Two former OpenAI employees, who resigned this year over safety concerns, have expressed their disappointment at the company's decision to oppose California's SB 1047 bill, which aims to prevent disasters related to artificial intelligence. Daniel Kokotajlo and William Saunders, who previously warned about OpenAI’s "reckless race" for market dominance, note that while the company's stance doesn’t surprise them, it does disappoint them.
Safety and Regulation Concerns
In a letter, Kokotajlo and Saunders criticize the apparent contradiction in OpenAI's position. "Sam Altman, our former boss, has repeatedly called for AI regulation," they write. "Now, when real regulation is on the table, he opposes it." The two researchers emphasize that with appropriate regulation, they still hope that OpenAI can fulfill its mission to develop artificial general intelligence (AGI) safely.
This letter comes at a time of tension within the AI sector, where the need for regulations to avoid potential risks is weighed against concerns about how these rules might impact innovation. The SB 1047 bill has become a focal point of this debate, especially in California, a key state for technological development.
OpenAI’s Response and Other Companies' Stance
In response to the former employees' criticism, an OpenAI spokesperson stated that the company "strongly disagrees with the mischaracterization of our position on SB 1047." The spokesperson explained that OpenAI supports the implementation of AI safety regulations at the federal level, given their implications for national security and global competitiveness, and has endorsed legislation in Congress that reflects this view.
On the other hand, Anthropic, a competitor of OpenAI, has expressed support for SB 1047, albeit with some reservations and suggested amendments. Some of these amendments have already been incorporated into the bill, leading Anthropic CEO Dario Amodei to write to Governor Gavin Newsom, stating that the "benefits of the bill likely outweigh its costs," although he did not give it full endorsement.
The debate over SB 1047 highlights the complex relationship between technological innovation and the need to establish regulatory frameworks that ensure public safety and well-being without stifling progress in the field of artificial intelligence.
