Rhode Island AI Ethics Rules for Lawyers

The official instruments

One order, two kinds of authority

On June 17, 2026, the Rhode Island Supreme Court entered an order that must be read in two registers at once. Its first act is rulemaking: it amends comment 6 to Rule 1.1 of Article V, the competence rule, so that a lawyer’s required competence includes keeping abreast of the benefits and risks of existing and developing technology. That is binding text in the conduct rules, the comment-amendment route Colorado took in January 2026, and it is why the tracker types the order as a court rule. Its second act is advisory: the order adopts an attached set of Interim Guidelines for the Ethical Use of Generative AI, expressly without the force of law, one set written for lawyers and another for judicial officers. Chief Justice Suttell signed, with Justices Robinson, Lynch Prata, and Long. The dual character is the whole story of how to cite this instrument: the comment binds, the guidelines advise, and conflating the two is exactly the sloppiness the methodology is built to prevent.

The lawyer guidelines, rule by rule

The attorney-facing guidelines move through the conduct rules the way a careful memo would. Every generative AI output gets verified before use, the duty whose breach fills the sanctions survey. Fees follow actual time, and generative AI costs sit in overhead rather than on the bill, unless the client has requested specialty tools, the same efficiency-belongs-to-the-client logic the billing analysis traces nationally. Confidentiality starts before first use: a tool’s confidentiality terms get vetted, the vendor-diligence step the Rule 1.6 breakdown walks through. Disclosure to the client has two triggers, when the client asks and when client information is input into the tool. And the guidelines reach a subject most states leave unspoken, directing lawyers to weigh Rule 8.3’s reporting duty where opposing counsel’s generative AI misconduct rises to the substantial level that rule addresses. For everything else they point lawyers to the court’s Ethics Advisory Panel, which, as of our July 1, 2026 check, has issued no AI opinion of its own, a gap worth knowing when the guidelines send you there.

The bench’s own lines, and the committee behind the order

The judicial-officer guidelines are shorter and harder-edged: generative AI must never undermine or replace independent judicial decision-making, and judges should never rely on GAI-produced content when drafting orders or decisions. A no-drafting position for the bench is among the firmest any state has committed to writing, stricter in kind than the accountability-only postures elsewhere on the states index. The machinery that produced all of this is worth recording as activity. Rhode Island created a Committee on Artificial Intelligence and the Courts by executive order on October 22, 2024, co-chaired by Justice Lynch Prata and Judge Stern, per the judiciary’s own press release; secondary trackers report the committee’s work extended by a further executive order of September 3, 2025. A committee is process rather than authority, so it stays out of the instrument count, but the June 2026 order shows the process producing law.

Practicing under an interim framework

Interim is a promise of revision, so Rhode Island lawyers should treat this page’s instrument as version one. The duties beneath it do not wait: supervision of AI use across an office per the supervision rules, a written policy per the policy template, and the national reasoning of ABA Formal Opinion 512, which the guidelines’ rule-by-rule structure resembles. Individual judges can add case-level requirements notwithstanding the statewide framework, the check the court orders guide covers, and the changelog will record the day the interim guidelines become final ones, or the comment amendment gains company in the black-letter rules.