More than one hundred AI workers have gone up against their employers by publicly supporting an AI safety bill.
The SB 1047 bill puts responsibility squarely with developers of AI models if they haven’t met the safety measures set out. However, not all AI companies on onboard, despite how their employees may feel.
The bill does have the support of Anthropic, which is working hard to catch up with OpenAI.
What is the Wiener Bill?
This AI safety bill has now passed both houses of California’s Legislature. It targets developers of AI models that cost $100 million or more to train; but also only developers based in California. It’s all about where the liability lies.
It was created by state Senator Scott Wiener but has the support of some tech hard-hitters including former Google Brain researcher and Turing Award winner, Geoffrey Hinton and Yoshua Bengio. Despite calling Wiener a “pedophile-apologist”, Elon Musk is also a proponent.
This just in! View
the top business tech deals for 2024 👨💻
The Bill will now sit with Governor Gavin Newsom, who has until September 30 to make his decision as to whether to sign or veto it.
Who is Against Weiner Bill?
Leading names among the opponents including Speaker Emerita Nancy Pelosi, San Francisco Mayor London Breed, and Trump-supporting venture capitalists Marc Andreessen and Ben Horowitz.
Congresswoman Zoe Lofgren who represents part of Silicon Valley is also against and wrote a letter against the bill, signing alongside seven other California House Democrats. Her concerns include the fact that the Bill was not informed by technology experts.
AI Employees Turn Against Their Bosses
The letter, which was published on Monday, has been signed by 113 people who currently or previously worked for AI big hitters.
And some of them are high up in the ranks. Anthropic co-founder Chris Olah; former OpenAI researcher Jan Leike; and OpenAI whistleblowers Daniel Kokotajlo and William Saunders have all signed.
The San Francisco Standard reports that “thirty-seven signatories are employees of companies opposing the bill: 19 work for Google DeepMind, 15 are at OpenAI, and three are at Meta”. On top of this, there are alumni of these companies as well as anonymous signatories.
The OpenAI employees could face the wrath of CEO Sam Altman who has taken questionable steps in the past to silence naysayers. According to Vox, the company tied outgoing employees’ vested equity hostage to their signing of lifetime NDA and non-disparagement agreements. OpenAI dialed this back after a public outcry.
What Are Employees Main Concerns?
The signatories warn in the letter “that the most powerful AI models may soon pose severe risks, such as expanded access to biological weapons and cyberattacks on critical infrastructure.” They argue that the Bill “represents a meaningful step forward” and it is “feasible and appropriate” to safeguard against the potential harms that AI could cause.
It follows a letter published in August by Dario Amodei, Chief Executive Officer of Anthropic, which gave its approval to changes made to the original bill. It stated that there were “still some aspects of the bill which seem concerning or ambiguous to us” but “some aspects have the potential to meaningfully improve our ability to prevent serious risks from AI systems”.
Voices on both sides are bound to become increasingly vocal as the deadline for Newsom approaches.
