No Guardrails: Why America's AI Wild West Should Alarm Anyone Who Works for a Living

What Happened

On Tuesday, Rep. Maxine Waters confronted Treasury Secretary Scott Bessent during a House Financial Services Committee hearing over the Trump administration's approach to artificial intelligence. According to The Hill, Waters criticized President Trump for "blindly" advancing AI development and data center construction "with absolutely no safeguards." The exchange highlights a fundamental disagreement over whether AI should race forward unfettered or develop under some form of regulatory guardrail.

This isn't a small technical dispute. It's about whether America will shape AI's impact on workers, energy use, and economic stability, or whether we'll let it happen to us.

What It Means for You

If you work for a living, this matters immediately. Thirty percent of hours worked in the American economy could be automated by 2030, according to research cited in the Common Good Party's position on the future of work. That's not a distant forecast. That's four years away. And here's the catch: only 24% of workers displaced by automation receive any retraining at all.

Compare that to Denmark, which spends 2% of its GDP on active labor market policies, job retraining, wage insurance, direct support for displaced workers. The U.S. spends 0.1%. We've built almost no safety net for what's coming.

Then there's the energy question. Every large AI model requires massive data centers that consume enormous amounts of electricity and water. In a country where the Colorado River sits at 25% of historical flow and 2 million Americans lack clean running water, deploying AI infrastructure without environmental safeguards isn't just sloppy policy, it's a choice to sacrifice communities' basic needs for corporate speed.

The Bigger Picture

The U.S. currently has no federal AI law. None. Meanwhile, six companies control more than 90% of frontier AI development, and 63% of Americans say AI needs more regulation. The European Union already passed its AI Act in 2024. We haven't started.

This isn't a partisan issue, though the framing often becomes one. The real question is: who decides how AI gets built, companies optimizing for speed and profit, or democratic processes that ask what's good for the majority of people who'll live with the consequences?

The administration's stance appears to prioritize velocity. Fewer rules mean faster deployment, faster returns, faster dominance in a technology race against China. That's the logic. But it comes at a cost most people won't pay. Workers in routine jobs pay. Communities with fragile water systems pay. The unretrained pay.

The timing is particularly sharp because data centers themselves are infrastructure projects with real footprints. They demand reliable electrical grids, massive water supplies, and often land in rural areas already struggling economically. A truly common-good approach would ask: where do these go, who benefits locally, and what happens to water and power in that region? Right now, the answer is: whatever the company wants.

Where This Goes

Waters' challenge to Bessent signals that this won't be a quiet debate confined to tech policy conferences. Congressional oversight is beginning to push back on the hands-off approach. But pressure from one committee hearing rarely stops a determined administration.

The real test is whether Congress acts before AI deployment becomes so embedded that regulation becomes nearly impossible. History suggests that's the pattern: technology races ahead, builds constituency, normalizes its presence, and then regulation struggles to catch up. We've seen it with social media, with gig work classification, with cryptocurrency. By the time rules arrive, the power is already concentrated and the business models are locked in.

The window to shape AI development, to build in worker protections, environmental safeguards, and competition rules before six companies turn into three, is closing. It's open now. After the next wave of deployment and consolidation, it likely won't be.