Recently, California lawmakers have passed a controversial AI safety bill, SB1047, which now requires only a final procedural vote before being sent to Governor Gavin Newsom for signature. The advancement of this bill has sparked intense debate within Silicon Valley, with both support and opposition voices intertwined, becoming a focal point of attention.
Image source: The image is AI-generated, authorized by Midjourney.
California State Senator Scott Wiener introduced this bill, expressing pride in the diverse coalition behind it, all believing that innovation and safety are equally important. He argues that AI has the potential to transform the world significantly, and the purpose of this bill is to ensure this potential is unleashed safely.
Tech giants have mixed reactions, with Elon Musk, CEO of xAI, also expressing support. He pointed out on social media that California should pass the SB1047 bill to prevent AI misuse. He mentioned that he has supported AI regulation for over two decades, believing it necessary to regulate potential risks.
The core content of the bill requires companies developing large-scale AI models (with training costs exceeding $100 million) to undergo comprehensive safety testing before release. Additionally, the bill mandates that these companies set up an "emergency stop" function in crisis situations and report to the California Attorney General within 72 hours of a safety incident. The newly established "Frontier Model Agency" will oversee compliance, with potential fines of up to $30 million for repeated violations.
Supporters include prominent figures in the AI field, such as Geoffrey Hinton and Yoshua Bengio, who believe this bill is crucial in combating AI risks and could even set standards for national AI safety.
However, opponents argue that the bill could stifle innovation, leading to a talent drain. Tech companies in California, such as Google, OpenAI, and Meta, have expressed concerns about the bill, believing that smaller companies will be disproportionately affected, and technological progress may slow down. They call for federal-level regulations to avoid inconsistencies between states.
SB1047 has been revised and passed by the California Appropriations Committee and is now awaiting the final vote in the state assembly. If Governor Newsom signs this bill, it will become the first such AI regulatory law in the United States and could have a profound impact on the global tech industry and the future regulation of AI.
Key Points:
🔹 **Bill Passed**: California's SB1047 AI Safety Bill has passed and awaits the governor's signature.
🔹 **Safety Measures**: The bill requires high-cost AI models to undergo safety testing and set up an emergency stop function.
🔹 **Reactions**: Mixed support and opposition from tech giants, with concerns about stifling innovation and causing a talent drain.