California’s AI Governance: New Framework Proposed After SB 1047 Veto

In September of last year, all eyes were on Senate Bill 1047 as it made its way to California Governor Gavin Newsom’s desk. Ultimately, the bill was vetoed, marking a significant moment in the ongoing discussion surrounding artificial intelligence regulation.

Senate Bill 1047 aimed to impose stringent testing requirements on developers of large AI models, particularly those with training costs exceeding $100 million. The legislation sought to ensure that these models were evaluated for specific dangers before deployment. The veto sparked discontent among AI industry whistleblowers, while many large tech companies expressed relief at the decision.

However, the narrative did not conclude with the veto. Governor Newsom, believing that the legislation was overly rigid and not adaptable to the diverse landscape of AI development, initiated a new approach. He convened a group of leading AI researchers to devise an alternative plan that would not only support the growth of generative AI in California but also establish necessary guardrails to mitigate associated risks.

On Tuesday, the findings of this research group were published in a comprehensive report titled “California Report on Frontier Policy.” This 52-page document outlines a proposed framework for AI governance that reflects the rapid advancements in AI capabilities since the veto of SB 1047.

The authors of the report highlighted that AI models, particularly their chain-of-thought reasoning abilities, have significantly evolved in a short period. Through a combination of historical case studies, empirical research, modeling, and simulations, the researchers advocated for a new governance structure that emphasizes increased transparency and independent scrutiny of AI models.

The report’s recommendations aim to create a balanced approach to AI regulation, fostering innovation while ensuring safety and accountability. By advocating for more rigorous testing and evaluation processes, the researchers hope to address the concerns raised by industry whistleblowers and the public alike.

As California continues to be a leader in technology and innovation, the proposed framework could set a precedent for other states and countries grappling with similar challenges in AI governance. The ongoing dialogue surrounding AI regulation is crucial as society navigates the complexities and implications of these powerful technologies.

For more details on the report and its implications for AI governance in California, visit Read the full story at The Verge..