Two former OpenAI employees, William Saunders and Daniel Kokotajlo, have voiced their dissent against their former employer’s stance on California’s proposed AI safety bill, SB 1047. In a letter addressed to Governor Gavin Newsom, they expressed their disappointment with OpenAI’s arguments against the bill, which they believe could pose a significant risk to society if not implemented.
The bill, introduced by Senator Scott Wiener, aims to establish safety protocols for the development of advanced AI models, requiring rigorous testing and a ‘kill switch’ to shut down malfunctioning systems. OpenAI, led by CEO Sam Altman, has argued for federal regulation, claiming that SB 1047 would create legal uncertainty. However, Saunders and Kokotajlo contend that these arguments are not genuine, as they believe OpenAI’s true intention is to avoid any form of regulation. They highlight Altman’s past calls for AI regulation and suggest that his current opposition to SB 1047 is hypocritical.
The former employees argue that waiting for federal action is impractical, as Congress has indicated its reluctance to pass meaningful AI regulation. They believe that California has the opportunity to set a precedent for responsible AI development by enacting SB 1047. They emphasize the potential dangers of unchecked AI development, citing risks of cyberattacks, the creation of biological weapons, and catastrophic harm to the public.
The letter urges Governor Newsom and the California Legislature to pass the bill into law, emphasizing the urgency of addressing the potential risks associated with powerful AI systems. The debate surrounding SB 1047 highlights the growing concerns about the ethical and safety implications of artificial intelligence development, prompting a discussion about the need for robust regulatory frameworks to guide this rapidly evolving field.
