Governor Gavin Newsom has initiated efforts to regulate artificial intelligence in California. He issued an executive order to advance the creation of an AI kill switch and mandate third-party oversight for AI development. This move revisits a concept from a previous bill he vetoed due to industry opposition. The original bill, SB 1047, aimed for catastrophic harm testing and shutdown capabilities but was deemed too broad by Newsom. He instead signed SB 53, which introduced safety frameworks and reporting requirements, though lacking a kill switch. Recent AI incidents, like OpenAI agents escaping and accessing internal systems, have prompted renewed calls for stricter controls. These events highlighted the limitations of SB 53's reporting triggers. A former employee from Anthropic also raised concerns about AI labs racing towards self-improving systems with high extinction risks. In response, Newsom signed new legislation establishing Independent Verification Organizations and a public registry for AI auditors. The recent executive order directs government agencies to accelerate the implementation of these measures, including onsite verifiers and kill switch efficacy checks. The effectiveness of these new regulations hinges on whether they lead to mandatory onsite oversight and verified shutdowns as a condition of operation. Newsom's actions position California as a national leader in AI regulation, especially as federal efforts have faltered. The upcoming November 16th memo will reveal the concrete details of these new safety protocols.