California’s AI law matters if review carries consequences

On October 1, California Governor Gavin Newsom signed into law state legislation setting rules for how lawyers may use generative AI in their work, Reuters Legal reports. Bloomberg Law, Business & Practice reports that the bill also addresses lawyer and arbitrator AI use. Legaltech News reports that SB 574 bars arbitrators from relinquishing their decision-making powers to generative artificial intelligence.

The important move is not that California permits or prohibits AI in the abstract. It is that California law places the use of generative AI inside obligations that still attach to legal work. Legaltech News describes stronger AI review responsibilities for lawyers, while Bloomberg Law, Business & Practice reports that the law gives firms difficult disclosure decisions about when their use of the technology requires disclosure. AI does not eliminate the consequences of bad legal work: it relocates the question from whether a system produced an output to whether a lawyer reviewed and disclosed its use as California law requires.

The enactment establishes a California rule framework, not a finding that AI-generated legal work is reliable. Reuters Legal says the legislation sets rules for how lawyers are permitted to use generative AI; it does not report that the law makes review unnecessary or that disclosure resolves every risk. Bloomberg Law, Business & Practice reports that firms must determine when use requires disclosure, which means the harder issue remains application to particular legal work. The arbitrator provision is narrower still: it concerns relinquishing decision-making powers, and does not by itself establish anything about the quality of lawyer review. A stronger reading would require California decisions applying the law to actual uses of generative AI and identifying the consequences of inadequate review or disclosure. That has not happened in the reported enactment.

The next fact to watch is a California court decision, filing or other public application of the law to a lawyer’s use of generative AI, especially one addressing whether disclosure was required or whether review was adequate. Such an application would show whether the new duties produce consequences in legal work rather than remaining rules on paper. We should measure the law by those applications, not by the existence of the technology or the legislation alone.