Governor Gavin Newsom of California signed over 20 AI-related bills into law in the final days of September. Several of these bills address topics such as frontier model and generative AI oversight, AI transparency, and new requirements for use of AI in particular industries, described below. In addition to these bills, Governor Newsom also signed into law a sweeping set of AI bills relating to minors’ safety and privacy, which are described in our prior blog post, and novel frameworks related to the use of ADMT in the workplace, which we previously summarized here.

These newly passed laws include the following:

• AI Transparency: Governor Newsom signed several laws focused on AI transparency, including amendments to the California AI Transparency Act and disclosure requirements for AI-generated advertisements.
  • California AI Transparency Act: Governor Newsom signed two bills amending the California AI Transparency Act — SB 1000 and AB 2713. Among other things, SB 1000 expands the scope of “covered providers” subject to the law by removing the requirement that the generative AI system has more than 1 million monthly visitors or users, updates requirements for the disclosure verification tools that must be made available to users, and imposes new requirements for latent disclosures. Additionally, SB 1000 imposes requirements where a covered provider knows that an identifiable third-party licensee modified a licensed generative AI system such that it no longer complies with the law. In contrast, AB 2713 amends requirements placed on “large online platforms,” including by requiring large online platforms to detect provenance data, rather than (under the prior law) requiring that the large online platform detect whether the provenance data is compliant with widely adopted specifications, among other requirements.
  • Synthetic Performers: Similar to the New York Synthetic Performer AI Disclosure Law,  SB 1050 will require a clear and conspicuous disclosure when an advertisement “prominently” includes a synthetic performer, which is defined as “a digital figure, voice, or representation created in whole or in part using generative artificial intelligence that creates the realistic impression of the audio, audiovisual, or visual performance of a human performer who is not recognizable as any identifiable natural person.” The disclosure must use wording “substantially similar” to “this performance features a synthetic performer” or “no human performer is depicted.”

• Customer Service Chatbots: In addition to the restrictions on operators of a companion chatbot in Adam’s Law (SB 1119), AB 1609 will prohibit large private businesses from representing that a customer chatbot is a human and will require these businesses to provide a clear and conspicuous disclosure that the customer service chatbot is AI-generated and is not human, if a reasonable person likely would be misled into believing that they are interacting with a human. In addition, large private businesses will be required to make a good faith effort to connect a customer to an agent within 15 minutes after a request is made, or schedule an appointment with the customer, among other requirements.

• Focus on Specific Industries: Several provisions will impose requirements on development and use of AI in certain industries or contexts.  For example:
  • Generative AI in the Legal Profession: SB 574 will prohibit an attorney from delegating the practice of law to generative AI, and it will impose specific requirements on the use of generative AI to assist with the practice of law, including taking reasonable steps to verify the accuracy of generative AI outputs and correcting any erroneous or hallucinated outputs, among other things.

• Healthcare: Governor Newsom signed several measures focused on healthcare, such as AB 1979, which will prohibit healthcare providers or facilities from using or deploying AI for certain purposes including direct performance of certain clinical functions. As another example, SB 903 will permit entities that provide or facilitate psychotherapy services to use AI tools or systems only to assist in providing administrative or supplementary support, subject to additional requirements, and it will prohibit advertising or purporting to offer psychotherapy services when the services are provided through the use of companion chatbots.