TL;DR: California has enacted SB 574, the nation’s first comprehensive statutory framework governing attorneys’ use of generative AI. Signed by Governor Newsom on September 30, 2026, the law obligates lawyers to verify AI-generated content, disclose AI use in court submissions, protect confidential information in AI systems, and prohibits delegating the practice of law to AI. The rules take effect January 1, 2027 and extend to arbitrators and other ADR professionals. For trial teams, this creates a clear policy discipline around AI-assisted drafting, evidence citation, and disclosures, with Objection Academy’s practice modules offering a practical way to train for courtroom objections and AI-informed evidence handling.
What SB 574 does
SB 574 imposes a multi-part regime aimed at ensuring accuracy, accountability, and data protection when genAI tools are used in litigation. Key provisions include: (1) a prohibition on delegating the practice of law to generative AI, meaning the attorney of record must retain substantive control and responsibility over filings and submissions; (2) a requirement that any brief, pleading, motion, or court document containing AI-assisted content must be personally verified by the attorney responsible for submission, including accuracy of any citations generated by the AI; (3) a duty to disclose AI usage to the court for all documents submitted to the court, and to consider disclosure for content provided to the public; (4) protections preventing nonpublic confidential information from being entered into public GenAI systems, with a mandate to take reasonable steps to verify outputs and correct errors; (5) an explicit directive that the California Code of Civil Procedure be amended to require citation verification in state filings; and (6) applicability to arbitrators and alternative-dispute-resolution providers in California, reflecting a broad approach to GenAI governance across the civil justice system. The text and summaries emphasize not only disclosure and verification but also the safeguarding of client information against AI platforms that may train on uploaded materials. These provisions become effective January 1, 2027. (legiscan.com)
Practical implications for trial teams
For trial lawyers, SB 574 signals a fundamental shift in how AI is used in the preparation and submission of court documents. First, firms should implement an internal GenAI policy that delineates when AI may assist, who reviews AI outputs, and how citations are checked. Because the law requires personal verification of citations by the submitting attorney, pipeline workflows should separate AI-generated drafting from the final sign-off, with a dedicated checklist to confirm that every authority is correctly cited and that all statements align with the record and controlling law. Second, the disclosure obligation means that counsel should standardize a practice note for each filing to document AI use and the tools involved, including version numbers and the scope of AI assistance. Third, the confidentiality rule requires robust data-handling protocols to prevent nonpublic information from entering public AI systems, including training a policy for when to redact sensitive data before any AI-assisted drafting or review. Fourth, with arbitrators now included, ADR practitioners should anticipate similar scrutiny of AI-generated materials in motions for ADR proceedings and related submissions. Finally, the California environment creates opportunities for cross-jurisdictional matters; while SB 574 is California-specific, it sets a potential benchmark that other states may study as they consider their own GenAI guidelines. For working litigators, these changes may necessitate additional review steps, new disclosure templates, and more careful preservation and privilege considerations when AI tools are used during discovery or case preparation. (gov.ca.gov)
The Doe v GitHub DMCA context and wider evidence implications
Against the backdrop of AI-enabled litigation, the Ninth Circuit’s September 16, 2026 decision in Doe v. GitHub, Inc. affirms the district court’s dismissal of DMCA Section 1202 claims tied to AI-generated outputs. The court held that Copilot and Codex outputs do not “remove or alter” Copyright Management Information from existing works, because the AI often creates new outputs rather than copying with CMI removed. The decision clarifies that at least as to the DMCA 1202 output theory, the materials generated by AI do not implicate the statute in the same way as copying or altering CMI in a traditional sense. This ruling—decided just over a month before SB 574’s signing—remains highly relevant for trial teams evaluating the evidentiary significance and discoverability of AI prompts and outputs in civil litigation. Practically, this decision underscores that AI-generated content can raise new kinds of issues, but it does not automatically transform AI outputs into weaponizable evidentiary artifacts under the DMCA framework. (cdn.ca9.uscourts.gov)
Actionable steps for California practitioners (the 30–60–90 plan)
- 30 days: appoint a GenAI task force to draft a firm-wide policy, mapping when AI tools may be used for drafting, research, or review. Create a standard disclosure template for filings that notes AI involvement, tool names, and version numbers. Begin a “citations verification” protocol to satisfy SB 574’s personal verification requirement.
- 60 days: implement a data-handling protocol that restricts entering confidential or nonpublic information into any public GenAI system. Institute a secure, approved list of AI providers, with access controls and client data safeguards. Begin practice using a dedicated reviewer or checklist to verify AI-generated citations before filing.
- 90 days: train attorneys and support staff on SB 574 requirements, including a short course on how to prepare GenAI-assisted submissions that still meet professional standards and ethical duties. Integrate Objection Academy modules as a practical training companion to drill objection handling and evidentiary authentication in AI-augmented workflows. Ensure that every filing contains a clear AI-use disclosure and a verification log summarizing who reviewed AI outputs and what was verified. (morganlewis.com)
Objection Academy in the GenAI era
As California engineers a GenAI regime across courts and ADR settings, trial teams benefit from disciplined rehearsal of objections, authentication, and evidence handling. Objection Academy’s practice drills for objections, cross-examination, and real-time evidentiary reasoning can help attorneys internalize the discipline SB 574 demands, particularly around verifying AI-generated authorities and ensuring transparent disclosure of AI involvement in filings. While SB 574 governs the practice of law in California, the broader emphasis on rigorous verification and ethical AI use aligns with modern trial skills training and can be a practical complement to California’s new rules. (ropesgray.com)
Cross-border and preparedness notes
SB 574 places California at the forefront of GenAI regulation for litigators, and firms with multi-state practices should prepare parallel playbooks for other jurisdictions. While no other state has enacted a law identical to SB 574, several jurisdictions are advancing AI ethics guidance, discovery protocols, or proposed rules addressing AI in evidence and discovery. California’s approach provides a concrete blueprint for practice readiness, compliance, and trial-readiness investments that can pay dividends as other states watch California’s implementation and outcomes. (gov.ca.gov)
Sources
- California SB 574 text and implementation details, including January 1, 2027 effective date: LegiScan, SB 574 text; California Governor press release; law firm analyses. (legiscan.com)
- California Governor Newsom signs SB 574; first-in-nation GenAI duties for attorneys and arbitrators: Governor of California, Newsom press release. (gov.ca.gov)
- Practical guidance and analysis on SB 574 from Morgan Lewis and Baker McKenzie: implications for legal departments and litigation practice. (morganlewis.com)
- Ninth Circuit Doe v GitHub, Inc. (Sept. 16, 2026): DMCA 1202 outputs claim discussion and input theory forfeiture. (cdn.ca9.uscourts.gov)
- California Courts Newsroom and related coverage of AI governance developments in California. (newsroom.courts.ca.gov)