California has turned guidance on AI in legal work into statutory requirements: lawyers must verify AI-generated material and cannot delegate the practice of law to generative AI. Senate Bill 574, signed by Governor Gavin Newsom on 30 September, takes effect on 1 January, according to PYMNTS’ account of an analysis by Mayer Brown attorneys.
Watch Desk analysis
What happened
The account describes rules for lawyers covering verification, confidentiality and disclosure. Attorneys must take reasonable steps to check AI output, including every case and statutory citation, and correct errors. Court documents must contain only citations the responsible attorney has personally read and verified, and attorneys must disclose generative AI use in documents submitted to courts.
The restrictions also cover arbitrators: they may not delegate any part of the decision-making process to generative AI. If they rely on AI-generated information outside the record, they must disclose it to the parties and, as far as practical, allow them to comment.
Why it matters
The practical line is not “AI forbidden” versus “AI allowed”. The rules described here permit assistance but keep responsibility with legal professionals, including responsibility for checking authorities and protecting confidential material. That makes tool access and review procedures part of legal practice, not merely a matter of convenience.
For arbitrators, disclosure of outside-record AI information gives parties a chance to respond before it influences a decision. The statute gives those expectations a legal footing, rather than leaving them only in professional guidance.
Our read
This is a meaningful policy change with concrete steps for legal teams: verify citations, control what information goes into AI systems, and disclose use where required. The detail matters, and so does the source boundary: PYMNTS reports the provisions through a Mayer Brown analysis, rather than quoting the statute’s full text. Lawyers and arbitrators should check the law and their own professional obligations before changing practice.
What to watch
- How California courts and legal professionals interpret the verification and disclosure requirements.
- What counts as delegating legal practice or arbitration decision-making to AI.
- Whether similar rules appear in other states or arbitration institutions.
Discussion spark: Should lawyers and arbitrators have to disclose every use of generative AI in submitted work, or only uses that shape legal reasoning or decisions?
Sources and evidence
- Source update (12 November 2026, 16:32 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.