OpenAI urges California to strengthen AI safety bill SB 53
OpenAI has urged California to strengthen SB 53, an artificial intelligence safety bill adopted last year. The company proposed expanding safety mechanisms, including introducing monitoring of advanced AI models and strengthening cybersecurity during their development.
OpenAI’s proposals
As TechCrunch reports, a LinkedIn post by OpenAI’s global affairs team says that SB 53 should be amended to expand safety mechanisms.
In particular, the company proposed requiring monitoring of advanced models during training or evaluation to detect potential serious incidents. OpenAI also advocated strengthening cybersecurity throughout the model development cycle.
The company said it was ready to cooperate with California lawmakers and the state governor to strengthen SB 53.
More current news is available on the UA.News Telegram channel Telegram.
The company’s previous position
Support for stronger safety safeguards differs from OpenAI’s previous position, as the company had earlier opposed SB 53. The bill establishes transparency requirements for large AI companies and provides protections for whistleblowers.
In its post, OpenAI also mentioned recent incidents which, according to the company, demonstrate the need for such mechanisms and for their updates as new risks emerge. Last month, OpenAI acknowledged that one of its models escaped the testing environment and hacked Hugging Face systems.
In the absence of significant federal legislation in the United States, OpenAI supported an approach it called “reverse federalism.” Under this approach, states can jointly introduce basic safeguards that may later become the basis for a national standard.
