Happy New Year!  New laws are taking effect as of January 1, including California SB 53.
https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53
SB 53 requires frontier model developers to publish an AI risk and safety framework and to disclose catastrophic risks, defined in the statute as risks that can materially contribute to large-scale loss of life or more than $1B in damage.

While parts of SB 53 may be challenged by DOJ litigation or narrowed through the President’s recent Executive Order, key elements are likely to survive, especially:
• requirements to document and publish AI risk policies
• expectations around risk assessment and escalation
• transparency around low-probability, high-impact outcomes

One practical question this raises for finance and governance teams:
How can we determine what is “catastrophic” if we don’t measure the size of AI risk at all?

SB 53 doesn’t mandate outcomes, but it does normalize the idea that AI risk can be defined, sized, documented, and disclosed. That framing is likely to persist well beyond California.

#AIlaw #SB53 #AI