AI employees are defying their employers to support SB 1047
In speaking out, they’re showing just how important the bill is
In the latest development in the saga of California's SB 1047, the AI regulation bill has found unexpected allies: the very employees building the technology it seeks to govern. More than 100 current and former staff from leading AI companies have signed a letter endorsing the bill and urging Governor Gavin Newsom to sign it into law. Most notably, the list includes 36 employees from OpenAI, DeepMind, and Meta — all of which are lobbying hard to kill the bill.
The letter pulls no punches, saying that “the most powerful AI models may soon pose severe risks such as expanded access to biological weapons and cyberattacks on critical infrastructure”. It goes on to argue that “it is feasible and appropriate for frontier AI companies to test whether the most powerful AI models can cause severe harms, and for these companies to implement reasonable safeguards against such risks”.
This stance isn't just a gentle disagreement with company policy: it's a direct contradiction of the party line. OpenAI's head of strategy, Jason Kwon, has argued — albeit with a conspicuous lack of detail — that the bill would hamper innovation and possibly lead companies like OpenAI to leave California. (It’s worth noting, though, that Kwon allowed employees to sign the letter disagreeing with the company.) Google and Meta have argued similarly, pushing the narrative that the bill’s requirements are unwieldy — despite them being incredibly similar to commitments companies have already made to the White House and international governments.
The employee letter, however, makes clear that a significant proportion of the people building advanced AI believe the bill is both urgent and necessary. They are deeply worried about the potential risks of advanced AI and, more importantly, they don't trust their employers to address these risks without government intervention.
Coming from technical and policy staff at the leading AI companies, the pleas are hard to ignore. When they’re joined by leading AI experts such as Geoffrey Hinton, Yoshua Bengio and Gary Marcus, a swathe of other academics, and the leadership of Anthropic, the warning becomes deafening.
And while those at the coalface of the technology support the bill, the opposition is primarily being led not by AI experts, but billionaire investors such as Marc Andreessen and Garry Tan, who have hired some of Gavin Newsom’s closest allies to lobby against it. They’re joined by trillion-dollar corporations like Google and Meta — groups with a track record of putting profit over the public interest. Even many of the seemingly independent scientists opposing the bill have significant conflicts of interest. The most notable, Fei-Fei Li, runs a $1 billion startup with investment from Andreessen Horowitz; Andrew Ng, meanwhile, runs a huge AI investment fund.
The question now is: who will Newsom listen to? Will he, like Nancy Pelosi, Zoe Lofgren, and London Breed, cave to the interests of the rich and powerful? Or will he heed the warnings of the workers actually building this technology, who are increasingly vocal about their concerns? With a deadline of September 30, we’ll know soon enough.