California SB 574 and AI in Legal Services: What Clients Should Know
California SB 574 would establish specific responsibilities for attorneys who use generative AI, including protecting confidential information, independently verifying AI-generated material, personally checking citations, and disclosing AI use in documents submitted to a court. The bill would also prohibit attorneys from allowing AI to replace their own legal judgment. SB 574 has passed the California Legislature but has not yet become law.
Artificial intelligence is becoming increasingly common in the legal profession. AI-powered tools may assist with organizing information, reviewing documents, conducting research, and preparing preliminary drafts. Their growing use, however, raises important questions about confidentiality, accuracy, transparency, and professional responsibility.
California Senate Bill 574 attempts to address some of those questions. If signed into law, the bill would add specific requirements governing how California attorneys may use generative artificial intelligence in their legal work.
What Would SB 574 Require?
As currently written, SB 574 would require attorneys using generative AI to:
- Retain responsibility for legal work rather than delegating the practice of law to AI.
- Protect confidential, personally identifying, and other nonpublic information entered into AI systems.
- Take reasonable steps to verify the accuracy of AI-generated material.
- Correct erroneous or fabricated AI output included in material used by the attorney.
- Personally verify citations contained in documents filed with a court.
- Disclose the use of generative AI to the court for documents submitted to the court.
- Consider whether AI use should be disclosed when it helps create content provided to the public.
The bill also addresses the use of AI by arbitrators. It would prohibit arbitrators from delegating any portion of their decision-making responsibilities to an AI tool or relying on AI-generated information outside the record without appropriate disclosure to the parties.
For clients, these proposed requirements center on several familiar expectations of legal representation: confidentiality, accuracy, transparency, and meaningful attorney oversight.
AI Can Assist a Lawyer, but It Cannot Replace One
One of the central principles of SB 574 is that an attorney may not delegate the practice of law to generative AI.
Technology may help an attorney complete certain tasks, but it cannot assume the attorney’s professional responsibilities. Attorneys must still exercise independent judgment, consider the individual circumstances of each matter, and remain accountable for work performed on a client’s behalf.
This reflects a broader principle already emphasized by the State Bar of California: A lawyer remains fully responsible for legal work produced with the assistance of AI.
Protecting Confidential Client Information
The proposed legislation also addresses the information attorneys enter into generative AI systems.
Legal matters frequently involve sensitive personal, financial, medical, or business information. Under SB 574, an attorney could not enter confidential or nonpublic information into a generative AI system unless access to that information is restricted to the attorney and other authorized individuals who are obligated to protect it.
The provision highlights an important consideration as legal technology evolves: The convenience offered by an AI tool must not come at the expense of client confidentiality.
Attorneys Would Remain Responsible for Accuracy
Generative AI tools can produce information that appears convincing but is incomplete, inaccurate, or entirely fabricated. These errors are sometimes called AI “hallucinations.”
SB 574 would require attorneys to take reasonable steps to verify AI-generated output and correct any erroneous or hallucinated material they use. It would also amend California’s rules concerning court filings to state that an attorney responsible for submitting a document may not include a citation that the attorney has not personally verified.
Significantly, the personal-verification requirement would apply to every citation in a covered court filing, not only citations supplied by AI.
Disclosure of AI Use in Court Documents
One of the bill’s most notable provisions would require attorneys to disclose their use of generative AI to the court for all documents submitted to the court.
Exactly how this requirement would operate remains uncertain. The bill does not establish a particular disclosure format or explain how the requirement would apply when AI plays only a limited role. For example, it does not expressly address whether the same disclosure would be required when AI helps revise a small portion of existing language as when it generates a substantial first draft.
These practical questions may require additional guidance if SB 574 becomes law. For clients, however, the underlying principle is straightforward: Attorneys remain responsible for the documents and legal work produced on their behalf, regardless of the technology used to assist them.
Part of a Broader Conversation About AI and the Law
SB 574 arrives as the State Bar of California continues to evaluate how existing professional obligations apply to rapidly developing AI tools.
Current State Bar guidance emphasizes duties involving competence, diligence, confidentiality, communication, supervision, billing, and candor to the court. It also addresses emerging “agentic AI” systems that can perform multistep tasks with greater autonomy.
SB 574 would build upon that broader conversation by placing certain AI-related responsibilities directly into California law.
What Happens Next?
SB 574 has passed both houses of the California Legislature and is awaiting action by the governor. Until that process is complete, its provisions remain proposed requirements rather than binding law.
The bill reflects a larger principle that will remain important as legal technology evolves: AI may assist with legal work, but it cannot replace an attorney’s judgment, accountability, or understanding of a client’s individual circumstances. The practice of law must remain fundamentally human.
For clients, responsible innovation should offer the benefits of useful technology without sacrificing the careful judgment, confidentiality, personal guidance, and trusted relationships at the heart of legal representation.