California has implemented a significant new law, Senate Bill 942, also known as the California AI Transparency Act, as of August 1, 2026. Championed by State Senator Josh Becker and signed by Governor Gavin Newsom in 2024, this legislation is designed to enhance user awareness and identification of content generated or altered by artificial intelligence. Its core objective is to counter the proliferation of deepfakes, online scams, and sophisticated disinformation campaigns by obligating major technology firms that develop generative AI models to provide accessible tools for verifying the authenticity and origin of images, videos, and audio shared online. The law applies to "covered providers," defined as companies with generative AI systems attracting over one million monthly users in California.
What changes from now on
The California AI Transparency Act (SB 942) is being rolled out in two distinct phases. The initial phase, which commenced on August 1, 2026, mandates that all "covered companies" must supply a free-to-use tool. This tool enables individuals to ascertain whether any given image, video, or audio file, or content accessed via an internet link, was produced or manipulated using their AI systems. A crucial aspect of this phase is the stipulation that these verification tools are prohibited from retaining submitted user files beyond the immediate analysis period and are forbidden from collecting personal user information. The second phase, set to begin on January 1, 2027, will require platforms to prominently display visible labels on all AI-generated content, thereby facilitating clear differentiation from authentic, human-created material for all users.
The importance of "digital provenance"
A cornerstone of this new legislation is the concept of "digital provenance data." This refers to a comprehensive set of embedded information or metadata within content, detailing its creation history. This includes identifying the original creator, the specific artificial intelligence system employed, the precise date of generation, and any subsequent modifications made to the content. Essentially, digital provenance functions as a verifiable digital record, crucial for confirming the authenticity of content, particularly in high-stakes contexts such as electoral processes, vital public health campaigns, or any information that could pose a risk to user security. The law further mandates that, where technically feasible, AI-generated content must explicitly incorporate details such as the provider's name, the version of the AI system used, the creation date, and a unique identifier to ensure its traceable origin.
It also applies to companies that license their technology
The scope of SB 942 extends beyond just the developers of AI models to include companies that license these systems. These licensing companies are now obligated to guarantee that the tools they provide continue to feature the identification mechanisms prescribed by the law. Should a licensing company discover that a third party has altered their AI system in such a way that it no longer complies with the mandatory tagging or provenance information requirements, they are compelled to revoke the license for that system within a strict timeframe of 96 hours. This provision aims to maintain a consistent standard of transparency across the AI ecosystem, regardless of how the technology is distributed or utilized.
Fines of up to USD 5,000 per violation
To ensure compliance, the legislation outlines a robust penalty framework for infractions. Authorities are empowered to levy civil fines of up to $5,000 for each individual violation. Critically, the law specifies that every day an act of non-compliance continues will be treated as a separate and independent violation. This cumulative penalty structure has the potential to dramatically increase the total amount of fines for prolonged non-adherence. Enforcement actions can be initiated by various legal entities, including the Attorney General of California, city attorneys, and legal counsels in different counties throughout the state. Through this assertive regulatory stance, California reinforces its role as a pioneering state in the field of artificial intelligence governance, prioritizing user transparency and actively working to diminish the harmful effects of content manipulated by generative AI technologies.
