California’s Groundbreaking AI Transparency Law: What You Need to Know

On September 19, 2024, California Governor Gavin Newsom signed the California AI Transparency Act, a comprehensive law that will require AI companies to be transparent about their technology and its outputs. Set to take effect on January 1, 2026, this law is designed to bring clarity to the growing world of generative AI, especially for companies with more than 1 million monthly users.

1. Free AI Detection Tools: Companies must provide users with tools to detect whether content—whether it’s an image, video, or audio—has been generated or altered by AI, along with detailed information about the content's origin.

2. Mandatory Disclosures:

• Manifest Disclosures: Clear, visible indicators for users that content was AI-generated.
• Latent Disclosures: Hidden in the metadata, providing crucial details like the AI system used, version, and creation date. These must be tough to remove and detectable by industry-standard tools.

3. Contractual Safeguards: AI developers must enforce compliance with these transparency measures even when licensing their technology. If a third-party fails to uphold these standards, the developer must terminate the agreement within 96 hours.

Unlike laws in other states, California’s act is the first to establish specific watermarking rules for AI content, pushing for greater accountability in the tech space. Violations come with hefty fines, ensuring companies take the new regulations seriously.

This law sets a new standard for transparency in AI, aiming to build trust as generative AI continues to grow in influence.

Source: Mayer Brown
Authors: Arsen Kourinia, Howard W. Waltzmn and Mikey Liebner
Edited by: Legal Data AI

Let us know about this, we will study the possibility of preparing a digest on your topic. In any case we will respond to your request in a letter if you left your email.