California Is Starting to Build Civic Infrastructure for AI

Most AI policy debates still sound like a choice between acceleration and regulation. Go faster, or slow down. Innovate, or restrain.

California may be pointing toward a better frame.

Over the past several months, the state has begun assembling three pieces of what looks like early civic infrastructure for AI. None of it is complete. None of it proves success. But taken together, the pattern is worth watching.

First, California has been building rules for responsible public-sector AI use. In March, Governor Gavin Newsom signed an executive order aimed at AI procurement and state contracting. The order directs the state to develop new vendor-certification standards around public safety, harmful bias, civil rights, civil liberties, and watermarking. In plain English: if companies want to sell AI systems to California, the state wants to ask harder questions before buying.

Second, California has launched a statewide public-engagement process through Engaged California, asking residents how AI is affecting work and the economy. The state says participants can share their experiences with AI at work, their concerns about the economy, and ideas for government action. That matters because civic infrastructure is not just policy. It is the machinery that helps people understand public problems, deliberate across differences, coordinate action, and hold decision-makers accountable.

Third, California has now moved toward worker and business transition planning. Newsom’s May 21 executive order directs state agencies to study AI’s labor-market effects, examine support for displaced workers, and launch a dashboard showing AI’s employment impacts across sectors. Reporting on the order notes that the state is looking at tools ranging from temporary cash assistance and work-sharing to stock-compensation and worker-support policies.

This is still scaffolding, not a finished building.

But the shape of the response is important. Civic infrastructure for AI likely needs many functions at once: sensemaking, translation, deliberation, coordination, legitimacy-building, feedback, accountability, and diffusion monitoring. California’s emerging approach touches several of those at the same time.

That is why this is a watch signal.

The state appears to be doing more than passing a single AI law or issuing a one-off task force report. It is beginning to connect public-sector procurement, public deliberation, and labor-market transition planning. That combination is closer to an operating system than a press release.