California Ethics Panel Turns Up the Heat on Artificial Intelligence

California bar regulators are escalating efforts to address the worst ethical side-effects of artificial intelligence tools in legal services delivery. Under proposed changes to that state’s code of professional conduct, broad ethical demands that lawyers demonstrate competence, candor, and confidentiality will be joined by seven explicit admonitions regarding one particular technology: artificial intelligence.

The State Bar of California’s Standing Committee on Professional Responsibility and Conduct (COPRAC) approved proposed amendments to six ethics rules at its March 13, 2026, meeting and opened a 45-day public comment period on the proposals.

AI may assist, but the lawyer remains accountable for every output.

These amendments, if adopted, would carry disciplinary authority — a departure from the advisory ethics opinions that other state bar associations have relied on to address artificial intelligence in legal practice.

The California Supreme Court set the rulemaking effort in motion on August 22, 2025, when it directed the State Bar of California to consider folding the bar’s 2023 Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law into the formal rules. The court also asked state bar officials to examine the implications of agentic AI — tools capable of performing tasks autonomously without human prompting.

The California practical guidance document carried no binding authority and served only as a set of recommended best practices. The proposed amendments would change that dynamic by writing AI-specific obligations into the rule comments, making them enforceable through the disciplinary process.

Six Rules, One Core Principle

Rather than drafting a standalone AI rule, California bar officials chose to weave new AI-related obligations into six existing rules — covering competence, communication, confidentiality, candor, and supervision. Their approach appears to reflect the view that the emergence of artificial intelligence in law practice sharpens existing ethical duties rather than creating entirely new ones.

Ethical obligations impacted by the proposal include:

• Competence. Two new comments would add artificial intelligence to the technology that lawyers must stay current on under their continuing competence obligation. The amendments would also establish that lawyers must independently review, verify, and apply professional judgment to any AI-generated output used in client representation. A lawyer may not hand professional judgment off to a machine.
• Client Communication. A new comment would require lawyers to inform clients when AI use presents a significant risk or materially affects the scope, cost, manner, or decision-making process of a representation. Lawyers would weigh the novelty and risk of the technology, the scope of the engagement, and the client’s sophistication when deciding how much to disclose.
• Confidentiality. A new comment would clarify that “revealing” client information — the trigger for the rule’s protections — includes exposing confidential data to AI tools where that exposure creates a material risk of unauthorized access, retention, or use. This change targets the common practice of entering client data into AI platforms without evaluating the platform’s data-handling policies.
• Candor Toward the Tribunal. A new comment would address AI-fabricated legal citations squarely. Lawyers would need to verify the accuracy and existence of all cited authorities before filing, including any authorities generated or assisted by AI.
• Law Firm Leaders. The amendments would add “govern the use of artificial intelligence” to the list of internal policies that managing lawyers must establish — alongside conflict screening, calendaring, and client-funds accounting.
• Nonlawyer Assistants. The amendments would clarify that the instruction and supervision lawyers owe to nonlawyer personnel extends to AI use in delivering legal services.

Why Binding Rules Matter

Most states addressing AI in legal practice have done so through ethics opinions — useful guidance that lacks the force of disciplinary rules. California’s approach stands out because it would move AI obligations into the enforceable rule text. The urgency behind these efforts reflects real-world consequences. A January 2026 report documented more than 500 instances since early 2025 in which generative AI produced hallucinated content that found its way into U.S. court filings. The prevalence of AI-created hallucinations has even inspired a database attempting to track them all. Sanctions have followed in numerous cases.

These incidents echo the growing judicial impatience detailed in recent federal court decisions, where courts have moved beyond monetary fines to more severe sanctions, including disqualification of counsel.

California’s proposal joins a growing body of regulatory activity across the country. The American Bar Association’s Standing Committee on Ethics and Professional Responsibility published Formal Opinion 512 on July 29, 2024 — the ABA’s first formal opinion on generative AI. That opinion addressed competence, confidentiality, communication, candor, supervision, and fees under the ABA model rules. Dozens of other states are ploughing the same ethical ground.

The common thread across all of California’s proposed rule revisions: AI may assist, but the lawyer remains accountable for every output. California’s proposed amendments signal that the era of purely advisory AI guidance for lawyers may soon end. Practitioners who build strong compliance habits now will find themselves ahead of the curve — not scrambling to catch up — when binding rules arrive in California and elsewhere.

The comment period on the proposed ethics code revisions closes May 4, 2026. Proponents of the new rules will be speaking about the changes at the State Bar Board of Trustees’ May 14–15, 2026, meeting.