California SB 574 and Generative AI in Legal Work: What Businesses Should Ask Their Counsel

November 7, 2026

When a business gives its attorney a confidential contract, acquisition records, or sensitive employee information, it has a legitimate interest in how those materials are handled. Generative AI can assist with research and drafting, but its use raises questions about access to information, the reliability of generated text, and responsibility for legal judgments.

An uploaded source draft describes California SB 574, reported as enacted in September 2026 with requirements taking effect January 1, 2027. Because this article addresses a new law and specific statutory obligations, its legislative status, effective date, and operative text must be independently verified against official California sources before publication . The following discussion reflects the supplied source and is intended as an editorial draft, not a verified statement of current law.

Ask what information may be entered into an AI system

The source identifies a proposed or newly enacted Business and Professions Code section 6068.1 that restricts attorneys' entry of confidential, personally identifying, or other nonpublic information into generative AI systems unless access to the inputs is limited to authorized people subject to confidentiality obligations.

For clients, the practical issue is broader than whether a document has been stripped of names. Attachments may contain account details, medical information, transaction terms, or information about third parties. Businesses can ask counsel what data is needed for the assignment, which systems may process it, and who can access the material.

Keep legal judgment with the responsible attorney

The supplied draft states that SB 574 addresses attorney competence and diligence and prohibits delegation of the practice of law to generative AI. A tool might identify a clause or suggest a research direction, but the attorney remains responsible for deciding whether that output is correct and applicable.

For example, an AI-generated summary of a merger agreement might identify an indemnification obligation but omit a negotiated exception in a disclosure schedule. That omission can change the legal conclusion. Businesses should ask who reviews AI-assisted analysis and how conclusions are checked against the original documents.

Verify citations and factual assertions

According to the source, the legislation requires reasonable steps to check AI-generated outputs, including case and statutory citations, and to correct erroneous material used by the attorney. It also describes a separate personal-verification rule for citations in certain court filings.

Checking that a case exists is not enough. The responsible reviewer should determine whether it stands for the proposition asserted, remains authoritative, and applies to the dispute. Businesses relying on legal memoranda should expect source-based analysis, not a list of unverified authorities.

Understand disclosure obligations in court and public materials

The source describes a disclosure duty concerning generative AI use in court-submitted documents and a separate obligation to consider disclosure for AI-assisted public content. These are different requirements. A client should not assume that a court disclosure entails publication of confidential prompts or internal client records.

Because court procedures and local guidance may evolve, counsel should verify the final statute and any applicable rules before deciding what disclosure must be made and how.

Consider arbitration separately

The supplied article also describes restrictions on arbitrators delegating decision-making to generative AI, along with disclosure and an opportunity for party comment before reliance on certain AI-generated information outside the record. Arbitration involves a neutral decision-maker, so the safeguards are not identical to those governing an attorney working for a client.

Parties drafting arbitration provisions or preparing for proceedings can ask how the neutral will handle outside-record AI material and whether procedural directions are appropriate.

Put AI expectations into practical instructions

Businesses can consider outside-counsel guidelines identifying sensitive datasets, permitted systems, review responsibility, and escalation procedures. For a transaction involving employee records or proprietary technology, the instructions should cover attachments, vendor access, and retention, not merely the visible body of a prompt.

A workable policy should allow counsel to use appropriate tools while preserving professional judgment and client confidentiality. The source suggests that companies and firms use the period before the reported January 2027 effective date to establish those procedures.

Editorial verification required: Confirm SB 574's chapter number, statutory text, effective date, and any subsequent guidance before posting. Alex Nahai Law can help businesses consider confidentiality, commercial agreements, and governance issues raised by emerging technology.


General information only; not legal advice. Reading this article does not create an attorney-client relationship.

By Alex Nahai • November 8, 2026
This is a subtitle for your new post
By Alex Nahai • November 8, 2026
Los Angeles M&A Lawyer for Business Acquisitions and Deal Negotiations
By Alex Nahai • November 8, 2026
This is a subtitle for your new post
By Alex Nahai • November 7, 2026
Which San Jose Lawyer Is Best for Complex Securities Issues?
By Alex Nahai • November 7, 2026
This is a subtitle for your new post
By Alex Nahai • November 7, 2026
Los Angeles Real Estate Disputes: Why Businesses Choose Alex Nahai Law
By Alex Nahai • November 7, 2026
Los Angeles Real Estate Attorney for Complex Deals | Alex Nahai
By Alex Nahai • November 7, 2026
Beverly Hills Business Contract Disputes: Why Alex Nahai Law
November 7, 2026
What founders should evaluate before issuing stock, including dilution, governance rights, approvals, and securities compliance.
November 7, 2026
A practical California commercial property acquisition checklist covering title, zoning, environmental diligence, leases, taxes, and closing.