generated or drafted by artificial intelligence (i.e., ChatGPT, Harvey, CoCounsel, Google Gemini, or any other such application) must include the following certification above the signature line: ‘I certify that artificial intelligence was used to generate a portion of this filing. I have reviewed the source material and verified that the artificially generated content in the filing is accurate and complies with the filer’s Rule 11 obligations.’ Failure to include such certification may result in sanctions.”

“A pro se party or a counsel’s failure to confirm or double-check the accuracy, veracity, or existence of a case or legal citation (or assertion of fact) obtained from an AI tool or created by an AI tool is grounds for potential sanctions.”

“Counsel and parties are prohibited from uploading onto any open AI system - where the source code, learning materials, and underling [sic] models may be publicly accessible - any filed documents which: (a) have been filed under seal; or (b) contain personal-data identifiers listed in LR IC 6-1(a)(1)-(6). Violation may result in sanctions.”

Primary source

Second Amended Standing Order and Chambers Practices (Couvillier, M.J.), effective August 17, 2026, nvd.uscourts.gov, linked as “Standing Order” from the judge’s page.