California Lawmakers Send 21 AI Bills to the Governor
Key highlights this week:
We’re currently tracking 678 bills in 45 states related to AI this year, 80 of which have been enacted into law.
California lawmakers sent nearly two dozen AI-related bills to the governor for his potential signature. The governor has until the end of the month to decide which will become law.
Task forces and committees in Georgia, Indiana, Texas, and Wisconsin are actively studying AI and discussing potential legislative action on the issue.
California has been an early leader in regulating artificial intelligence. In 2018, the Golden State passed transparency laws for “bot” communications, and in 2019 the state enacted laws to prevent potential bias in AI, limit facial recognition use, and address sexual and political deepfakes. California rulemakers have also been active in regulating the emergent technology. But this year, lawmakers sent an entire menu of AI-related bills to the governor’s desk for his approval. By our count, Gov. Newsom has 21 bills related to AI sitting on his desk. While SB 1047 has garnered the majority of attention, this week we’re providing an overview of each of the bills that made it out of the legislature this year. Gov. Newsom has until Sep. 30 to decide whether they become law or not.
Broad AI Regulation
CA SB 1047: Senator Scott Wiener’s (D) proposed Safe and Secure Innovation for Frontier Artificial Intelligence Models Act has been the center of attention for months and the object of intense lobbying on both sides. The bill has been amended several times since its introduction in February and it is now up to Gov. Newsom (D) whether to sign it into law or not. The governor has been careful not to show his hand, and political betting markets are placing the odds at just above 50% that the governor signs it. As we’ve chronicled in detail, the bill would require developers of frontier AI models to implement safety and security protocols and have the ability to shut down closed models in case of catastrophic safety concerns.
CA AB 2013: This bill would require the developer of a generative AI system or service to publish documentation regarding the data used to train the AI system.
CA AB 2855: This bill would define the term "artificial intelligence” to mean an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments. This definition would then apply to a number of current AI-related laws on the books in the state, including those related to studies on deepfakes and the impact of AI on the workforce, requiring inventories of state government use of AI, directing university research to AI, and required reports for social media platforms.
Deepfakes
Not surprisingly, deepfakes were a major concern for California lawmakers this year, particularly around elections, sexual material, and digital replicas of professionals in the entertainment industry. California already has a law on the books with civil penalties for nonconsensual sexual deepfakes. Political deepfakes are the only AI-related area in which Gov. Newsom has explicitly supported this session, despite the state being one of the first to address the issue in a 2019 law.
Sexual Deepfakes
CA AB 1831: This bill would amend the child pornography laws to include content digitally altered or generated by the use of AI. The law would only take effect if CA SB 1381 also passes and becomes law before the end of the year.
CA SB 896: This bill would make it a crime for a person 18 or older to intentionally distribute or cause to be distributed certain sexual images, including digital and computer-generated images, that cause a reasonable person to believe the image is authentic, cause reasonably foreseeable distress, and actually cause distress. The bill incorporates additional changes to the Penal Code proposed by AB 1874, AB 1962, and SB 1414, to be operative only if this bill and some or all of those other bills are enacted and this bill is enacted last.
CA SB 981: This bill would require a social media platform to provide a mechanism that is reasonably accessible to users to report digital identity theft to the social media platform. The bill would require immediate removal of a reported instance of sexually explicit digital identity theft from being publicly viewable on the platform if there is a reasonable basis to believe it is sexually explicit digital identity theft.
CA SB 1381: This bill would expand child pornography laws to include matter that is digitally altered or generated by the use of AI. Eighteen states have enacted similar laws.
Political Deepfakes
CA AB 2355: This bill would require a committee that creates, originally publishes, or originally distributes a qualified political advertisement to include in the advertisement a specified disclosure that the ad was generated or substantially altered using AI.
CA AB 2655: The Defending Democracy from Deepfake Deception Act would require a large online platform to block the posting of materially deceptive or created content related to an election within 120 days of an election and up to 60 days after an election, and require the online platform to label certain additional content inauthentic, fake, or false during specified periods before and after an election.
CA AB 2839: This bill would prohibit an entity from knowingly distributing an advertisement or other election communication with malice within 120 days of an election and 60 days after an election that contains materially deceptive deepfake content if the content is reasonably likely to harm the reputation or electoral prospects of a candidate. When accompanied by a disclaimer, materially deceptive content can be used for parody or satire or a candidate may use a deepfake of themselves. The bill only becomes law if CA AB 2655 is enacted beforehand.
Transparency
CA SB 942: The California AI Transparency Act would require a provider of a generative AI system (with over 1 million monthly visitors) to make available an AI detection tool, at no cost to the user, and offer the user an option to include a disclosure that identifies content as AI-generated content and is clear, conspicuous, appropriate for the medium of the content, and understandable to a reasonable person.
CA AB 2905: This bill would require a call from an automatic dialing-announcing device to inform the caller if the prerecorded message uses an artificial voice, that is generated or significantly altered using AI.
Digital Replicas
CA AB 1836: This bill would create liability against a person who produces, distributes, or makes available the digital replica of a deceased personality’s voice or likeness in an expressive audiovisual work or sound recording without specified prior consent.
CA AB 2602: This bill would require a contract between an individual and any other person for the performance of personal or professional services is unenforceable only as it relates to a new performance by a digital replica of the individual if the provision meets specified conditions relating to the use of a digital replica of the voice or likeness of an individual in lieu of the work of the individual.
Health
CA AB 3030: This bill would require a health facility, clinic, physician’s office, or office of a group practice that uses generative AI to generate written or verbal patient communications to ensure that those communications include both a disclaimer that indicates to the patient that the communication was generated by AI and clear instructions permitting a patient to communicate with a human health care provider.
CA SB 1120: This bill would require a health or disability insurer that uses an AI, algorithm, and other software tools for utilization review or management decisions to comply with requirements pertaining to the approval, modification, or denial of services, inclusive of federal rules and guidance regarding the use of AI, algorithm, or other software tools. Such software must be equitably and fairly applied across the patient population.
Government Use of AI
CA SB 892: This bill would require the Department of Technology to create an automated decision procurement standard and would prohibit a state agency from entering into a contract for an automated decision tool until the department has adopted regulations creating an automated decision procurement standard that provides a completed risk assessment, requires adherence to appropriate risk controls, and provides procedures for adverse incident monitoring.
CA SB 896: The Generative AI Accountability Act would require a state report examining significant, potentially beneficial uses for the deployment of generative AI tools by the state and a joint risk analysis of potential threats posed by the use of generative AI to California’s critical energy infrastructure. State agencies using generative AI would need to disclose that fact, with clear instructions on how to contact a human employee.
CA SB 1220: This bill would require certain state agencies and local government agencies that provide call center services to provide services performed solely with and by workers employed in California and prohibits using a call center that uses, AI or automated decision systems (ADS) that would eliminate or automate core job functions of a worker. An agency that utilizes AI or ADS that impacts the core job functions of workers would need to notify the workers, their collective bargaining representatives, and the public within a specified timeframe about prescribed information, including a general description of the AI or ADS system.
Education
CA AB 2876: This bill would require the Instructional Quality Commission to consider incorporating AI literacy content into the mathematics, science, and history-social science curriculum frameworks after 2025 and to consider including AI literacy in its criteria for evaluating instructional materials when the state board next adopts mathematics, science, and history-social science instructional materials.
CA SB 1288: This bill would establish a working group related to AI in public schools, to provide guidance for local educational agencies and charter schools on the safe use of AI in education, and to develop a model policy regarding the safe and effective use of AI in ways that benefit, and do not negatively impact, pupils and educators.
Final Thoughts
Notably, two bills that garnered attention this session but didn’t make it out of the legislature and to the governor were Assemblymember Bauer-Kahan’s bill requiring impact assessment for automated decision-making tools (CA AB 2930) and Assemblymember Wicks’s bill (CA AB 3211) to mandate watermarking synthetic content. Overall, California legislators introduced 55 AI-related bills this session, about 30 of which made it out of their legislative chamber of origin. That number has narrowed to the 21 bills sent to the governor, who now has until the end of the month to decide which of those will become law. But don’t expect the Golden State to slow down on AI regulation when the legislature reconvenes in January.
Recent Developments
In the News
Australia: This week, the Industry and Science Minister introduced ten voluntary AI guidelines to begin a discussion on whether there should be mandatory guidelines. The proposed guidelines would set accountability, establish a risk management process, implement data governance to manage quality and provenance, require testing and monitoring models, enable human intervention, inform end-users, set processes to challenge outcomes, require transparency and record-keeping, and set compliance certification.
Major Policy Action
Congress: Last Tuesday, Sen. Mike Rounds (R-SD) introduced five AI-related bills that would promote the use of AI in the financial services and defense spaces and create a national AI literacy strategy. Senate Majority Chuck Schumer (D-NY) has suggested including AI bills in must-pass legislation before the end of the year, perhaps as amendments to the National Defense Authorization Act.
Georgia: The Senate Study Committee on Artificial Intelligence will hold a meeting on September 12 in conjunction with the House Technology and Infrastructure Innovation Subcommittee on AI. This will be the fourth meeting held by the study committee, with future meetings scheduled for October 2, October 23, November 8, and December 4.
Indiana: Gov. Eric Holcomb (R) made four appointments to the Artificial Intelligence Task Force to serve through June 30, 2025. The group held its first meeting on August 21.
Texas: The Senate Business and Commerce Committee held a hearing last Tuesday with a wide range of AI-related topics, including data privacy concerns, algorithmic bias, the threat of deepfakes, the harms posed to artists, and the use of the technology in state government. Rep. Giovanni Capriglione (R) plans to draft a package of bills addressing AI-related concerns for next year’s legislative session.
Wisconsin: The Legislative Council Study Committee on the Regulation of Artificial Intelligence in Wisconsin held its second meeting last Thursday in Green Bay to discuss the use of the technology in health care. The group will meet again in Wausau on September 11 to discuss the need for AI-skilled workers and the benefits and risks of generative artificial intelligence.