The State Bar of Arizona issued Advisory Opinion 24-01 in June 2024, making it one of the early states to formally address AI ethics for lawyers. The opinion provides clear guidance on competence, confidentiality, billing, and disclosure — giving Arizona attorneys a concrete framework for using AI tools in practice.
Arizona's opinion is notable for its direct stance on billing: attorneys may not bill for time saved by AI at the rate of human work. This positions Arizona among the more progressive bars on the question of who benefits from AI efficiency gains — the answer is the client.
What the Bar Says
Advisory Opinion 24-01 (June 2024) addresses AI through the lens of existing ethics rules while adding AI-specific interpretation. The opinion holds that Rule 1.1 (competence) requires attorneys to understand how AI tools function before using them, including their limitations and tendency to hallucinate. Rule 5.3 (supervision) applies to AI as nonlawyer assistance — attorneys must review and verify all AI-generated work product. The opinion emphasizes that AI output is a starting point, not a final product. Attorneys bear full responsibility for accuracy.
Billing Implications
Arizona's opinion takes a clear position: attorneys may not bill for time saved by AI at the rate of human work. If AI drafts a motion in 10 minutes that would have taken an attorney 3 hours, billing 3 hours violates Rule 1.5's reasonableness requirement. The opinion directs attorneys to bill based on actual time spent reviewing, editing, and verifying AI output. AI tool subscription costs may be passed to clients as expenses with appropriate disclosure, but cannot be marked up as attorney time. This is one of the most specific billing positions any state bar has taken.
Confidentiality Rules
The opinion requires attorneys to ensure AI tools do not compromise client confidentiality under Rule 1.6. Specifically, attorneys must conduct due diligence on AI vendor data handling practices before using any tool with client information. This includes reviewing terms of service, understanding whether data is used for model training, and evaluating data storage and transmission security. The opinion stops short of banning consumer AI tools outright but makes clear that using them without vetting constitutes a confidentiality violation.
What's Still Unclear
Advisory Opinion 24-01 recommends disclosure when AI materially contributed to work product but does not mandate it. The line between "material" and "immaterial" AI contribution remains undefined. The opinion also does not address AI use in specific practice contexts like criminal defense, family law, or immigration — areas where AI errors carry heightened consequences. Arizona has not yet addressed whether AI-generated legal analysis constitutes the unauthorized practice of law when provided directly to clients. See Arizona's AI regulation landscape for broader context.
The Bottom Line: Arizona's Advisory Opinion 24-01 provides clear guidance: understand your AI tools, protect client data, bill honestly for actual attorney effort, and disclose material AI use.
AI-Assisted Research. This piece was researched and written with AI assistance, reviewed and edited by Manu Ayala. For deeper takes and the perspective behind the research, follow me on LinkedIn or email me directly.