California has moved from guidance to legislation: the state is codifying rules that govern how attorneys may use artificial intelligence in the practice of law, converting what was previously advisory ethics material into binding professional obligations.

The shift comes amid mounting evidence that lawyers are turning to generative AI tools for research, drafting and client communications — and that the technology sometimes produces fabricated case citations and other errors. The new legislative approach in California directly addresses that risk.

Why did California act now?

The push follows a wave of reported incidents in which attorneys cited non-existent authorities generated by AI tools in filings submitted to courts. Those episodes, widely covered in the legal press, exposed a gap: existing professional-conduct rules did not explicitly address AI-generated work product.

California's response replaces that ambiguity with codified requirements. For practitioners, the practical consequence is straightforward — reliance on general ethical duty of competence is no longer sufficient. Attorneys using AI must now comply with specific, legislated standards, and failure to do so carries the prospect of professional discipline rather than merely informal guidance.

What does the legislation cover?

The codified rules target the core stages of legal work where AI now plays a role:

• Research and citation. Attorneys must verify AI-generated citations and legal authorities before submitting them to a court.
• Client disclosure. The framework addresses when lawyers must inform clients that AI tools are being used on their matters.
• Confidentiality. The rules engage with the duty to protect client information when data is entered into AI systems.
• Supervision. Responsibility for AI-assisted work remains with the licensed attorney, not the tool.

The operative principle is that AI output is treated as a starting point, not a finished product. The lawyer — not the software — answers for accuracy.

What does this change for practitioners?

For California attorneys, the change moves AI governance from the margins of ethics commentary into the center of daily compliance. Law firms of every size will need documented workflows: verification steps for AI-generated research, disclosure protocols for client engagements, and training so that junior lawyers and paralegals understand where human review is mandatory.

Out-of-state lawyers advising on California matters should also take note. Professional-conduct exposure does not stop at the state line when the representation touches California clients or courts.

How does this fit the national picture?

California is not acting alone. Courts across the United States have issued standing orders requiring disclosure of AI use in filings, and some judges have sanctioned lawyers who submitted hallucinated citations. Bar associations in several states have published formal opinions on generative AI and the practice of law.

What distinguishes California's move is its form. Guidance from a bar association can be updated or reinterpreted; codified rules carry legislative weight and a clearer enforcement pathway. That gives regulators a firmer basis for discipline and gives practitioners a clearer standard against which to measure their own conduct.

What should law firms do immediately?

Practitioners advising on compliance should consider several near-term steps:

• Audit current AI use across research, drafting and client communication functions.
• Establish written verification procedures for all AI-generated citations and factual assertions.
• Update engagement letters and disclosure language to address AI use where required.
• Confirm that vendor tools handling client data meet confidentiality obligations.
• Train all timekeepers on the codified requirements before reliance on AI becomes routine.

The legislative turn in California signals that AI in legal practice has moved past the novelty phase. Regulators are watching, and the rules are now written down.