The question of how to govern artificial intelligence has moved to the center of American politics, and the answer, for now, is being written not in Washington but in state capitals.

Congress has produced no comprehensive AI law. Lawmakers remain divided over how far regulation should go, and the debate heading into the midterm elections is unsettled, with no clear consensus on a federal framework. Into that opening have stepped governors and candidates for governor, who increasingly argue that states should set their own rules while the national picture remains uncertain.

California has been the most active. Governor Gavin Newsom signed legislation aimed at preparing Hollywood workers for the disruptions AI may bring to their industry, and he advanced an executive order addressing an AI "kill switch" — a mechanism intended to allow powerful systems to be shut down in an emergency. Newsom is among the Democrats often mentioned as potential 2028 presidential candidates, along with Illinois Governor J.B. Pritzker, and both have made AI safety a prominent part of their agendas while federal action lags.

The policy fights are unfolding on several fronts at once. One centers on safety itself: how to prevent advanced systems from causing catastrophic harm. A group of experts published a public letter calling for truly independent bodies to evaluate the safety of companies such as Anthropic and OpenAI, rather than leaving those judgments to the firms building the technology. Some of the AI developers, including Anthropic, OpenAI and Google, face a lawsuit connected to calls to "pace" the speed of AI development.

Another front is physical and financial: the rapid construction of data centers to power AI. The boom has fueled debate over electricity costs, mounting debt, and whether the buildout serves a coherent national strategy. Analysts disagree sharply about the scale and consequences, and some coverage has argued the public conversation has drifted from the underlying facts.

Underneath the policy debate runs a deeper argument about how much danger AI actually poses. Some scientists warn of severe, even existential, risks. Others, including business leaders gathered at Salesforce's Dreamforce conference, said existing models are already capable enough for most commercial needs and questioned the urgency of racing toward more powerful systems. Technology journalist Kara Swisher framed the concern less around the technology destroying humanity and more around the people building it.

The debate is not confined to the United States. In China, officials and workers are grappling with AI's implications for employment and safety, echoing anxieties visible in the West.

Even the courtroom has become a testing ground. In one trial, an expert witness's use of AI drew scrutiny after questions arose about the technology's biases.

Across these fronts, a country is debating rules for a technology that is already reshaping industries while its governing institutions work to respond.

References

1. 1.State policy coverage — Newsom's Hollywood AI legislation and 'kill switch' executive order
2. 2.Political reporting — Newsom and Pritzker as 2028 contenders emphasizing AI safety
3. 3.Expert open letter — call for independent safety evaluations of Anthropic and OpenAI
4. 4.Legal reporting — lawsuit involving Anthropic, OpenAI and Google over 'pace' of AI development
5. 5.Business/tech coverage — data center buildout debate over costs, debt, and strategy
6. 6.Dreamforce coverage — business leaders on current model capabilities; Kara Swisher's framing of AI risk
7. 7.International reporting — AI employment and safety concerns in China
8. 8.Court coverage — scrutiny of an expert witness's AI use and technology bias