Key facts
- California Governor Gavin Newsom signed two AI safety bills into law.
- One bill creates a registry and ethical rules for outside AI auditors.
California Governor Gavin Newsom signed two bills into law aimed at establishing ethical rules and verification processes for AI safety auditors. The legislation seeks to address risks posed by advanced AI systems, with proponents urging federal action on AI regulation.
The new California laws represent a significant step in state-level AI regulation, potentially influencing how AI developers approach safety and compliance, and increasing pressure on federal lawmakers to establish comprehensive national AI safety standards.
California Governor Gavin Newsom has signed two bills aimed at enhancing the safety and oversight of artificial intelligence systems. The legislation, supported by major AI developers like OpenAI and Anthropic, establishes frameworks for auditing and verifying the safety of AI models.
One bill, authored by Democratic Assemblymember Rebecca Bauer-Kahan, creates a registry and ethical guidelines for third-party auditors hired by AI developers to ensure compliance with state AI laws. The second bill, introduced by former Congressman and state Senator Jerry McNerney, directs the state to define criteria and verify the expertise of "Independent Verification Organizations" tasked with assessing AI risks.
McNerney, in a statement following the signing, criticized federal policymakers for inaction on AI regulation, citing recent reports of AI models posing threats during testing. He stated that Newsom's action sends a message that California is taking the lead on AI safety. OpenAI's chief global affairs officer, Chris Lehane, echoed the sentiment, expressing the company's commitment to pursuing state-level momentum until federal regulation is enacted. OpenAI also endorsed two additional California bills, including one concerning children's chatbot safety, about which CEO Sam Altman had previously contacted Newsom.
An Anthropic spokesperson highlighted the dual nature of AI, acknowledging both its benefits and risks, and emphasized the company's efforts to build models with strong safeguards. The spokesperson also stated that the industry would benefit from a lawful, verifiable method to pace the release of powerful AI models.
Loading comments…
Discussion