California’s AI law works by staying narrow
The law takes a narrow, transparency-first approach to regulating advanced “frontier” AI models, creating room for experimentation, while requiring timely disclosures that give the state the data it needs to address risks as they emerge.