CALIFORNIA SB 574: DID CALIFORNIA NEED TO CREATE ANOTHER LAYER OF ATTORNEY REGULATION?
By Dina Blikshteyn
On September 30, 2026, Governor Gavin Newsom signed California Senate Bill 574 into law. The law takes effect January 1, 2027. SB 574 is the first law of its kind, and poses a question for the legal community: Does the responsible use of artificial intelligence (AI) in legal practice really require new legislative mandates when existing professional duties, judicial rules, and disciplinary mechanisms already address many of the same concerns?
The impetus behind SB 574 is understandable. Since OpenAI released ChatGPT in November 2022, courts have seen attorneys submit briefs with fabricated citations and other inaccurate AI-generated content. Despite sanctions, reputational harm, and potential disciplinary consequences, these incidents keep happening at firms of every size, from small practices to AM 100 firms.
These incidents raise a fundamental question: Do they expose a gap in existing professional rules, or do they show that some attorneys are failing to comply with duties that already exist?
This article is written by Dina Blikshteyn, a partner and Co-Chair of AI and Deep Learning Practice Group and Haynes Boone, LLP. Dina focuses her practice on intellectual property. Dina also regularly counsels clients on various issues related to AI.
_______________________________________________________________________________
CALIFORNIA SB 574: DID CALIFORNIA NEED TO CREATE ANOTHER LAYER OF ATTORNEY REGULATION?
On September 30, 2026, Governor Gavin Newsom signed California Senate Bill 574 into law. The law takes effect January 1, 2027. SB 574 is the first law of its kind, and poses a question for the legal community: Does the responsible use of artificial intelligence (AI) in legal practice really require new legislative mandates when existing professional duties, judicial rules, and disciplinary mechanisms already address many of the same concerns?
The impetus behind SB 574 is understandable. Since OpenAI released ChatGPT in November 2022, courts have seen attorneys submit briefs with fabricated citations and other inaccurate AI-generated content. Despite sanctions, reputational harm, and potential disciplinary consequences, these incidents keep happening at firms of every size, from small practices to AM 100 firms.
These incidents raise a fundamental question: Do they expose a gap in existing professional rules, or do they show that some attorneys are failing to comply with duties that already exist?
What SB 574 Requires
SB 574 sets new statutory requirements for the use of generative AI by attorneys and arbitrators. It also changes oversight of alternative dispute resolution providers. Its main provisions are as follows:
Attorneys’ use: New Business and Professions Code section 6068.1 (1) prohibits attorneys from delegating the practice of law to generative AI, (2) requires attorneys to protect confidential and nonpublic information, and (3) requires attorneys to take reasonable steps to verify AI-generated material and to correct erroneous or hallucinated output.
Mandatory disclosure: Section 6068.1 also requires attorneys who use generative AI to “disclose the use of generative artificial intelligence to the court for all documents submitted to the court.” It further requires attorneys to consider whether to disclose AI use in content provided to the public.
Verification of legal citations: SB 574 amends Code of Civil Procedure section 128.7 to require attorneys to personally verify every citation in a court filing, whether or not AI was used. The language tracks the “personally read and verified” standard from Noland v. Land of the Free, but the Legislature removed the word “read” before enactment. Violations are subject to sanctions under section 128.7, including its 21-day safe-harbor procedure.
Restrictions on arbitrators: New Code of Civil Procedure section 1282.1 prohibits arbitrators from delegating any part of their decision-making process to AI. It also restricts arbitrators from relying on AI-generated information outside the evidentiary record unless they disclose it in advance and, where practical, give the parties an opportunity to comment.
Judicial standards: The legislation directs the Judicial Council to publicly revisit, and revise as necessary, Standard 10.80 of the California Standards of Judicial Administration, which governs judicial officers’ use of generative AI.
Alternative dispute resolution oversight: SB 574 also revises the State Bar’s existing voluntary certification program for ADR providers. Business and Professions Code section 6173 has required that program since January 1, 2025. The new law adds a confidential process under which the State Bar or an approved professional organization receives, investigates, and resolves complaints.
These provisions aim to promote accuracy, confidentiality, and accountability. But how much of SB 574 fills genuine gaps in existing law, and how much simply codifies duties that attorneys and other legal professionals already have to observe?
Existing Professional Duties Already Govern AI-Assisted Legal Work
The California Rules of Professional Conduct already address many of the risks of generative AI. Rules 1.1, 1.3, and 1.6 govern competence, diligence, and confidentiality, and Rule 3.3 requires candor toward the tribunal. These duties apply regardless of the technology used.
California appellate decisions show that meaningful remedies already exist for AI-related misconduct. In Noland v. Land of the Free, L.P. (2025) 114 Cal.App.5th 426, the court imposed $10,000 in sanctions for fabricated citations and quotations and referred the matter to the State Bar. In People v. Alvarez (2025) 114 Cal.App.5th 1115, the court imposed $1,500 in sanctions for fabricated and misrepresented legal authorities and notified the State Bar. In Shayan v. Shakib (Cal. Ct. App., 2d Dist., Div. 1, Dec. 1, 2025), the court imposed $7,500 in sanctions, struck the defective brief, required a corrected filing, and directed that the matter be reported to the State Bar.
The State Bar also addressed the ethics of AI before SB 574. It issued its Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law on November 16, 2023, and updated it in May 2026.
The Judiciary Also Addresses Generative AI Issues
California’s judicial branch has not ignored the risks of generative AI either.
Standard 10.80 of the California Standards of Judicial Administration addresses judicial officers’ use of generative AI in adjudicative work. It covers confidentiality, discriminatory effects, accuracy, and correction of erroneous output. It also calls for considering disclosure when AI is used to create material provided to the public.
Separately, California Rules of Court, rule 10.430, requires courts that do not prohibit generative AI use by court staff or judicial officers to adopt use policies. Those policies must address confidentiality, accuracy, bias, disclosure, and compliance with applicable law and professional standards.
Arbitrators and Alternative Dispute Resolution
SB 574 prohibits arbitrators from delegating any part of their decision-making to AI and regulates their use of AI-generated information outside the evidentiary record. However, California already has safeguards. The in the California Rules of Court, Ethics Standards for Neutral Arbitrators in Contractual Arbitration. Standard 5 requires arbitrators to maintain impartiality and uphold the integrity and fairness of arbitration, Standard 7 sets out disclosure obligations, and Standard 10 addresses disqualification. Code of Civil Procedure sections 1281.9 and 1281.91 also govern disclosure and disqualification, and section 1286.2 provides statutory grounds for vacating awards.
SB 574 makes the restrictions on AI use more explicit, but the question remains whether these existing safeguards were inadequate to address the underlying concerns.
What Does the New Law Add?
SB 574 requires disclosure of AI use in court filings and personal verification of legal citations. Both address legitimate concerns, but courts already have tools to handle them through standing orders and court rules.
AI disclosure: SB 574 does not say what the disclosure must contain or where it must appear, and it can be read to cover only AI-assisted filings or every filing by an attorney who uses AI at all. Courts, by contrast, can tailor disclosure requirements, define what level of AI assistance triggers disclosure, and exclude routine tasks such as formatting and proofreading.
Citation verification: Attorneys should already verify legal authorities before relying on them in court submissions. California Rules of Court, rule 1.200, lets citations follow either the California Style Manual or The Bluebook: A Uniform System of Citation, which gives attorneys established standards for consistent citation. Citation style does not establish whether a case exists or supports a particular proposition, so attorneys remain responsible for checking the underlying authorities. Courts have reinforced the verification duty through local rules. For example, Mendocino County Superior Court Local Rules 1.54–1.55, effective July 1, 2026, require the filer of an AI-assisted document to check every citation, fact, and argument for accuracy, and provide that this duty cannot be delegated to AI.
Courts can therefore impose tailored verification and disclosure requirements through their existing rulemaking authority, distinguishing substantive AI assistance from routine editing.
Conclusion
SB 574 addresses legitimate concerns about AI in legal practice, but those concerns do not automatically establish a need for more legislation. California already has professional conduct rules governing competence, diligence, confidentiality, and candor, judicial standards addressing AI use, and established mechanisms for sanctions and attorney discipline. Given these safeguards, the question is what additional protection SB 574 provides and whether that benefit justifies another layer of statutory regulation.
The strongest argument for SB 574 is uniformity. Court-by-court rules create a patchwork that attorneys practicing in several counties must track. Whether that concern justifies statewide statutory mandates is the question.
Another question is whether other states will follow California and enact AI-specific legislation for attorneys and arbitrators. Alternatively, they may continue to rely on guidance from the American Bar Association, state bars, and the judiciary. Either way, these choices will shape how attorneys can use AI and how much flexibility they keep to adopt new technologies while meeting their ethical duties to clients and the courts.
_____________________________________________________________
Collaborate with us!
As always, we appreciate you taking the time to read our blog post.
If you have news relevant to our global WAI community or expertise in AI and law, we invite you to contribute to the WAI Legal Insights Blog in 2016! To explore this opportunity, please contact WAI editors Silvia A. Carretta - WAI Chief Legal Officer (via LinkedIn or silvia@womeninai.co) or Dina Blikshteyn (dina@womeninai.co).
Silvia A. Carretta and Dina Blikshteyn
- Editors
Comments