Arizona AI Ethics Rules for Lawyers

The official instruments

Guidance from the court, not the bar

Arizona is easy to misfile, so start with the label. The state’s generative AI instrument is “Generative AI: Ethical Best Practices for Lawyers and Judges,” published November 14, 2024. It is guidance, not a numbered ethics opinion, and the body that produced it is not the State Bar’s ethics committee. It is the Arizona Supreme Court’s Steering Committee on Artificial Intelligence and the Courts, the AISC, established by Administrative Order 2024-33. The State Bar of Arizona hosts and published the document, which is why it is sometimes cited as a bar product, but its authorship sits on the court-administration side. For a tracker whose whole premise is honest labeling, that provenance is the first thing worth getting right: cite it as steering-committee guidance published via the bar, not as an Arizona bar ethics opinion, and not as a binding rule.

The format follows from the source. Rather than reasoning to a holding the way a formal opinion does, the document sets out best practices across a defined set of areas, which makes it read more like a framework than an adjudication. That is useful for firm-policy drafting, because the areas map cleanly onto policy sections, but it also means the instrument carries the persuasive weight of guidance rather than the citable weight of a formal opinion. Arizona lawyers relying on it should understand both halves of that trade.

One document for both sides of the bench

The distinctive feature of the Arizona guidance is its audience. Because the AISC is a court committee, its “Ethical Best Practices” speak to judges as well as lawyers, in a single instrument. Most state AI instruments are written for practitioners and leave judicial conduct to separate channels; Arizona articulates expectations for both at once. Lead with that when you describe the document, because it is what genuinely sets Arizona apart from the bar-issued opinions and guidance in neighboring jurisdictions. It signals that, in Arizona, the AI conversation is being run from the courts outward rather than from the bar inward.

For practicing lawyers, the substance lands across seven areas. Competence and diligence: understand the tool, and check every material proposition and source in its output, because generative AI can produce confident, fabricated text. Confidentiality: do not enter client or other nonpublic information into tools that lack sufficient guarantees about handling, and obtain informed consent otherwise. Communication: disclose chatbots to clients and supervisors where appropriate, without a duty to flag every routine AI use. Billing: do not charge for time saved by AI, and disclose AI costs in writing in advance. Court disclosure and candor: do not let AI output reach a tribunal unverified. Bias mitigation: account for the possibility of biased outputs. And supervision: adopt a written, acknowledged firm AI policy. The policy template implements the supervision and confidentiality pieces directly, and the state comparison table places Arizona’s consent posture among the states.

What Arizona lawyers should do now

Treat the seven areas as a checklist with artifacts behind each item. Maintain an approved-tools register that records the confidentiality guarantees behind every tool you permit, which is how you evidence the “sufficient guarantees” standard. Default to anonymizing inputs, and capture informed consent where residual risk remains. Build a verification workflow that treats “check every material proposition and source” as a hard gate before anything is filed or sent. Put your AI-cost disclosure into the engagement letter so the billing rule is satisfied at intake, not in arrears. And adopt the written, acknowledged firm policy the guidance calls for, because in Arizona the supervision expectation is stated in terms of exactly that document. This site’s home page tracks Arizona with the honest “guidance” label, and the changelog will record any move from guidance toward a binding rule.

One layer the guidance frames but does not resolve at the filing level: the courtroom. Even though Arizona’s instrument speaks to judges, there is no single statewide court rule dictating AI disclosure in every Arizona filing, so the obligations that bite on a given case are still driven case by case, including judge-by-judge standing orders, particularly in the federal districts. The court orders guide covers the per-filing check that Arizona practice still requires, and the Ropes & Gray tracker is the place to confirm a specific judge’s order before you file.