Back
Gibson Dunn: Regulating the Future -- Eight Key Takeaways from SB 1047
webAuthor
Richard Manfredi
Legal analysis from Gibson Dunn on California's SB 1047, a landmark but ultimately vetoed AI safety bill; useful reference for understanding state-level AI governance debates and the regulatory challenges of overseeing frontier AI models.
Metadata
Importance: 58/100opinion pieceanalysis
Summary
Gibson Dunn law firm analyzes California's SB 1047, a significant AI safety bill that would have imposed safety requirements on developers of large frontier AI models, which was ultimately vetoed by Governor Newsom in 2024. The analysis covers the bill's key provisions, compliance obligations, and implications for AI developers. It provides a legal practitioner's perspective on the regulatory landscape for frontier AI.
Key Points
- •SB 1047 would have required developers of large AI models (above compute thresholds) to implement safety protocols and conduct hazard assessments before deployment.
- •The bill proposed creation of a Frontier Model Division within California's Department of Technology to oversee compliance and enforcement.
- •Governor Newsom vetoed the bill citing concerns it was too broad and could stifle innovation, particularly by applying to models regardless of their actual risk profile.
- •The law would have created significant liability exposure for AI developers, including potential civil penalties and whistleblower protections for employees.
- •The veto does not end state-level AI regulation efforts; California and other states continue to pursue alternative AI governance frameworks.
Cited by 2 pages
| Page | Type | Quality |
|---|---|---|
| Safe and Secure Innovation for Frontier Artificial Intelligence Models Act | Policy | 66.0 |
| US State AI Legislation Landscape | Analysis | 70.0 |
Cached Content Preview
HTTP 200Fetched Apr 3, 202615 KB
Regulating the Future: Eight Key Takeaways from California’s SB 1047, Vetoed by Governor Newsom - Gibson Dunn
Regulating the Future: Eight Key Takeaways from California’s SB 1047, Vetoed by Governor Newsom - Gibson Dunn
Skip to content
Regulating the Future: Eight Key Takeaways from California’s SB 1047, Vetoed by Governor Newsom
Client Alert | September 30, 2024
Newsom committed to working with legislators, academics, and other partners to “find the appropriate path forward, including legislation and regulation.”
Update : On September 29, 2024, Governor Newsom vetoed SB 1047 by returning it to the legislature without his signature, criticizing the bill as “a solution that is not informed by an empirical trajectory analysis of AI systems and capabilities.” [1] Building on concerns he previously expressed about the bill, [2] Newsom explained in a statement accompanying his veto that SB 1047 regulates models based only on their cost and size, rather than function, and fails to “take into account whether an Al system is deployed in high-risk environments, involves critical decision-making or the use of sensitive data.” [3]
Despite the veto, Newsom voiced support for AI regulation and California’s role in these efforts, stating, “Safety protocols must be adopted. Proactive guardrails should be implemented, and severe consequences for bad actors must be clear and enforceable,” and that California “cannot afford to wait for a major catastrophe to occur before taking action to protect the public.” [4] Newsom committed to working with legislators, academics, and other partners to “find the appropriate path forward, including legislation and regulation.” [5]
On August 28, 2024, the California State Assembly passed proposed bill SB 1047 , the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act, through which California seeks to regulate foundational AI models and impose obligations on companies that develop, fine-tune or provide compute resources to train such models. The bill purports to regulate only the most powerful AI models, trained using large computing capacity, but its requirements are likely to have a broader impact, including on open source models.
SB 1047 currently sits with Governor Newsom. As of September 24, it is unclear whether the Governor will sign the bill or veto it; on September 17, Newsom signaled some discomfort with the bill, but stated that he remained undecided even as he signed several other AI-related bills into law. [6] Gov. Newsom has until the end of September to sign or veto the bill; if he does not veto or return the bill to the legislature, SB 1047 will become law and take effect on January 1, 2026, even if he does not sign it.
Controversial since its introduction, SB 1047 represents a major shift in
... (truncated, 15 KB total)Resource ID:
006e7685ee710738 | Stable ID: MzNkNmZhOD