Trump signed an Executive Order (Dec 2025) preempting state AI safety laws, threatening to withhold $42.5B in broadband funding from states that refuse to comply (specifically targeting Colorado and California).
In response, New York signed the "RAISE Act" after the EO was issued. It has strict safety, transparency, and reporting protocols for frontier models.
California is enforcing its "Transparency in Frontier AI Act" (Sept 2025) regardless of the Federal threat. It requires developers of large AI models (over 10^26 FLOPS) to publicly disclose safety frameworks, report "catastrophic risk" incidents, protect whistleblowers.. etc
Big Tech (OpenAI, Google, Andreessen Horowitz) is siding with Trump on this one. They prefer one weak federal law to 50 strict state laws.
This post:
Red states are creating deregulation areas. If a big tech company has data centers in Montana, and CA tries to impose an audit on their model, the company can sue, claiming their "Civil Rights" in Montana are being infringed by California's overreach.
Red states are tying "Compute" to the First Amendment (free expression), basically anticipating the Supreme Court.
Future implications:
The US continues to split into two distinct operating environments. https://www.economist.com/interactive/briefing/2022/09/03/am...
Seems pretty vague to me, but IANAL.
"So, hypothetically, in a state with a right-to-compute law on the books, any bill put forward to limit AI or computation, even to prevent harm, could be halted while the courts worked it out. That could include laws limiting data centers as well.
“The government has to prove regulation is absolutely necessary and there’s no less restrictive way to do it,” Wilcox said. “Most oversight can’t clear that bar. That’s the point. Pre-deployment safety testing? Algorithmic bias audits? Transparency requirements? All would face legal challenge. "
My take: This sounds incredibly pro-industry and anti-democratic.
While I appreciate bringing attention to ongoing changes in the tech/legal landscape, I'll get my rundowns from a source that doesn't blindly repeat this broken assertion. Doesn't speak well of their research practices.
Hasn't stopped every authoritarian from parroting the quote whenever they want to censor something.
Is it though? If you're putting on a play, and there is a fire in the script, e.g. in a play criticizing that decision, can the government punish you for putting on the play because of the risk it could cause a panic? If there is actually a fire in the theater, can they punish you for telling people? What if there isn't actually a fire but you believe that there is?
Not only is it useless as an analogy for doing any reasoning, the thing itself is so overbroad that even the unqualified literal interpretation is more of a prohibition than would actually be permissible.