The Ivo Blog
→
California’s New AI Rules for Lawyers: What SB 574 Means for Contract Work
AI in the legal profession

California’s New AI Rules for Lawyers: What SB 574 Means for Contract Work

Kelsey Eisen
Kelsey Eisen

Starting January 1, 2027, a new statute will go into effect governing the use of generative AI by lawyers. The measure passed unanimously and was signed into law by Governor Newsom on September 30, 2026. SB 574 is the country’s first comprehensive statutory scheme for governing attorneys’ use of generative AI. Read on for a breakdown of SB 574, what to check before the law takes effect, and how Ivo can help you stay in compliance.

What SB 574 requires

The core requirements of SB 574 are found in a new Business and Professions Code section, § 6068.1. The statute doesn’t prohibit attorneys from using AI, but imposes duties of confidentiality, verification, and disclosure, and bars attorneys from delegating the practice of law to AI. SB 574 also explicitly calls out in-house counsel as being covered under the statute.

No delegating the practice of law to AI. An attorney may not “delegate the practice of law to generative artificial intelligence” (§ 6068.1(a)(2)). The statute doesn’t explicitly define “delegate” nor does it define “the practice of law.”

Confidentiality. Attorneys may not enter confidential, personal identifying, or other nonpublic information into a generative AI system unless access is limited to the attorney and people the attorney has authorized who are bound by confidentiality obligations (§ 6068.1(a)(3)(A)).

Verification. Attorneys must take reasonable steps to verify the accuracy of AI output, including all case and statutory citations, and correct any erroneous or hallucinated content in material they use (§ 6068.1(a)(3)(B)).

Disclosure. Attorneys must disclose AI use to the court for every document submitted to it, and consider whether to disclose it for content provided to the public (§ 6068.1(a)(3)(C)). 

How this affects contract work

Nothing in the bill mentions contracts. The relevant inquiry is then what “delegating the practice of law” covers, as the statute doesn’t define it. California courts treat preparing contracts as the practice of law (People v. Merchants Protective Corp., 1922), so the delegation language could be reasonably read to reach contract work. ABA Formal Opinion 512 lists contract review and due diligence among the tasks AI tools can assist with, but overall, no guidance interpreting SB 574 beyond the plain language of the statute has been issued. 

Two other duties in the statute do apply to contract work more directly. Confidentiality covers any nonpublic information an attorney enters into an AI tool, and counterparty contracts typically qualify. The verification duty can be read as reaching transactional work, so attorneys who use AI to redline, summarize, or analyze contracts should expect to review that output before relying on it. 

In practice, teams affected by SB 574 will want to confirm who can access their AI tools, document who reviewed AI-assisted work, and define where attorney judgment sits in their workflow. Consult counsel on how the law applies to you. 

How Ivo supports your compliance

SB 574 puts its duties on attorneys for compliance, so no vendor can make a team compliant simply by virtue of the tool used. What a vendor can do, however, is make the duties easier to meet and easier to show. Here is how Ivo does that.

Attorney review. Ivo Review produces surgical redlines that track to your team’s exact language and playbook, so attorneys review changes against standards they wrote. Every suggested change arrives with the playbook position it came from and a short explanation of why it was raised, which means the attorney is reviewing against a known standard rather than reverse-engineering someone else's judgment. From there it's the attorney’s call: accept the markup as drafted, adjust it, or pass on it entirely and move to the next item. Nothing is applied to the document and nothing is finalized until someone on the attorney’s team does it.

Output you can check. Every recommendation Ivo generates is tied back to what produced it. You can see which playbook position it came from, read the explanation of why it was raised, and look at the clause in the agreement that triggered it, so a reviewer can trace any suggestion from the final markup back to the standard behind it without taking the output on faith. When Ivo Research surfaces legal authority, each citation links to the underlying source so an attorney can open it and confirm the proposition before relying on it.

A record of review. Ivo keeps a workspace-level audit log, available to administrators in settings, covering sign-ins and authentication events, user and workspace administration, changes to playbooks including who created, edited, imported or viewed them, deletions and moves of contracts in your repository, and changes to API keys and integrations. Within a review, each recommendation carries its own history, so a reviewer can see how the suggested language changed during the session and revert to an earlier version. 

Documentation for your diligence. Ivo holds:

  • SOC 2 Type II
  • ISO/IEC 27001
  • ISO/IEC 42001
  • Cyber Essentials

Further details are available at trust.ivo.ai. 

This post gives general information, not legal advice. Consult counsel about how SB 574 applies to your team.