San Francisco-based AI company Anthropic has endorsed California’s revised AI regulation bill, SB 1047, citing its benefits likely outweigh its costs. The bill, advanced by State Senator Scott Wiener, aims to ensure the safety of advanced AI models by requiring developers to implement kill switches and outlining methods for handling AI malfunctions.
Key Changes and Industry Reactions:
The revised bill removes the proposed government AI oversight committee, a change that Anthropic and others in the tech industry welcomed.
However, major tech firms like Alphabet and Meta have expressed concerns, arguing the bill could stifle AI innovation in California.
Anthropic Position:
In a letter to California Governor Gavin Newsom, Anthropic CEO Dario Amodei acknowledged improvements in the bill, though he noted some remaining ambiguities.
He highlighted that the updated version addresses initial concerns about potentially hindering AI development.
The bill’s provisions include safety testing for costly or high-computational AI models and give the state attorney general authority to enforce compliance.