Political Tech
A growing conflict between federal centralization efforts and state-level AI legislation is reshaping who actually controls the rules of the AI economy.
NewsOnScale Staff
June 29, 2026
There's a familiar pattern in American technology policy: Washington declares that it will take charge, and the states ignore it. That dynamic is now playing out at full speed in artificial intelligence governance, and the consequences for the AI agent economy — for workers, consumers, and the platforms themselves — could be enormous.
The Biden-era executive order on AI safety set the tone for federal primacy, and the Trump administration, despite its deregulatory instincts in other areas, has shown its own appetite for centralized AI oversight, particularly where national security and economic competitiveness are framed as justifications. The message from Pennsylvania Avenue has been consistent across administrations: let Washington lead.
The states have a different message. They're already legislating.
## A Patchwork That Isn't Going Away
More than a dozen states have passed or are actively advancing AI-related legislation, covering everything from algorithmic hiring decisions and automated content disclosure to healthcare data governance and deepfake prohibitions. Colorado's AI Act, modeled loosely on the EU AI Act's risk-tiered framework, is among the most comprehensive. California has moved on multiple fronts simultaneously. Texas, Illinois, and New York are not far behind.
For companies deploying AI agents at scale — the core constituency of the emerging AI agent economy — this fragmentation isn't a minor compliance headache. It's a structural problem. A customer service AI operating in all 50 states must now contend with dozens of overlapping, sometimes contradictory disclosure requirements, bias audit mandates, and consumer opt-out mechanisms. Legal teams are being rebuilt around this reality.
But the fragmentation argument cuts both ways. Critics of federal centralization argue that state experimentation is precisely how democratic governance is supposed to work — that waiting for a unified federal standard means waiting indefinitely while harms accumulate. They point to financial regulation, environmental law, and consumer protection as precedents where states moved faster than Congress and ultimately forced federal action by raising the floor.
## The Accountability Gap at the Center
What's worth scrutinizing in the push for federal primacy is the question of what federal oversight would actually look like in practice. There is currently no dedicated federal AI regulatory agency. The responsibilities are distributed — and often contested — among the FTC, NIST, sector-specific regulators like the CFPB and FDA, and executive branch offices with limited enforcement teeth.
When the White House talks about centralizing AI regulation, it is describing a policy goal, not an existing institutional capacity. States, by contrast, already have attorneys general with enforcement authority, consumer protection bureaus, and in some cases dedicated technology oversight offices. The institutional readiness argument, rarely surfaced in coverage of this debate, actually favors the states in the near term.
There's also a political economy dimension worth naming. Large AI platform companies have, almost universally, lobbied for federal preemption of state AI laws. A single federal standard — especially one shaped with significant industry input — is vastly preferable to 50 separate compliance regimes. When companies and the executive branch align on centralization, the accountability journalist's instinct should be to ask who benefits from foreclosing local enforcement options.
## What Comes Next
The most likely near-term outcome is continued legal conflict. States will pass laws, companies will challenge them in federal court, and preemption doctrine will be litigated in ways that could take years to resolve. In the interim, the rules of the AI agent economy will be written not by any single deliberative body but through an accumulation of lawsuits, agency guidance, and local enforcement actions.
For the platforms and agent developers building on top of these systems, regulatory uncertainty is itself a form of market power — those with large legal teams survive it better than startups or independent operators. That dynamic should be part of the governance conversation, too.
The battle over who controls AI rulemaking is, at bottom, a battle over who is accountable to whom. That's not a technical question. It's a political one, and it deserves political scrutiny.