California's AI politics intensify

California AI politics just got more real. On May 4, Xavier Becerra rolled out his first detailed AI platform in the governor’s race, and it wasn’t a side note — it was an 11-point governing plan. At the same time, Jim Steyer, the Common Sense Media chief and brother of candidate Tom Steyer, is already helping define the state’s hardest-edged AI fights around kids, chatbots, audits, and liability. Put those together and you get the real story: in California, AI is no longer just a tech policy debate. It is becoming campaign politics. (politico.com) ### What did Becerra actually propose? Becerra’s memo tries to split the difference between boosterism and regulation. He wants California to use AI more aggressively in education and government, expand AI literacy through schools, libraries, and community colleges, invest in worker upskilling, and build a state-run cloud computing cluster. He also wants fa(politico.com)— but only if it fits the climate frame. (politico.com) ### Why is that a political move? Because Becerra has been climbing fast, and AI gives him a way to look future-facing without sounding captured by Silicon Valley. The memo frames AI as something that should serve “every Californian,” not just enrich a few companies. That puts him near Tom Steyer on the populist critique, but with a more technocratic tone —(politico.com), that matters. (politico.com) ### Where does Jim Steyer fit in? Jim Steyer matters because he is not just a commentator. He runs Common Sense Media, has deep Sacramento relationships, and has already moved from advocacy into ballot-measure politics on AI. One initiative he filed in late 2025 would restrict certain AI chatbot uses by minors, limit data-sharing involving kids, and create (politico.com) rules. Basically, he is helping turn “protect kids online” into a concrete AI regulatory agenda. (news.ballotpedia.org) ### Why does Tom Steyer’s campaign make that more important? Because Tom Steyer is also running on AI, and his platform overlaps with that same worldview. He says companies profiting most from AI should not be the ones writing the rules. His campaign calls for workplace guardrails, independent safety testing, st(news.ballotpedia.org)family’s policy lane is obvious — child safety, audits, accountability, and redistribution of AI gains. (tomsteyer.com) ### What changed in California before this? The backdrop is that California already spent 2025 trying to figure out what frontier-AI governance should even look like. Newsom’s working group published a major report last June with researchers from Stanford, Berkeley, Harvard, and elsewhere. That did the vocabulary-building part — model risks, governance options, state capacity. What is changing now is who owns the (tomsteyer.com)orts into campaign platforms and ballot threats. (gov.ca.gov) ### What does that mean for startups? It means the compliance target may start moving faster. If California’s AI debate gets driven by gubernatorial politics and kid-safety ballot measures, startups should expect more pressure around audit trails, disclosures, model testing, and proving they considered foreseeable harms before launch. The catch is that even companies far from “frontier AI” can get pulled in if they touch education, minors, work, or public services. (politico.com) ### So what is the bottom line? California is starting to treat AI the way it treats energy, housing, and health care — as a governing issue with factions, constituencies, and election incentives. Once that happens, the center of gravity shifts. The question stops being whether the state will regulate AI, and becomes who gets to define what “safe” and “fair” mean first. (politico.com)