What are the key requirements of the CA SB 1047?

Would have required safety testing before training and deployment Would have required 'kill switch' capability to shut down AI systems Would have required third-party audits for models over $100M training cost

US State·Vetoed

California Safe and Secure Innovation for Frontier Artificial Intelligence Models Act (SB 1047)

Dates and status re-checked against primary sources on September 30, 2026. How we verify

Would have required developers of large AI models (over $100M in training costs) to implement extensive safety testing, 'kill switch' capabilities, and third-party audits. The legislature gave final passage on August 29, 2024, and Governor Newsom vetoed it on September 29, 2024 on grounds that the bill was too broad and could inhibit AI development.

What this means for your team

SB 1047 was vetoed and did not become law. It is referenced here because it shaped the AI policy landscape significantly: California's later SB 53 took a narrower, transparency-based approach to frontier models. The principles behind SB 1047 — safety testing, kill switches, third-party audits — continue to influence policy discussions even without the law.

Key requirements

Would have required safety testing before training and deployment
Would have required 'kill switch' capability to shut down AI systems
Would have required third-party audits for models over $100M training cost
Vetoed — does not currently create any legal obligations

California Safe and Secure Innovation for Frontier Artificial Intelligence Models Act (SB 1047)

What this means for your team

Key requirements

Further reading