CA’s New AI Laws

The governor of CA recently announced new initiatives to advance safe and responsible AI, with the objective of protecting Californians, through signing 17 bills covering regulation and deployment of generative AI, as well as vetoing SB 1047.

In the next couple posts in my AI series, I go through the key bills that were signed that will have the most impact on tech companies and implications for them going forward: AB 2013 and SB 942. I’ll also briefly share thoughts on Governor Newsom’s veto of SB 1047. 

AB 2013: AI Training Data Transparency

• Objective & Summary: In an effort to create greater transparency around training data, developers of gen AI models must post on their websites certain disclosures about datasets used to train their models. These disclosure must be publicly available.
• When it goes into effect: Jan. 1, 2026, and afterward every time a gen AI system is modified or created, and released for public use.
• Who it affects: “Developers” of gen AI systems made available to Californians to use, on or after Jan. 1, 2022. This is broad coverage, and covering models released over 2 years ago.
• Disclosure requirements: a high level summary of datasets used to train these systems, including the below info
  • Sources or owners of the datasets
  • Summary of how the datasets further the intended purpose of the gen AI system
  • General range of the number of data points used in the datasets
  • A description of the types of data points within the datasets.
  • Indication of whether the datasets include any data protected under IP laws (copyright, patent, trademark) or are in the public domain
  • Whether the datasets were purchased or licensed
  • Whether the data sets include personal information (names, addresses, etc.)
  • Whether the data sets include aggregate consumer information
  • Whether there was any processing or modification by the developer
  • The time period during which the data was collected, as well as if data collection is ongoing
  • The dates the datasets were first used during development of the gen AI system 
  • Whether synthetic data is used in the development of the gen AI system

Gen AI developers do NOT have to make these disclosures if their gen AI systems’ sole purpose is to help ensure security and integrity, operation of aircraft in national airspace, or were developed for national security, military or defense purposes and made available only to a federal entity.

‍