The California legislature has taken a bold step into artificial intelligence regulation with the passage of California AI Transparency Act. This groundbreaking legislation establishes new requirements for major AI providers operating in California. It is set to take effect in January 2026.
At its core, the Act targets providers of generative AI systems that have more than one million monthly users or visitors in California. These “covered providers” will face significant new obligations designed to increase transparency and accountability in AI-generated content.
Perhaps the most notable requirement is that covered providers must offer a free AI detection tool to the public. This tool must be capable of analyzing images, videos and audio to determine whether they were created or modified by the provider’s AI system. The detection tool needs to be accessible via API and must handle both direct uploads and URL submissions, though providers can implement reasonable security limitations.
The Act also creates a two-tiered system for content disclosures. First, providers must offer users the option to include a “manifest disclosure” – a clear, conspicuous marking that identifies content as AI-generated. Second, and more significantly, providers must implement mandatory “latent disclosures” in all AI-generated content. These latent disclosures must contain key provenance information, including the provider’s name, system details and creation timestamp, and be technically difficult to remove.
The legislation includes important provisions for third-party licensing relationships. Providers who license their generative artificial intelligence systems must ensure their licensees maintain the disclosure capabilities. If a licensee disables these features, providers have just 96 hours to revoke the license once they become aware of the modification. Licensees must immediately cease using the system after revocation.
Enforcement comes with real teeth: violations carry a civil penalty of $5,000 per violation per day. The Attorney General, city attorneys and county counsel all have enforcement authority, and successful actions can include recovery of attorney fees.
Notably, the Act carves out exceptions for non-user-generated content in video games, television, streaming services, movies and interactive experiences. This exclusion helps focus the law’s impact on generative AI tools while avoiding unintended consequences for the entertainment industry.
This legislation represents California’s latest effort to establish guardrails around emerging technologies while promoting transparency and accountability in artificial intelligence. As other states watch this development, California’s approach may well become a model for AI regulation across the country.