“If states cannot act, who will stand between powerful platforms and the public?” That question haunted lawmakers and technologists this spring when, amid the chaos of a sprawling budget debate, Senator Ted Cruz quietly proposed a ten‑year federal moratorium that would bar states from regulating artificial intelligence. The proposal — a political fuse lit inside a much larger bill — forced a stark choice: preserve a single, uniform market for AI development, or leave states free to experiment, restrict, or protect their citizens from the harms that increasingly powerful systems can create.
At stake is more than market design. As Bruce Schneier and other observers have noted, the governance of AI is where questions of energy consumption, data control, copyright, employment, democratic trust, and public safety collide. The moratorium would freeze the policy space for a decade, a long stretch in a field advancing by the month, and it would remove a primary mechanism for local experimentation and accountability that democracies have relied on for other fast-moving technologies .
Background: what the moratorium would do and how it surfaced
Senator Cruz’s amendment — introduced during negotiations over a mammoth appropriations and budget package — sought to preempt state laws governing AI, effectively consolidating regulatory authority at the federal level for ten years. Proponents framed the measure as a way to avoid a patchwork of state rules that could hamper innovation and create compliance headaches for companies operating nationwide. Opponents warned it would hand de‑facto regulatory power to the largest AI firms, lock in a permissive federal stance, and curtail democratic experimentation at the state level.
Why the debate cuts deep
- Speed and concentration of technological power. Modern AI systems have become concentrated among a few very large companies that control massive compute resources and datasets, and who can therefore set de‑facto norms for design and deployment. That concentration raises conventional monopolistic concerns — and practical governance problems when systems affect energy use, labor markets, or civil rights. Schneier and others argue the stakes are institutional as much as technical: choices about transparency, auditing, and oversight will determine whether AI amplifies public goods or private power .
- Local harms need local remedies. States often encounter the first tangible consequences of new technology — from housing markets disrupted by gig platforms to local labor displacements and neighborhood‑level surveillance. A federal moratorium would remove a key lever for states to test rules tailored to these localized harms and to develop enforcement practices that can later scale up or inform federal policy .
- Democratic legitimacy and experimentation. Regulatory laboratories have long been a feature of federal systems: states test reforms, refine them, and provide models for national adoption. Blocking that route for a decade would slow adaptive governance precisely when rapid learning is most needed, according to civil‑society and academic observers who stress civic engagement, audits, and public scrutiny as central to legitimacy .
Arguments in favor of a federal moratorium
- Uniformity and certainty for industry. Supporters say a single federal standard reduces compliance costs and uncertainty, which can encourage investment and slower‑moving regulation that is more technocratic and consistent.
- Preventing a regulatory race to the bottom or chaotic fragmentation. Some policymakers fear wildly divergent state rules that complicate interstate commerce and create enforcement gaps that adversarial actors might exploit.
- Centralized capacity. The federal government can marshal resources for national security, cross‑state coordination, and large‑scale technical standards in ways individual states cannot.
Arguments against the moratorium
- Entrenchment of incumbent power. A decade of federally imposed silence can allow already dominant firms to set norms by default — in training data, model designs, and business models — with little external contest or public scrutiny, reinforcing incentives that favor opacity and proprietary control over public accountability .
- Loss of democratic experimentation and responsiveness. States acting as laboratories of democracy can innovate faster than a lumbering federal process; their ability to impose spatially targeted constraints (for example, on public‑sector procurement, deployment in schools, or local surveillance systems) would be eliminated for a long period .
- Practical harms are local, enforcement is local. Energy strain from massive datacenter loads, neighborhood surveillance, or local labor market dislocations manifest unevenly. Blocking state remedies eliminates a tool for immediate mitigation and meaningful recourse for affected citizens.
How technologists, policymakers, users, and adversaries see it
- Technologists: Views are split. Some industry leaders welcome uniform standards that allow them to scale products quickly. Other researchers and engineers favor strong auditing, access to reproducible benchmarks, and requirements for transparency that can be implemented at state or sectoral levels to ensure safety in high‑risk deployments .
- Policymakers: Federal officials argue for national frameworks; many state legislators, attorneys general, and governors want the right to protect their constituents. Experts caution that good governance will require both federal baseline rules and the flexibility for states to act where federal policy lags or is insufficiently granular .
- Users and civic groups: Civil‑society organizations stress transparency, contestability, and public auditing — tools that are more likely to arise from a pluralistic policy ecosystem rather than a single, prolonged federal freeze. Public trust, they argue, is earned through oversight and participation, not by delegating choices to corporate actors or a distant bureaucracy .
- Adversaries: Nation‑state or commercial actors seeking to exploit regulatory gaps would benefit from a uniform, permissive regime if it reduces enforcement complexity. Conversely, fragmented rules with rigorous enforcement in some jurisdictions could complicate malign actors’ strategies.
Policy alternatives and middle paths
The extremes are not the only options. Pragmatic alternatives could include:
- Short, conditional federal pauses tied to milestones. Rather than a decade‑long moratorium, a time‑limited pause linked to specific technical benchmarks or to the existence of minimum federal safeguards could balance certainty with accountability.
- Preemption with carve‑outs. The federal government could set national minimums while explicitly preserving state authority in well‑defined areas, such as consumer protection, labor regulation, and energy siting.
- Cooperative federalism models. Federal baseline rules plus funding and technical support to states for compliance, auditing, and enforcement would combine scale with local responsiveness.
- Mandated transparency and auditability. Regardless of venue, requiring independent audits, access to provenance metadata, and impact assessments for high‑risk systems would give both citizens and regulators the information they need to act — a point underscored by experts urging model cards, documentation, and public scrutiny .
What this moment reveals about governance
The Cruz amendment surfaced a deeper unease: whether institutions can keep pace with technology that shifts power as quickly as it changes code. Schneier and others have argued that technology governance is less about finding a single perfect law than about building durable practices — audits, contestability, public participation, and cross‑jurisdictional learning — that preserve democratic choice amid rapid change .
Concluding analysis: the risks of locking the policy field
Put simply, a ten‑year federal moratorium risks freezing a policy landscape in a configuration that favors concentration and opacity just when experimentation, transparency, and local remedies are most needed. Federal uniformity has value, but not at the cost of eliminating democratic avenues for accountability and rapid local response. If the goal is durable, trustworthy systems that serve the public, policymakers should avoid blunt, long‑term preemption and instead design layered governance: federal baselines, state flexibility, enforceable transparency requirements, and funded public audits.
So we return to the opening dilemma: who will stand between powerful platforms and the public good? If we answer with a decade of silence from states, we must accept that much of the regulatory heavy lifting will fall to the entities that profit from the status quo. If we instead preserve the right of states to experiment while pushing for federal standards of transparency and safety, we keep open the democratic tools — trial, error, and improvement — that have long been the strength of federated systems. Which path will best protect citizens from the unintended — and sometimes devastating — consequences of rapidly deployed AI?
Source: https://www.schneier.com/blog/archives/2025/12/against-the-federal-moratorium-on-state-level-regulation-of-ai.html
