- Criticism from Former OpenAI Staff
- OpenAI's Response and Position
- Anthropic Supports SB 1047 with Amendments
Two former OpenAI researchers, Daniel Kokotajlo and William Saunders, have expressed their discontent with the company’s stance on California’s AI regulation bill, known as SB 1047. They had quit the company earlier this year, citing safety issues arising from the high rate of AI development.
Criticism from Former OpenAI Staff
In a letter to Politico, they expressed their disagreement with OpenAI’s stance on this bill, saying that it contradicts the company’s earlier appeals for AI regulation. In their letter, Kokotajlo and Saunders raise a broader concern for the management and safety of AI. According to them, OpenAI's opposition to SB 1047 reveals the organization’s irresponsible attitude toward the creation of AI.
OpenAI's Response and Position
In response to the letter from the former researchers, an OpenAI representative shared a statement denying the allegations of mischaracterization. The spokesperson also noted that OpenAI backs some of the AI safety bills in Congress and claims that these should be regulated at the federal level because they pose a national security threat and affect the country’s competitiveness on the world stage.
Anthropic Supports SB 1047 with Amendments
Meanwhile, OpenAI’s competitor, Anthropic, has welcomed SB 1047 but has proposed some changes. The company’s CEO, Dario Amodei, who used to work at OpenAI, was more moderate in his opinion about the bill. In a letter to Governor Gavin Newsom, Amodei stated that although the bill has its pros, the current version with amendments is cost-effective and beneficial.
Anthropic’s endorsement of SB 1047 and the proposed changes signal a different approach to AI governance. Their stance contrasts with OpenAI’s federal regulation approach. The fate of SB 1047 will likely impact future debates at both the state and federal levels.
