Vermont AI Ethics Rules for Lawyers

The official instruments

The bench moved before the bar

In Vermont the generative AI guidance came from the court system’s side of the profession. The Vermont Supreme Court created a dedicated body, the Vermont Judiciary Committee on Artificial Intelligence and the Courts, and that committee’s First Annual Report, published in March 2025, is the state’s operative instrument; the material addressed to lawyers sits in its Appendix G. Vermont has no bar ethics opinion on AI, so the report is not a supplement to bar authority, it is the authority, and it is a judiciary document rather than bar guidance. Our dataset types it as a report at month precision, and this page follows both choices: no invented day, no inflated label.

The committee’s name repays a second read. “First Annual Report” describes a body built to keep watching, not a task force that answered one question and dissolved. For a technology that changes quarterly, a standing committee with a reporting rhythm is itself a policy decision, and it means Vermont’s position has a scheduled opportunity to evolve. Successor reports are the concrete thing to watch for in this state, and the changelog will record one as soon as it verifies.

Deciding not to legislate, and what got issued instead

The report’s structural holding is restraint: no amendments to the conduct rules are warranted, because generative AI use falls within existing rules. Vermont looked at the same technology that pushed other jurisdictions toward new text and concluded its rulebook already reached it. What issued instead is five-point guidance, and the points are notably concrete. Keep abreast of AI risks and benefits, a continuing competence expectation rather than a one-time briefing. Review a tool’s terms of service and understand how confidential information is used or retained, which turns confidentiality from an abstraction into a reading assignment aimed at the vendor’s own documents. Never bill hourly clients for time saved, as unqualified a statement of the efficiency-belongs-to-the-client principle as any instrument in the dataset makes; the billing analysis shows how that principle runs through other states’ fee holdings. Confirm output accuracy, the verification duty that stands between a helpful draft and a sanctionable filing. And consider an office AI-use policy as part of supervision duties, which folds AI governance into the management obligations a practice already carries.

Read together, the five points assume a small-office audience, terms-of-service review and an office policy are things a two-lawyer firm can actually do this month, and Vermont practice skews small; the solo and small-firm starter sequences the same moves for exactly that setting.

Working under a judiciary report

For a Vermont lawyer the compliance posture is existing rules plus Appendix G. The office policy the report recommends is the natural container for the rest of it: tool vetting records that show terms-of-service review, a verification step before AI-assisted work is relied on, billing practices that never charge for saved time, and training that keeps the risks-and-benefits knowledge current. The policy template provides that container, and ABA Formal Opinion 512 supplies the fuller national reasoning behind each duty the five points compress.

One distinction keeps the report in focus: it is guidance from a judiciary committee, not a filing rule. Nothing in our dataset imposes a Vermont-wide certification or disclosure requirement for AI-assisted court papers, so the filing layer works the way the court orders guide describes everywhere, through general candor duties and the standing orders of the individual judge, confirmed before each submission. And because the issuing committee reports annually, Vermont is less a settled jurisdiction than a scheduled one; the tracker will keep its entry synchronized with what the committee publishes next.