A little over six weeks ago, Governor Gavin Newsom signed Executive Order N-6-26 in Sacramento. For a moment, people in the tech world took notice. Not in a dramatic way where the gavel falls on the marble. It was more like a quiet acceptance that someone in power had finally said out loud what many workers and economists had been thinking behind closed doors: AI is not a faraway threat. California isn’t ready for what’s going to happen now.
There are no new rules for private employers in the order itself, which was signed on May 21, 2026. That point needs to be made again because it gets lost in the news. On May 22, nothing changed in the human resources department of a medium-sized logistics company in Fresno or in the customer service departments of a financial services company in downtown Los Angeles. The official position of the state changed, which shows that Sacramento wants to get ahead of this, or at least try to.
What the order really does is set off a bunch of deadlines. The Labor and Workforce Development Agency has 90 days to do research on which industries and groups of people are most likely to be displaced. Within 180 days, or until around mid-November, which is after Election Day, the same agency is supposed to suggest changes to the California WARN Act and look into what a safety net for workers who lose their jobs because of AI could look like. Some examples of this are standards for severance pay, models for sharing profits, and more unemployment insurance. It’s possible that none of it really leads anywhere. Also, it might lead to something important.
As I watch this from afar, I get the impression that the order was made with two groups of people in mind. One group is workers, especially women, who are called out by the executive order as being more likely to be hurt by automation. The second group is the business community, which is being told in measured Sacramento language that the reckoning is coming but will be looked into first before it is made law. There are already 33 of the top 50 private AI companies in the world based in California. It’s clear that Newsom doesn’t want to scare them. Also, it’s clear that he doesn’t want to be the governor who does nothing while workers are laid off.
Most likely, the WARN Act review is the part that most employers should be paying attention to. Large companies already have to give notice before laying off a lot of workers in California because of the way the law is written. If the LWDA suggests that the framework be expanded to include AI-driven displacement, and the Legislature does so, it could mean big changes for employers when it comes to compliance. SB 951 is already trying to do the same thing and is still in the Assembly. The review window in the executive order might make that bill move more slowly. After all, there’s no reason to rush legislation before the state’s own study is finished.
Another thread that runs through all of this doesn’t get enough attention. During the same time period, California quietly made a deal with Anthropic to offer the Claude AI assistant to state agencies at a 50% discount. As a bonus, the deal included free training for state employees to use the AI. It’s not hard to see what the arrangement means: the government will use these tools and wants its employees to know how to use them too. That doesn’t go against the executive order. That’s almost the point.
After six weeks, the most honest assessment is that this is still mostly a plan for a conversation that hasn’t begun yet. The research window of 90 days ends at the end of August. In November, the 180-day window will be over. Either what comes out of those reviews will be used to make real policy, or it will just be another report that is filed and forgotten. Still, it’s not clear which way this goes. But there is order, the clocks are ticking, and California is often a reminder that things don’t end there. It’s where most of them start.
