The State Bar of California published its Practical Guidance for the Use of Generative AI in November 2024, establishing one of the most comprehensive AI ethics frameworks in the country. California's guidance goes beyond a single opinion — it functions as a practical manual for attorneys integrating AI tools into their practice.
As the largest state bar in the U.S. with over 270,000 active members, California's position carries outsized influence. The guidance addresses billing, confidentiality, competence, disclosure, and supervision with specificity that smaller state bars are already using as a template. If your state has not issued AI guidance, California's framework is the de facto standard.
What the Bar Says
California's Practical Guidance for the Use of Generative AI (November 2024) treats AI as a transformative technology requiring updated interpretation of existing rules. Rule 1.1 (competence) now demands that attorneys understand how generative AI works, its limitations, and its tendency to produce hallucinated citations. Rule 5.3 (supervision) requires that all AI-generated work product be reviewed with the same rigor applied to junior associate work. The guidance explicitly states that "I relied on AI" is not a defense to filing inaccurate information. Attorneys are the final checkpoint.
Billing Implications
California requires that AI-assisted work reflect actual value under Rule 1.5 (reasonable fees). The guidance does not ban billing for AI-assisted work but requires that fees correspond to the value delivered and the effort expended. Passing AI tool costs to clients requires transparency and must be reasonable. The guidance encourages value-based billing models as AI reshapes the time-effort equation. Firms billing purely on hourly rates face increasing scrutiny when AI reduces task time by 70-90%. California expects the efficiency gains to flow, at least partially, to clients.
Confidentiality Rules
California provides the most detailed confidentiality framework of any state. Attorneys must evaluate vendor terms of service before using any AI tool with client data. The guidance specifically flags risks with consumer AI tools that use input data for model training. Attorneys must confirm that AI vendors provide adequate data protection, do not retain client information beyond the session, and do not use inputs to train models. The guidance also addresses the risk of prompt injection attacks and data leakage through AI tool interfaces. Firms should maintain a vetted list of approved AI tools and prohibit use of unapproved platforms.
What's Still Unclear
California's guidance recommends disclosure but stops short of a blanket mandate — it notes that disclosure is required in certain court filings while recommending transparency with clients more broadly. The guidance does not define specific AI competence benchmarks or CLE requirements. It also does not address AI use in specialized contexts like criminal defense (where AI errors affect liberty interests) or immigration law. The interaction between California's AI guidance and the California Consumer Privacy Act (CCPA) for AI vendor data handling remains unexplored. See California's AI regulation page for the broader regulatory landscape.
The Bottom Line: California's November 2024 guidance is the most comprehensive state bar AI framework — covering competence, billing, confidentiality, and disclosure with practical specificity that sets the national standard.
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.