概述
Governor Gavin Newsom vetoed the bill on September 29, 2024, arguing its size-based approach did not sufficiently account for deployment context and could cover basic functions. The veto ended that bill; later legislation such as SB 53 took a different path.
深入探討
SB 1047, the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act, passed the California Legislature in 2024 but did not become law. Governor Gavin Newsom returned it unsigned on September 29, 2024. The proposal would have established obligations for developers of models above specified compute and cost thresholds, required safety and security protocols, addressed catastrophic-risk testing and incident response, and created state oversight mechanisms. It also included provisions concerning shutdown capability and whistleblower reporting. Because the governor vetoed the bill, those provisions did not take effect as SB 1047. In his veto message, Newsom said the bill’s size-based trigger did not adequately account for whether a system was deployed in a high-risk environment, involved critical decision-making, or used sensitive data. He argued the approach could impose stringent standards on systems with basic functions merely because they belonged to a large model. The message also acknowledged catastrophic risks and argued California should use a more flexible, evidence-based approach. The veto was therefore framed as a disagreement over scope and design, not a claim that AI risks were unimportant. Debate over SB 1047 involved competing concerns about catastrophic risk, innovation, open-source development, state oversight, and whether compute thresholds reliably identify dangerous capabilities. Supporters emphasized safety protocols and accountability for powerful models. Critics argued the bill could be overly broad, create uncertainty, or burden development. Those arguments are context for the legislative debate; they are not a substitute for the bill text or the governor’s message. California later enacted SB 53, the Transparency in Frontier Artificial Intelligence Act, in 2025. SB 53 created a different framework with transparency, incident-reporting, and whistleblower provisions for covered frontier developers. It should not be described as SB 1047 taking effect under a new name. To understand the policy shift, compare each enacted text, its scope and effective dates, then distinguish operative law from proposed measures and executive statements.
戰略影響
風險與安全
災難性和日常的人工智慧危害都取決於誰了解風險以及誰能夠採取行動。
更明確的決策
民眾和專業素養決定強而有力的安全政策在政治上是否可行。
突破炒作
清晰的解釋可以減少炒作、實驗室公關和模糊道德劇場的影響。
The Future of Why California’s SB 1047 Was Vetoed
The debate continues through later statutes, implementation, and federal policy. SB 53 is now law and may be amended or interpreted through regulations and enforcement. SB 1047 remains a vetoed proposal and a reference point in the policy debate, not an active compliance regime. Recheck official legislative sources before describing California’s current frontier-model requirements. Monitor implementation guidance and amendments to SB 53, while treating SB 1047 as historical legislative context. Verify each claim against official bill records. Distinguish proposals from enacted requirements.
現實世界的實施
A reader distinguishes a bill’s proposed model thresholds from the requirements of a law that actually took effect.
A policy team compares the governor’s veto message with the enrolled bill to separate stated objections from the bill’s text.
A company tracks SB 53 as a later enacted law with its own definitions and reporting duties rather than calling it a re-enactment of SB 1047.
A journalist explains that the veto did not eliminate all California AI rules, including laws enacted separately on employment, content transparency, and consumer protection.
風險與防護欄
將存在風險視為科幻小說，同時能力複合。
混淆了表面產品安全與高度自治下的對準。
只給非英語和非專業觀眾留下低品質的資源。
實施路線圖
單獨的產品危害、誤用和失控/失調風險。
詢問哪些證據會改變您對時間表和嚴重性的看法。
比起行銷主張，更喜歡主要來源和具體評估。
確定一條行動路徑：職業、政策、資金或技能——而不僅僅是意識。
不斷探索
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Why California’s SB 1047 Was Vetoed quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
常見問題
What is Why California’s SB 1047 Was Vetoed?
California SB 1047 was a 2024 proposal that would have imposed safety and security duties on developers of very large AI models and created state oversight mechanisms. Governor Gavin Newsom vetoed the bill on September 29, 2024, arguing its size-based approach did not sufficiently account for deployment context and could cover basic functions. The veto ended that bill; later legislation such as SB 53 took a different path.
Did California SB 1047 become law?
Governor Newsom returned SB 1047 without signature on September 29, 2024.
What did the veto message criticize about the bill’s approach?
The governor argued the bill’s thresholds did not sufficiently consider context and could reach systems with basic functions.
What did SB 1047 propose for covered developers?
The proposal included safety and security requirements for developers of models meeting its thresholds.
How does SB 53 relate to the vetoed SB 1047?
SB 53 later enacted transparency and incident-reporting provisions under its own terms.
Which source explains Newsom’s reasons for the veto?
The official message records the governor’s stated reasons for returning SB 1047 unsigned.
繼續學習
相關指南
為此主題精選的更多指南