California’s SB 1047, the “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act,” has passed the state’s Assembly and Senate and now awaits Governor Gavin Newsom’s decision by September 30. This legislation aims to regulate large language models by introducing stringent testing, safety, and enforcement standards.
Whistleblower Protections
One of the critical aspects of SB 1047 is its provision for whistleblower protections. Employees who report risks of significant harm related to AI models to the California Attorney General are shielded from retaliation. This move encourages transparency and accountability in AI development, ensuring that potential ethical and safety concerns are promptly addressed. Companies must also establish internal processes for anonymous disclosures and educate employees about their rights and responsibilities regarding these protections.
Scope of the Law
SB 1047 targets AI models that exceed specific computational and financial thresholds:
- Models using more than 10^26 operations and costing over $100 million.
- Fine-tuned models requiring over three times 10^25 operations and costing over $10 million.
These criteria ensure that only the most resource-intensive and potentially impactful AI models are subject to the bill’s regulations.
Support and Opposition
The bill has garnered support from over 120 current and former employees of major AI companies like OpenAI, Anthropic, Google’s DeepMind, and Meta. They argue that powerful AI models could pose severe risks, such as facilitating biological weapons or cyberattacks, and thus require stringent oversight.
Academic experts have also endorsed the bill, describing it as a significant step forward in mitigating AI-related risks. However, the bill faces opposition from industry leaders like OpenAI, which argues that it could stifle innovation. Political figures such as Representative Nancy Pelosi and Mayor London Breed, along with organizations like the U.S. Chamber of Commerce, also oppose the bill.
Business Implications
For corporate decision-makers, SB 1047 represents a critical juncture in AI governance. On one hand, the bill aims to ensure the safe deployment of AI technologies, which could prevent catastrophic misuse. On the other hand, the regulations could impose additional compliance costs and slow down innovation.
Conclusion
SB 1047 is a landmark piece of legislation that could set a precedent for AI regulation nationwide. Business leaders must weigh the benefits of enhanced safety and transparency against the potential constraints on innovation. As Governor Newsom’s decision looms, companies should prepare for the implications of this bill and consider how to adapt their AI strategies accordingly.
