Here’s what’s illegal under California’s 18 new AI laws (and counting).

In September, California Governor Gavin Newsom considered 38 bills related to AI, including the highly controversial SB 1047, which the state legislature sent to his desk for final approval. He voted for SB 1047 on Sunday, marking the end of the road for California’s controversial AI bill that tried to prevent AI disasters, but he signed more than a dozen other AI bills into law this month. These bills try to address the most pressing problems in artificial intelligence: everything from Al risk, to deep nudity created by AI image generators, to Hollywood studios creating AI clones of dead actors.

“Home to many of the world’s leading AI companies, California is working to harness this transformative technology to help address pressing challenges while studying the risks involved,” Governor Newsom’s office said in a press release.

To date, Governor Newsom has signed 18 AI bills into law, some of which are America’s most far-reaching generative AI legislation yet. Here’s what they do.

The danger of AI

On Sunday, Governor Newsom signed SB 896 into law, requiring the California Office of Emergency Services to conduct a risk analysis of potential threats posed by artificial intelligence. CalOES will work with frontier modeling companies, such as OpenAI and Anthropic, to analyze potential AI threats to critical national infrastructure, as well as threats that could lead to mass casualty events.

Training data

Another law Newsom signed this month requires artificial intelligence providers to disclose the data used to train their AI systems in documents published on their websites. AB 2013 comes into effect in 2026, and requires AI providers to publish: the sources of their data sets, a description of how the data is used, the number of data points in the set, whether copyrighted or licensed data is included, the time the data was collected, among others standards.

Privacy and AI systems

Newsom also signed AB 1008 on Sunday, which specifies that California’s existing privacy laws are extended to artificial intelligence systems. That means that if an AI system, like ChatGPT, reveals someone’s personal information (name, address, biometric data), California’s existing privacy laws will limit how businesses can use and profit from that data.

Education

Newsom signed AB 2876 this month, which requires the California State Board of Education to consider “AI literacy” in its math, science, and history curriculum and teaching materials. This means California schools may begin teaching students the basics of how artificial intelligence works, as well as the limitations, implications, and ethical considerations of using the technology.

Another new law, SB 1288, requires California superintendents to form task forces to examine how AI is used in public school education.

It defines AI

This month, Newsom signed a bill that establishes a uniform definition of artificial intelligence in California law. AB 2885 states that artificial intelligence is defined as “an engineering or machine-based system that varies in its level of autonomy and can, for explicit or implicit purposes, derive from its inputs how to produce outputs that can have a tangible or observable effect. nature.”

Health care

Another bill signed in September is AB 3030, which requires healthcare providers to disclose when they use artificial intelligence to communicate with a patient, especially if those messages contain the patient’s clinical information.

Meanwhile, Newsom recently signed SB 1120, which sets limits on how health care providers and health insurers can automate their services. The law ensures that licensed physicians supervise the use of AI tools in these settings.

AI robocalls

Last Friday, Governor Newsom signed a bill into law requiring robocalls to disclose whether they used AI-generated voices. AB 2905 aims to prevent another instance of a deep-pocketed robocall similar to the Joe Biden voiceover that confused many New Hampshire voters earlier this year.

Deep porn

On Sunday, Newsom signed AB 1831 into law, which expands the scope of existing child pornography laws to include material generated by AI systems.

Newsom signed two pieces of legislation dealing with the creation and spread of extreme nudity last week. SB 926 criminalizes this act, making it illegal to report AI-generated nude images to someone like them.

SB 981, which also became law on Thursday, requires social media platforms to establish channels for users to report extreme nudity like them. The content should then be temporarily blocked while the forum investigates it, and permanently removed if confirmed.

Watermarks

Also on Thursday, Newsom signed a bill to help the public identify AI-generated content. SB 942 requires widely used AI-generated systems to disclose that they are AI-generated in their content origination data. For example, all images created by OpenAI’s Dall-E now require a small tag in their metadata that says they were created by AI.

Many AI companies are already doing this, and there are several free tools out there that can help people read this raw data and visualize AI-generated content.

Election deepfakes

Earlier this week, California’s governor signed three laws cracking down on AI deepfakes that could influence elections.

One of California’s new laws, AB 2655, requires major Internet platforms, such as Facebook and X, to remove or label deep AIs related to elections, and to create channels to report such content. Candidates and elected officials can seek emergency relief if a large online platform does not support action.

Another law, AB 2839, targets social media users who post, or repost, AI deepfakes that could potentially deceive voters in upcoming elections. The law went into effect immediately on Tuesday, and Newsom suggested that Elon Musk could be at risk of violating it.

AI-generated political ads now require direct disclosure under a new California law, AB 2355. That means going forward, Trump may not be able to get away with posting AI deepfakes of Taylor Swift endorsing her on Truth Social (she endorsed Kamala Harris). The FCC has proposed a similar disclosure requirement at the national level and has already made robocalls using AI-generated voices illegal.

Players and AI

Two laws Newsom signed earlier this month — which SAG-AFTRA, the nation’s largest actors’ union, was pushing for — created new standards for California’s media industry. AB 2602 requires studios to get permission from an actor before creating an AI-generated voice or likeness.

Meanwhile, AB 1836 prohibits studios from creating digital likenesses of deceased actors without permission from their estates (eg, legally cleared likenesses used in the movies “Alien” and “Star Wars”, as well as other films).

SB 1047 gets a veto

Governor Newsom still has several AI-related bills to decide on before the end of September. However, SB 1047 is not one of them – the bill was closed on Sunday.

In a letter explaining his decision, Newsom said SB 1047 focuses too much on large-scale AI programs that would give the public a false sense of security. California’s governor noted how dangerous small AI models like those targeted by SB 1047 can be, saying a more flexible regulatory approach is needed.

While speaking with Salesforce CEO Marc Benioff earlier this month during the Dreamforce 2024 conference, Newsom may have tipped his hat to SB 1047, and how he thinks about regulating the AI ​​industry more broadly.

“There is one bill that is like a big one in terms of public discourse and awareness; this is SB 1047,” Newsom said on stage this month. “What are the perceived risks of AI and what are the perceived risks? I can’t solve everything. What can we solve? So that’s how we’re taking the whole spectrum on this one. “

Check back to this article for updates on the AI ​​laws the California governor signs, and what he doesn’t.