California Sets National Standard With 18 New AI Laws

California Governor Gavin Newsom has recently signed 18 artificial intelligence (AI) laws, signaling a significant push to address the emerging risks and ethical challenges posed by AI technologies. The new laws span a wide range of concerns, including deepfake images, privacy violations, healthcare automation, and educational reforms. Although Newsom vetoed SB 1047, a controversial AI disaster prevention bill, he has made headway with other AI-focused regulations, positioning California as a leader in the national debate on AI governance.

Regulating AI Risk and Infrastructure Threats

One of the cornerstone laws passed this month, SB 896, mandates that California’s Office of Emergency Services (CalOES) conduct risk assessments related to generative AI. These assessments will evaluate the potential threats AI poses to critical state infrastructure and its capacity to lead to mass casualty events. CalOES will collaborate with major AI firms like OpenAI and Anthropic to study how AI could impact public safety, making this bill one of the more forward-looking approaches in California’s legislative package.

Transparency in AI Training Data

Another law, AB 2013, set to go into effect in 2026, will compel AI companies to reveal the data used to train their AI models. This law emphasizes transparency by requiring providers to publish detailed reports on their datasets, including the sources, volume, and use of the data, as well as any involvement of copyrighted material. This initiative is aimed at ensuring public oversight over how AI systems are trained, particularly in terms of ethical data use.

Privacy Protections for AI Systems

In a move that extends existing privacy regulations, AB 1008 clarifies that California’s robust privacy laws also apply to generative AI systems. If an AI tool like ChatGPT mishandles personal data—such as names, addresses, or biometric information—businesses will face limitations on how they can profit from that data. This law expands consumer protection in the age of AI, focusing on the critical area of data privacy.

AI in Education

Governor Newsom also signed AB 2876, which introduces “AI literacy” to California’s State Board of Education curriculum frameworks. This new requirement ensures that students will be educated on the principles of AI, including its limitations, ethical concerns, and broader impacts on society. Additionally, SB 1288 creates a working group tasked with exploring how AI is already being integrated into public school education, a proactive measure aimed at understanding and shaping the role AI plays in the classroom.

Defining AI in California

One of the foundational laws passed this month, AB 2885, sets a clear and uniform definition of artificial intelligence for the state. This law defines AI as a machine-based system capable of using input to generate outputs that influence physical or virtual environments. Establishing a legal definition of AI will provide a consistent framework for future regulations and legal cases involving AI technologies.

Healthcare and AI

The healthcare sector is also under new scrutiny with AB 3030, which mandates that healthcare providers disclose when generative AI is used in communications with patients, particularly when discussing clinical information. Additionally, SB 1120 imposes restrictions on the use of AI in healthcare settings, ensuring that licensed physicians supervise AI tools. These new regulations seek to safeguard the doctor-patient relationship and ensure that AI does not replace essential human oversight.

Tackling AI-Generated Robocalls and Deepfakes

In response to growing concerns about AI-generated content in election campaigns, AB 2905 was signed to regulate robocalls using AI-generated voices. This law requires that any robocall generated by AI must clearly disclose that fact, preventing confusion similar to the deepfake robocalls that impersonated public figures like Joe Biden earlier this year.

AI-generated deepfakes, particularly in the realm of pornography, are also addressed under AB 1831 and SB 926. These laws make it illegal to blackmail individuals using AI-generated nude images and establish procedures for social media platforms to quickly block and remove such content. Platforms are required to set up reporting channels and investigate AI-generated deepfake nudes, ensuring that victims can take swift action to protect themselves.

AI in Political Campaigns

Three additional laws passed this month aim to prevent the use of AI-generated deepfakes in political campaigns. AB 2655 requires social media platforms like Facebook to remove or label AI-generated deepfakes related to elections. AB 2839 targets users who post AI-generated deepfakes intended to deceive voters, and AB 2355 mandates that AI-generated political ads include clear disclosures. These laws aim to preserve the integrity of elections by ensuring that AI-generated content cannot be used to mislead voters.

The Veto of SB 1047

While Governor Newsom has signed a variety of AI-related bills, he vetoed SB 1047, which sought to focus narrowly on large AI systems that posed the greatest risks. Newsom’s reasoning was that this bill could give a false sense of security by ignoring smaller AI systems, which could be just as dangerous. He argued for a more flexible and comprehensive approach to AI regulation, hinting at future legislative efforts to tackle AI challenges more broadly.

These 18 laws signal California’s proactive stance in regulating AI, addressing a variety of issues from consumer privacy to election integrity. By establishing some of the most comprehensive AI laws in the United States, the state aims to balance the promise of AI technologies with the potential risks they pose to society.