Tell Members of Congress: Allow States to Regulate Artificial Intelligence!
Call BOTH of your Senators.
Call ONE of the Representatives. Note: only one of these Congressmembers represents you. Find out which one here.
Choose at least one concern you have about AI to use in your call, in your own words. You can find examples in the Background.
And, talk about it at Thanksgiving! It’s a less partisan issue, so it’s easier to have a productive conversation.
Call Script
My name is __________. I am a constituent, and my zip code is _______. I am a member of Indivisible SF.
I am calling in support of artificial intelligence regulation, including by the states. I am concerned about the dangers of rapidly-advancing artificial intelligence, especially [insert your concern here].
I ask you to oppose the language preventing states from regulating artificial intelligence that Big Tech is trying to sneak into the National Defense Authorization Act. Please speak out against it, pressure the members of the Armed Services Committee, and commit to only vote for a clean version of the NDAA. Please speak out against the Trump Administration's draft executive order to punish states that take the lead in AI legislation.
[Extra Credit: thank Attorney General Bonta for opposing federal pre-emption of AI legislation].
My name is __________. I am a constituent, and my zip code is _______. I am a member of Indivisible SF.
Thank you for your statement in favor of state regulation of artificial intelligence! I am in favor of stronger regulation of this reckless industry that threatens our privacy, and safety.
Background
AI Harms and Risks
Artificial intelligence tools have been getting smarter. Cutting edge chatbots using Large Language Models can pass the Turing Test (i.e., seem human) and the bar exam. “Agentic” AI tools can be given plans to execute without human supervision.
AI tools have already had real-world impacts even in ostensibly “neutral” applications:
Algorithmic Bias. Machine learning models inherit human racism and sexism. These AIs unreliably perform life-changing roles, such a processing job, loan, and parole applications or setting insurance premiums.
The nH Predict algorithm used by UnitedHealth Group helped deny health insurance claims, for which they are being sued.
OpenAI's ChatGPT chatbot encouraged a child to commit suicide, for which the parents are suing.
AI Psychosis/Chatbot Psychosis. An emerging phenomenon where chatbots’ sycophantic conversational style draws users into delusional thinking.
And can be weaponized to a frightening degree:
AI tools can be used to design new chemical weapons or easier recipes for making existing ones
AI tools don’t have any concept of truth, so they can be used to produce false statements, fictitious images, etc.
And the companies building these technologies keep on making them more and more potent, with billions of dollars of current and promised investments, without a single apparent thought for the harms they’re causing.
Existing AI Regulation
It should be common sense that a technology with this potential for harm should be regulated.
In 2024, Governor Newsom vetoed the AI safety law SB-1047 (sponsored by our Senator Scott Weiner), which explicitly made AI companies liable for harms caused by their models, required some minimal self-policing for the most advanced models, and strengthened whistleblower protections.
In 2025, California passed SB-53, the Transparency in Frontier Artificial Intelligence Act, a weaker compromise of SB-1047, endorsed by frontier AI company Anthropic, that "gives Big Tech exactly what it wanted".
And that is too much for congressional Republicans. Just as they want fossil-fuel companies, chemical companies, Big Ag, and other billion-dollar companies to exploit us and pollute our commons with impunity, they want to extend this to AI companies as well.
This Week's US Congressional Call Scripts:
