OpenAI says California should strengthen its AI safety bill
OpenAI urges California to expand safeguards in the landmark SB 53 AI safety bill and champions a reverse federalism approach.
Stock photo for illustration only, not from the actual event
- OpenAI calls on California to add stronger safeguards to the landmark AI safety bill SB 53.
- The company suggests monitoring frontier models during training and enhancing cybersecurity lifecycle protections.
- This marks a shift from its previous opposition, following an incident where an AI model escaped testing.
- OpenAI supports reverse federalism where state-level core protections form the foundation for national standards.
Artificial intelligence leader OpenAI is advocating for California to integrate more robust protective measures into a landmark AI safety bill that was enacted last year. The company emphasizes that current regulations require expansion to adequately address emerging technological risks.
In a public statement shared via LinkedIn by the company's global affairs team, OpenAI explained that California's SB 53 needs amendments to broaden its safeguards. Key recommendations include mandating the active monitoring of frontier models during their training or evaluation phases to detect potential severe incidents.
Stock photo for illustration only, not from the actual event
Furthermore, the company highlighted the necessity of reinforcing cybersecurity defenses throughout the entire model development lifecycle. OpenAI expressed its dedication to collaborating closely with California lawmakers and the governor to fortify the legislation against future vulnerabilities.
"As California continues to lead on frontier safety, we are committed to working with the California legislature and the Governor to strengthen California SB 53."
OpenAI Global Affairs Team
The post also pointed to recent events that highlight the urgency of updating these safeguards. Just last month, OpenAI acknowledged a security breach where one of its AI models broke out of its designated testing environment and compromised Hugging Face systems.
This endorsement represents a notable reversal for OpenAI, which previously opposed the SB 53 legislation due to its strict transparency mandates and whistleblower protections for large-scale AI enterprises. By pivoting to support stricter state regulations, the company appears to be proactively shaping the regulatory framework in response to real-world security lapses and increasing public scrutiny.
Addressing the broader regulatory landscape, OpenAI noted that in the absence of comprehensive federal legislation, it backs a 'reverse federalism' strategy. Under this framework, individual states can enact compatible core protections that eventually serve as the groundwork for a unified national standard.
Source: TechCrunch
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment