So here's California, doing what California does: stepping into a national policy debate with a big, splashy move of its own. On Monday, Governor Gavin Newsom decided it was time to put some guardrails on the AI wild west, at least for anyone who wants to do business with the state.
He signed an executive order that basically tells AI companies: if you want a contract with California, you're going to have to play by some new rules. The goal is to curb misuse—think things like creating illegal content, baking harmful bias into systems, or violating civil rights. It's a classic use-the-state's-purchasing-power move. California is a massive customer, so its rules tend to matter.
Watermarks, Please: A New Rule for Deepfakes
One of the more concrete parts of the order is about transparency. State agencies are now directed to clearly label AI-generated images and videos with watermarks. It's a direct shot at the growing problem of deepfakes and AI-powered misinformation. The idea is simple: if the state is putting something out, you should know if a machine helped make it.
California Charts Its Own Course Amid Federal Turmoil
This whole thing is happening against a pretty messy federal backdrop. Just recently, the Pentagon slapped a "supply-chain risk" label on Anthropic, the company behind the Claude AI. That label effectively bars contractors from using Anthropic's tech in military work.
Anthropic, which is backed by Amazon.com, Inc. (AMZN) and Alphabet Inc.'s (GOOG) Google, didn't take that lying down. Last week, they got a preliminary injunction from U.S. District Judge Rita Lin. The court said the government's classification was likely unlawful and might have been retaliatory, putting a temporary pause on a directive linked to the previous administration.
California, however, is signaling it might not just fall in line with that court decision. The state is taking an independent stance, suggesting its own risk assessments and rules might differ from what's happening in Washington.
Building a New Rulebook for AI Vendors
The order isn't just a one-off decree. It kicks off a process. Within the next 120 days, the state's Department of General Services and Department of Technology have to come up with a proposal for a certification framework. This framework would require AI vendors to formally attest that they have responsible governance and public safety protections in place.
It's part of a broader push to create formal, standardized accountability for the AI industry when it interacts with government—a way to move from vague promises to something a bit more concrete.
The Politics of AI: Newsom Draws a Line
Perhaps the most pointed part of this came from Newsom himself. In a statement, he didn't just talk about policy; he framed it as a philosophical battle.
"I just signed an executive order to actively make sure AI companies working with the state protect privacy and civil liberties," he said.
Then he drew the contrast: "While Trump pressures companies to deploy AI for autonomous weapons and domestic surveillance, California is using our power … to raise the bar on privacy and security."
It's a clear political statement, positioning California's approach—focused on consumer protection and civil liberties—in direct opposition to a vision of AI centered on national security and surveillance. In the evolving debate over how to govern this powerful technology, California has just planted its flag firmly on one side of the field.
