California.
Attorney AI ethics guidance status for California as of May 9, 2026. Practical Guide (Nov 2023) + proposed amendments to Rules 1.1, 1.4, 1.6, 3.3, 5.1, 5.3 (Mar 2026 comment period).
California bar AI ethics status. #
California issued the influential 2023 COPRAC Practical Guidance for Generative AI in November 2023, framing competence, confidentiality, and supervision under existing Rules of Professional Conduct. In March 2026, COPRAC approved proposed amendments to Rules 1.1, 1.4, 1.6, 3.3, 5.1, and 5.3, putting AI-specific obligations into the binding rules rather than non-binding practical guidance. The 45-day public comment period opened March 13, 2026. California Rule of Court 10.430 separately requires courts to either ban AI use by judicial officers or adopt a formal AI policy with verification and disclosure.
Authoritative source: https://www.calbar.ca.gov/Portals/0/documents/ethics/Generative-AI-Practical-Guidance.pdf
Last verified: . If this row text disagrees with the linked source, the source controls.
What this means in practice. #
For practitioners admitted in California, the formal guidance above is the binding professional-conduct anchor for AI use. The recurring duties: (1) maintain competence in the AI tool, (2) protect client confidentiality (do not submit confidential material to public-version AI tools), (3) supervise AI output, and (4) verify all factual and legal content before submission.
Cross-jurisdictional consideration: ABA Formal Opinion 512 (July 2024) is broadly consistent with the California guidance and serves as a federal anchor for multi-state practitioners. Where California guidance is more specific or more restrictive than ABA Op. 512, the California guidance controls within California.
Comparison to ABA Op. 512. #
Closely tracks ABA Op. 512 with two California-specific extensions. The 2023 COPRAC Practical Guidance pre-dates ABA Op. 512 (the ABA opinion drew partly on California's framework). Both documents map AI obligations onto existing rules rather than creating a new AI-specific rule regime.
Where California is broader:
- Technological architecture competence. California's Practical Guidance specifically requires understanding of large language models — how they work, their limitations, their hallucination tendencies — before use. ABA Op. 512 frames competence more abstractly.
- Cost transparency. California addresses both billable-hours and prompt-cost allocation specifically; ABA Op. 512 treats AI costs as a generic Rule 1.5 reasonableness question.
- Pending move to binding rules. The March 13, 2026 COPRAC-approved amendments would convert AI-specific obligations from non-binding "Practical Guidance" into amendments of binding Rules 1.1, 1.4, 1.6, 3.3, 5.1, and 5.3. Once adopted, California will become the first state to put AI into the rule text rather than commentary or guidance.
- Court-officer overlay. California Rule of Court 10.430 separately requires courts either to ban judicial-officer AI use or to adopt a formal AI policy with verification and disclosure. ABA Op. 512 does not address judicial-officer AI use at all.
Federal court AI standing orders affecting California practitioners. #
No widely-noted court-wide or individual-judge AI standing orders affecting California's federal districts have surfaced as of May 2026. Individual judges may have issued orders that are not in our representative list. Practitioners should consult each assigned judge's standing orders before filing in any federal matter. The comprehensive corpus of federal-judge AI orders is maintained by Bloomberg Law and Law360 (both paywalled).
Last verified: . The comprehensive corpus of federal-judge AI orders is maintained by Bloomberg Law and Law360.
AI case-law context for California. #
California sits within the 9th Circuit (United States Court of Appeals for the 9th Circuit). Decisions of the 9th Circuit are binding on the federal courts within California; state-court decisions of the California appellate system bind California's state courts; Supreme Court decisions are binding on everyone. The Supreme Court has not yet issued a decision directly addressing attorney AI use, so the controlling case law is at the circuit and district levels:
Case No. 23-cv-03223-AMO, 2024 WL 3748003 (N.D. Cal. Aug. 8, 2024) (Martínez-Olguín, J.) — held that attorney-crafted prompts to AI systems can constitute opinion work product. The leading 9th Circuit decision on AI in legal practice; persuasive nationally.
have issued individual AI standing orders. The 9th Circuit has not issued a circuit-wide rule.
Mata v. Avianca and Park v. Kim are routinely cited as persuasive authority in 9th Circuit federal courts.
Practitioners in California should treat the cases above as the operating framework for AI-related conduct until California-specific binding precedent or California bar guidance controls a particular question. The verification duty under Mata and Park v. Kim has been routinely cited in sanctions matters across all federal circuits including the 9th Circuit.
California legal landscape. #
- Northern District of California (N.D. Cal.)
- Eastern District of California (E.D. Cal.)
- Central District of California (C.D. Cal.)
- Southern District of California (S.D. Cal.)
California has the largest legal market in the US. The State Bar is a state agency under direct legislative oversight, distinct from the voluntary California Lawyers Association. The federal court system mirrors California's scale: four federal districts, with N.D. Cal. (covering Silicon Valley) generating an outsized share of technology-related federal litigation.
Most US states have adopted the substance of ABA Model Rule 1.1, Comment 8 — that the duty of competence requires lawyers to keep abreast of the benefits and risks of relevant technology. Whether California has adopted Comment 8 verbatim, in modified form, or relies on the underlying Rule 1.1 text alone, the substantive duty of technological competence applies to AI-related practice in California.
California-specific FAQ.
Has the California State Bar issued formal AI ethics guidance?
Yes — two distinct instruments. The Committee on Professional Responsibility and Conduct (COPRAC) issued Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law in November 2023, mapping AI obligations onto the existing California Rules of Professional Conduct. On March 13, 2026, COPRAC approved proposed amendments to Rules 1.1, 1.4, 1.6, 3.3, 5.1, and 5.3 themselves — putting AI-specific obligations into binding rule text rather than non-binding guidance. The 45-day public comment period opened March 13, 2026. If adopted, California becomes the first US state to put AI obligations into the rule text rather than commentary.
What does the 2023 COPRAC Practical Guidance specifically require?
Five operational duties under existing California Rules of Professional Conduct: (1) competence (Rule 1.1) — including understanding of large language models, their training data, and their hallucination tendencies before use; (2) confidentiality (Rule 1.6) — including evaluating AI vendor data-handling practices before submitting any client information; (3) supervision (Rule 5.3) — applying the nonlawyer-assistance framework to AI vendor relationships; (4) candor toward the tribunal (Rule 3.3) — verification of AI-generated authority before submission; and (5) reasonable fees (Rule 1.5) — addressing both billable-hours allocation when AI saves time and direct AI prompt costs passed through to clients.
How does California Rule of Court 10.430 affect AI in court proceedings?
Rule of Court 10.430 requires every California court to either prohibit AI use by judicial officers and research attorneys or adopt a formal AI policy with verification and disclosure requirements — a binary that creates a default-prohibition posture for any court that has not yet adopted a formal policy. This is a court-system overlay separate from the bar's COPRAC guidance, and it applies to judicial AI use rather than counsel use. Practitioners appearing before any California state court should verify each court's posted AI policy as part of routine pre-filing diligence; the policy controls what the chambers may do with the practitioner's submissions.
Which federal courts cover California, and what AI rules apply there?
California has four federal districts: Northern (N.D. Cal., San Francisco / Silicon Valley — the busiest technology-related federal docket in the country), Eastern (E.D. Cal., Sacramento), Central (C.D. Cal., Los Angeles), and Southern (S.D. Cal., San Diego). Multiple individual judges in N.D. Cal. and C.D. Cal. have issued AI standing orders requiring disclosure and verification. The 9th Circuit has not issued a circuit-wide AI rule. The leading binding case-law is Tremblay v. OpenAI, Inc., Case No. 23-cv-03223-AMO, 2024 WL 3748003 (N.D. Cal. Aug. 8, 2024) (Martínez-Olguín, J.), which held that attorney-crafted prompts to AI systems can constitute opinion work product. Tremblay is the most-cited 9th Circuit AI-in-legal-practice decision.
How does the 2026 proposed amendment affect California practice?
If adopted as drafted, the amendments incorporate AI-specific obligations directly into the binding text of Rules 1.1, 1.4, 1.6, 3.3, 5.1, and 5.3. Three practical consequences: (1) violations of the AI-specific provisions become directly disciplinable through the State Bar Court rather than merely advisory under non-binding practical guidance; (2) the verification duty under Rule 3.3 explicitly extends to AI-generated authority, with non-verification being a candor-toward-the-tribunal violation rather than a Rule 11-equivalent procedural matter; (3) the vendor-supervision duty under Rule 5.3 explicitly covers AI vendors, aligning California with North Carolina's 2024 FEO 1 framing. Practitioners should track the public-comment record because the final adopted text may differ from the March 13, 2026 draft.
Why is the State Bar of California structurally different from other state bars?
The State Bar of California is a state agency under direct legislative oversight, not a voluntary association — distinct from the California Lawyers Association (which is voluntary). The Board of Trustees is appointed (partly by the legislature, partly by the California Supreme Court). The State Bar Court is the sole disciplinary forum, with appeals running to the California Supreme Court. This structure means COPRAC guidance and the proposed Rule amendments operate as direct regulatory instruments under the State Bar Act, not advisory recommendations from a private association. With approximately 250,000 active attorneys — the largest US legal market — California rule changes shift the national center of gravity for AI legal-practice regulation.
What about AI vendor diligence in California?
The 2023 COPRAC Practical Guidance specifically requires evaluating AI vendor data-handling practices before submitting any client information — explicitly framing this as a Rule 1.6 confidentiality duty rather than only a Rule 5.3 supervision duty. California treats both framings as applicable: confidentiality (Rule 1.6) for the data-flow analysis, supervision (Rule 5.3) for the ongoing oversight obligation. North Carolina's 2024 FEO 1, which frames AI vendors explicitly under Rule 5.3, is treated as persuasive authority in California. The proposed March 2026 amendments to Rule 5.3 are expected to formalize this dual-framework approach in California rule text.
Why California is different. #
California's State Bar is structurally distinct from most state bars in the US: it is an integrated, mandatory state agency rather than a private association. The Board of Trustees is appointed (partly by the legislature, partly by the Supreme Court). The State Bar Court is the sole disciplinary forum, with appeals to the California Supreme Court.
Three implications for AI regulation:
- Bar guidance carries direct regulatory weight. Where many state-bar opinions are advisory or persuasive, California's COPRAC guidance and the proposed Rule amendments operate as direct regulatory instruments under the State Bar Act. The March 2026 proposed amendments would put AI obligations into binding rule text — first state to do so.
- Cal Rule of Court 10.430 covers the courts. A separate court rule requires courts either to ban AI use by judicial officers or to adopt a formal AI policy with verification and disclosure. This is a court-system overlay no other state has.
- Volume effect. California has more practicing attorneys (≈250,000) than any other state. Bar guidance from COPRAC affects roughly one in five US lawyers. The proposed Rule amendments will shift the national center of gravity for AI regulation in legal practice once adopted.