AI ethics for lawyers in 2027: what Opinion 512 and SB 574 require

If your attorneys use generative AI on client matters, the 2027 ethics question is whether the firm can show a client, a court or the State Bar how each use was checked, protected and disclosed. For California firms that question now has a date: SB 574 takes effect January 1, 2027.

This article looks at AI ethics for lawyers from the 2027 side: what ABA Formal Opinion 512 still settles, what California has added since, and the controls a managing partner can put in place this quarter.

What does ABA Formal Opinion 512 say about generative AI?

ABA Formal Opinion 512, issued July 29, 2024, applies the existing Model Rules to generative AI tools and creates no new rules. It walks through competence, confidentiality, communication, candor and meritorious claims, supervision and fees, and it is the ABA's formal ethics opinion on generative AI.

In summary, the opinion makes six points:

• Competence (Rule 1.1). Lawyers need a reasonable understanding of the capabilities and limitations of the tools they use, and the level of review must fit the task. Uncritical reliance on output is inconsistent with the duty.
• Confidentiality (Rule 1.6). Before entering information relating to a representation into a self-learning tool, a lawyer needs the client's informed consent. The opinion says boilerplate language in an engagement letter is not enough.
• Communication (Rule 1.4). Some uses must be disclosed to the client, for example when the client asks or when AI output will influence a significant decision in the matter.
• Candor (Rules 3.1, 3.3 and 8.4(c)). Lawyers remain responsible for every citation and factual assertion they submit.
• Supervision (Rules 5.1 and 5.3). Managerial lawyers should set clear policies on permitted use and make sure lawyers and staff are trained.
• Fees (Rule 1.5). A lawyer billing by the hour may bill for time actually spent, including time reviewing output, but generally may not bill a client for learning a tool the lawyer will use across matters.

The opinion is advisory. It does not bind lawyers on its own, and state rules control. That distinction matters more in 2027 than it did in 2024, because California has moved well past guidance.

What has changed in AI ethics for lawyers in California since Opinion 512?

Since Opinion 512, California has added a statute, pending rule amendments and revised bar guidance, all built on the same principles. Four developments matter most.

SB 574 becomes law. Governor Newsom signed SB 574 (Stats. 2026, ch. 858) on September 30, 2026, and it takes effect January 1, 2027. It adds Business and Professions Code section 6068.1 and amends Code of Civil Procedure section 128.7.

Rule amendments are pending. The State Bar's ethics committee has proposed AI-specific comments to rules 1.1, 1.4, 1.6, 3.3, 5.1 and 5.3, following an August 22, 2025 request from the California Supreme Court. The proposals would require verifying cited authorities, treat exposing client information to an AI system as a possible "reveal" under rule 1.6, and call on managerial lawyers to set internal AI policies. The second public comment period closed August 6, 2026. The amendments take effect only if the Board of Trustees and the Supreme Court approve them, and as of October 2026 the State Bar's page did not show that approval.

The bar's guidance was updated. On May 14, 2026, the Board of Trustees approved a revised Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law. The 2026 version replaces the November 2023 guidance and adds material on agentic AI tools that perform tasks without human prompting, including that lawyers must not let AI systems file documents or communicate with the court on their own.

Courts are sanctioning fabricated citations. In Noland v. Land of the Free (2025), the Court of Appeal found that 21 of 23 case quotations in an opening brief were fabricated, sanctioned counsel $10,000 and forwarded the opinion to the State Bar.

How does SB 574 compare with Opinion 512?

SB 574 turns several of Opinion 512's recommendations into statutory duties for California attorneys, and amended section 128.7 adds court sanctions for unverified citations. The main differences:

• Verification. Opinion 512 asks for review that fits the task. Section 6068.1 requires reasonable steps to verify generative AI output, including the accuracy of all case and statutory citations, and to correct erroneous or hallucinated output. Amended section 128.7 bars filing any paper with a citation the responsible attorney has not personally verified, whatever its source. Our guide on how to verify AI citations before filing covers the workflow.
• Confidentiality. Opinion 512 centers on informed consent for self-learning tools. Section 6068.1 bars entering confidential, personally identifying or other nonpublic information into a generative AI system unless access is restricted to the attorney and persons the attorney authorizes who are obligated to keep it confidential.
• Disclosure to the court. Opinion 512 sets no general court disclosure duty. Section 6068.1 requires attorneys to disclose generative AI use to the court "for all documents submitted to the court," and to consider whether to disclose it when AI is used to create content provided to the public. The statute does not prescribe the form, and commentators note that courts and the State Bar have not yet clarified whether incidental uses, such as proofreading, count.
• Delegation. Section 6068.1 states that an attorney may not delegate the practice of law to generative AI.
• Firm responsibility. Under section 128.7, a law firm is jointly responsible for its lawyers' and employees' violations, absent exceptional circumstances.

Opinion 512 remains useful for the questions SB 574 leaves open, such as client communication and fees. Firms should confirm with ethics counsel how the new statute applies to their own practice.

What does ethical AI use look like in a California firm in 2027?

Ethical AI use in a California firm in 2027 means documented controls that a partner can show a client, a court or the State Bar. A written law firm AI policy is the starting point. The firm also needs a way to show the policy is followed.

In practice that means:

• An approved tool list. Each tool is reviewed for how it retains inputs, whether it trains on them and who can access them. Because section 6068.1 limits entering confidential information unless access is restricted to the attorney and authorized persons bound to keep it confidential, access controls are a core part of that review. Our guide to reviewing an AI vendor contract under ABA Formal Opinion 512 walks through the clauses to check.
• A verification step with a name on it. Before anything is filed, the responsible attorney personally checks each citation against the source, and the matter file records that it happened.
• Standard disclosure language for court filings that used generative AI, approved once and used consistently.
• Specific client communication. Engagement letters describe how the firm uses AI on matters, in plain terms, with informed consent obtained where Opinion 512 or state rules require it.
• A billing rule that matches Opinion 512's fee analysis and is explained to clients before the first invoice.
• Training and supervision for associates, paralegals and staff, refreshed when tools change.
• Logs. Records of which tools were used, by whom and on which matters, so supervising attorneys can review use under rules 5.1 and 5.3.

What should a law firm do before January 1, 2027?

A law firm should start with the changes that affect daily practice before January 1, 2027: tool access, citation checks, court disclosure and engagement letters.

• Inventory every AI tool lawyers and staff use today, including AI features built into research, document and practice management platforms. Sign-in and email records can show AI tools attorneys signed up for with firm email.
• Block or restrict tools that do not limit access to the firm and its authorized users.
• Add a citation verification sign-off to the filing checklist for every matter.
• Draft and approve the firm's court disclosure language with ethics counsel.
• Update engagement letter language on AI use and billing.
• Brief every attorney on SB 574 and on the California Rules of Professional Conduct as they stand, and track the pending amendments.

How can s90 help?

s90 works with law firms in Los Angeles and Orange County to put technical controls behind an AI policy: limiting AI use to approved, access-restricted platforms, keeping client data out of public tools with data loss prevention, and keeping prompt and usage logs that supervising attorneys can review. If you want a second set of eyes on your firm's AI setup before January, we are glad to talk it through.