This is not California state Senator Scott Wiener’s first attempt at addressing the dangers of AI.
In 2024, Silicon Valley mounted a fierce campaign against his controversial AI safety bill, SB 1047, which would have made tech companies liable for the potential harms of their AI systems. Tech leaders warned that it would stifle America’s AI boom. Governor Gavin Newsom ultimately vetoed the bill, echoing similar concerns, and a popular AI hacker house promptly threw a “SB 1047 Veto Party.” One attendee told me, “Thank god, AI is still legal.”
Now Wiener has returned with a new AI safety bill, SB 53, which sits on Governor Newsom’s desk awaiting his signature or veto sometime in the next few weeks. This time around, the bill is much more popular or at least, Silicon Valley doesn’t seem to be at war with it.
Anthropic outright endorsed SB 53 earlier this month. Meta spokesperson Jim Cullinan tells TechCrunch that the company supports AI regulation that balances guardrails with innovation and says “SB 53 is a step in that direction,” though there are areas for improvement.
Former White House AI policy advisor Dean Ball tells TechCrunch that SB 53 is a “victory for reasonable voices,” and thinks there’s a strong chance Governor Newsom signs it.
If signed, SB 53 would impose some of the nation’s first safety reporting requirements on AI giants like OpenAI, Anthropic, xAI, and Google — companies that today face no obligation to reveal how they test their AI systems. Many AI labs voluntarily publish safety reports explaining how their AI models could be used to create bioweapons and other dangers, but they do this at will and they’re not always consistent.
The bill requires leading AI labs — specifically those making more than $500 million in revenue — to publish safety reports for their most capable AI models. Much like SB 1047, the bill specifically focuses on the worst kinds of AI risks: their ability to contribute to human deaths, cyberattacks, and chemical weapons. Governor Newsom is considering several other bills that address other types of AI risks, such as engagement-optimization techniques in AI companions.
Techcrunch event
San Franci …