When Governor Newsom introduced his Executive Order on generative AI last summer, he highlighted California’s leadership in all things artificial intelligence. The Order noted that the Golden State is home to 35 of the world’s 50 largest AI companies, while San Francisco and San José produce a quarter of all AI patents, conference papers, and global AI companies.
The California legislature has taken this leadership seriously, introducing a myriad of AI bills in the first weeks of the legislative year.
Here’s a quick recap of some of the most prominent legislation:
AB 331 (Bauer-Kahan) Automated decision tools. This bill – held over from 2023 – would require developers and deployers of automated decision tools (ADTs) that use AI to complete an annual assessment for the Civil Rights Department to describe the purpose, uses, and context of the technology in making “consequential decisions” impacting natural persons. These decisions are those impacting the matters of employment, education, finance, legal, healthcare, and others. Further, no such tools would be allowed where their use results in algorithmic discrimination. Natural persons impacted by the tools could request to opt out of decisions made solely by ADTs.
AB 1791 (Weber) Artificial intelligence: technical open standards and content credentials. In this bill, California-based companies in generative artificial intelligence would be required to implement the Coalition for Content Provenance and Authenticity’s technical open standards and content credentials into their tools and platforms. The intent is to provide users with greater transparency into the origin of the digital content they encounter, by using a widely adopted industry standard as the threshold for clarity.
AB 1824 (Valencia) Artificial intelligence: disclosure. Existing law requires the Department of Technology to conduct an inventory of all high-risk automated decision systems that are presently in use or contemplated for use by state agencies. This bill would state the intent of the Legislature to enact legislation that would create a disclosure requirement for content generated through artificial intelligence.
SB 896 (Dodd) Artificial Intelligence Accountability Act. This Senate bill seeks to accomplish several objectives regarding AI. First, it would require a state agency report detailing the risks and benefits of GenAI utilized by the state. Next, the Director of Emergency Services would partner with other agencies to analyze potential threats to critical infrastructure posed by GenAI. Finally, any state agency considering the use of automated decision systems would first need to complete a risk evaluation before its adoption.
SB 942 (Becker) Consumer protection: generative artificial intelligence. SB 942 would state the intent of the Legislature to enact legislation that would establish a mechanism to allow consumers to easily determine whether images, audio, video, or text were created by generative artificial intelligence.
The list of bills here is not comprehensive of possible AI legislation, and legislators have until February 16 to introduce additional bills that may regulate this technology. Nonetheless, the early breadth and scope of bills – ranging from consumer protection to infrastructure security to risk analysis – all seem to orbit the same hub in legislator’s minds: disclosure.
Stakeholders and the public have a role to play in helping elected officials navigate the brave new world of balancing disclosure with innovation. The Institute of California AI Policy (I-CAP) – housed within the Silicon Valley Leadership Group – is uniquely positioned to assist. It consists of industry experts, nonprofit leaders, and research pioneers who combine their knowledge to identify and promote policy solutions for developing responsible AI systems. Historically, California policy sets the trajectory for rules in other states and at the federal level. Getting AI “right” will require thoughtful input from all professional sectors, and I-CAP is ready to do its part.
