California First to Mandate AI Transparency with New Law
California Governor Gavin Newsom has signed the nation's first AI transparency law, Senate Bill 53. The law aims to enhance safety and accountability in the development and deployment of advanced AI models, specifically targeting 'frontier models' trained with more than 1026 operations. It imposes extra obligations on firms with annual revenues over $500 million, requiring them to publish and update AI frameworks, include detailed disclosures in transparency reports, report critical safety incidents, and protect whistleblowers from retaliation.
Newsom previously vetoed a tougher bill, SB 1047, due to its limitations in managing AI models only for the largest ones. The new law addresses these concerns, taking effect on January 1, 2026, with additional time allocated for setting up the public computing cluster, CalCompute.
CalCompute, established by the law, will provide resources for AI startups and researchers to develop safe, ethical, equitable, and sustainable AI. However, critics argue that the law's model size threshold and focus on paperwork may not effectively regulate AI.
The signing of SB 53 marks a significant step in AI regulation. It remains to be seen how effective the law will be in balancing transparency and innovation. Major tech companies, including Meta, Amazon, Google, and Microsoft, who have criticized stricter transparency regulations, will be among those affected by the new law.
Read also:
- Trade Disputes Escalate: Trump Imposes Tariffs, India Retaliates; threatened boycott ranges from McDonald's, Coca-Cola to iPhones
- Aquatech purchases Koch's Direct Lithium Extraction business, merging Li-ProTM DLE technology into the PEARLTM Technology Platform.
- Nepal's Journey: Evolution from Street Life to Political Power
- Li Auto faces scrutiny after crash test involving i8 model and a truck manufacturer sparks controversy
