The Consent Deficit - How America's Deregulatory Maximalism May Be Losing the AI Race It Thinks It's Running

There is a particular kind of policy failure that masquerades as boldness. It moves fast, it makes large gestures, and it mistakes the absence of guardrails for the presence of a strategy. We are watching one unfold in real time — and its consequences may be far more consequential than the people engineering it, seem to understand.

The Trump administration has made artificial intelligence a centerpiece of its second-term agenda. The Stargate Project, the deregulatory posture, the open hostility to any framework that might slow AI development — all of it is framed as a race against China. Win first. Regulate later. Trust the market.

It is worth asking, seriously, and without partisan noise, whether this approach is working. Not whether it is philosophically coherent (it isn’t), but whether it is achieving its stated objectives. The data suggests it is doing something closer to the opposite.

The Rebellion Nobody Predicted

Something has shifted in American public life over the past eighteen months, and it hasn’t received nearly the attention it deserves. A genuine grassroots rebellion against AI infrastructure has emerged — not from Berkeley or Brooklyn or the usual quarters of tech skepticism, but from Sand Springs, Oklahoma. From rural Michigan. From Warrenton, Virginia, where more than 500 residents showed up to a town council meeting, and nearly 130 people lined up to speak. From communities across 24 states where 142 distinct activist groups have organized against data center construction.

The numbers are striking. More than $64 billion in data center investment has been blocked or delayed by local opposition. Twenty-five projects were canceled in 2025 following sustained community pushback — and 21 of those 25 cancellations happened in the second half of the year, meaning the movement is accelerating. Opposition groups have learned from each other, shared legal strategies, and built regional coalitions that have turned isolated complaints into coordinated campaigns.

The concerns driving this rebellion are not abstract. They are the concerns of people watching their electricity bills climb while server farms consume the equivalent of 100,000 households per facility. They are the concerns of communities in drought-prone regions watching their water tables drawn down to cool machines they derive no benefit from. They are the concerns of rural towns that were promised economic revitalization and received instead noise, heat, and a smaller share of a utility grid that now serves a very different master.

“Trump wants data centers and Kevin Stitt wants data centers, but these things don’t affect them,” one Trump voter in Sand Springs told the Washington Post. “You know, this affects us.”

That sentence deserves to sit on its own for a moment. A Trump voter in Oklahoma — not a DSA organizer, not a climate activist — articulating the core logic of a populist revolt against the tech industry’s AI ambitions. This is not a left-wing phenomenon. It never was. Ron DeSantis has introduced an AI “bill of rights” giving Florida communities the right to block data center construction. Bernie Sanders has called for a national moratorium. Maine is on the verge of becoming the first state to impose one. New York, South Carolina, Oklahoma, and Vermont are considering similar measures. The politics here do not follow any conventional map.

What they follow, instead, is the map of a population that was never consulted.

The Trust Vacuum and What Fills It

Here is the uncomfortable truth about deregulation as a governance philosophy: it does not actually eliminate the question of trust. It simply relocates it — from institutions that can be held accountable to a market that cannot.

When a government says, in effect, “we will not regulate this technology; we will let the industry self-govern,” it is not creating neutrality. It is making a choice. And that choice, over time, produces a specific outcome: people who do not trust the technology, and who do not trust the industry behind it, and who have no institutional mechanism through which to express or resolve that distrust except through the bluntest available instrument — opposition.

The polling here is unambiguous and worth stating plainly. Only 5% of Americans say they trust AI a great deal. Forty-one percent express outright distrust. Seventy-seven percent believe AI could pose a threat to humanity. When asked how their trust in AI has changed over the past year, more Americans report a decrease than an increase. This is not a static picture of uncertainty. It is a picture of eroding confidence — in a technology that its most vocal proponents insist is the most transformative in human history.

More striking still: 60% of Americans say there should be more regulation of AI, including 59% of Republicans and 52% of independents. Even among people who view AI favorably, a majority want more regulation. The demand for governance is not coming from AI skeptics. It comes from across the ideological spectrum, including from the people who are actually using the technology.

And yet the administration has made deregulation a signature platform position. The absence of federal regulatory action has not built confidence. It has built a vacuum, and that vacuum is being filled — not by rational actors deliberating calmly, but by fear, by local politics, by the blunt instrument of moratorium legislation, and by a generalized sense that the people in charge of this technology are not particularly interested in the concerns of the people who have to live with it.

This is a systems-level failure. When legitimate governance channels are closed, pressure does not dissipate. It finds other outlets.

The China Paradox

And now we arrive at the part of this story that should genuinely unsettle the people who have framed AI dominance as a race against Beijing.

In China, 87% of people say they trust artificial intelligence. In the United States, 32% do. China’s rate of employee AI adoption — 93% using it regularly — dwarfs the global average. Fifty-four percent of Chinese respondents say they welcome greater AI use in daily life. Seventeen percent of Americans say the same.

Let that contrast settle. The country that the Trump administration has explicitly positioned as the adversary in the AI race has a population that is more than twice as trusting of the technology, more than three times as welcoming of its expansion into daily life and adopting it at rates that the United States is not approaching.

The instinctive response to this data — and it is worth naming the instinct precisely so we can examine it — is to attribute the Chinese numbers to authoritarian conformity. The population trusts AI because it has been conditioned to trust what the state endorses. The surveys may be contaminated by social desirability bias. Chinese respondents may say they trust AI because it is not safe to say otherwise.

There is something to this. Cross-national polling in authoritarian contexts requires methodological humility. But the behavioral data is harder to explain away. When 93% of Chinese employees are actively using AI tools at work, something more than performative compliance is happening. The technology is entrenched in daily life — in mobile payments, urban management, healthcare, and education — in ways that reflect genuine integration, not mere ideological endorsement.

And crucially, this integration appears to be at least partly a function of governance confidence — the belief, warranted or not, that the state has done the work of ensuring that the technology serves the population rather than exploiting it. More than 63% of Chinese respondents support national and international AI regulation. The Chinese are not against oversight. They are in favor of it. Their confidence is downstream of that governance framework, not despite it.

The United States has no equivalent. It has a regulatory landscape so fragmented and contested that most of its own citizens do not trust either political party to manage AI. Twenty-three percent trust Democrats to handle it. Twenty-five percent trust Republicans. The rest trust neither, or don’t know.

This is not a posture of confident leadership. It is a posture of institutional collapse — and it is accelerating precisely now when the country needs its population to be a willing participant in technological transformation.

The Neo-Luddite Misreading

It is tempting, particularly in certain tech-optimist circles, to frame the emerging resistance to AI infrastructure as Luddism — as fear-driven irrationality, the timeless story of humans flinching from the future.

This framing is historically illiterate and strategically dangerous. The original Luddites were not afraid of machines. They were skilled craftsmen who understood machines perfectly well. What they were opposing was the deployment of machinery by factory owners specifically designed to destroy their livelihoods, their wages, and their communities — without their consent, without their participation in the gains, and with the explicit connivance of a state that had chosen the owners over the workers.

That is not an irrational position. It is a political position. And what we are watching in Sand Springs and rural Michigan and Warrenton and Memphis and San Marcos, Texas — where hundreds marched against data center development last August — is not fear of the future. It is a political position. These are people saying, with increasing clarity and coordination: not like this. Not without us. Not without our consent.

The failure to hear that signal, to treat it as mere irrationality to be overcome by better PR and “targeted advertising,” is itself a kind of institutional failure. The AI Infrastructure Coalition can blast statements. Tech companies can run ad campaigns. But as one industry analyst noted, “it’s harder to buy public support than it is to buy a power plant.” That is not a solvable communications problem. It is a governance problem.

The Strategic Inversion

Here, then, is the full shape of the irony. The United States has adopted a policy posture premised on the idea that maximum speed and minimum regulation will secure AI dominance. The strategy assumes that governance is friction — that it slows development, that it concedes ground to competitors, that it is a luxury of the already-victorious.

But governance is not only friction. It is also legitimacy. And legitimacy, it turns out, is not a soft concern. It is load-bearing infrastructure for any technology that requires mass adoption to realize its economic and strategic potential.

China, by deploying AI within a framework of state governance — however imperfect, however authoritarian, however deserving of separate critique — has produced a population that trusts the technology and uses it. The United States, by refusing to build a comparable framework of public accountability, has produced a population that distrusts the technology, opposes its infrastructure, and is generating political pressure that will increasingly constrain the very buildout the administration is trying to accelerate.

The deregulatory maximalism did not eliminate the politics. It did not transcend the question of social consent. It created a power vacuum at the federal level that has been filled, messily and expensively, by 142 local activist groups, by moratorium legislation, by DeSantis and Sanders finding unlikely common ground, and by a Trump voter in Oklahoma who knows, with complete clarity, that the people building data centers next to his property are not particularly concerned with his wellbeing.

What we are watching is not a communication failure. It is not a perception problem that better messaging will solve. It is the predictable consequence of a governance philosophy that treats legitimacy as a cost rather than a precondition — and that is now discovering, at scale, that the preconditions were real.

The race against China is real. The urgency is real. The strategic stakes could not be higher. But races are not won by runners who have lost the consent of the road. At some point, the question of whether the American public is a willing participant in this transformation — not a passive recipient, not a collection of obstacles to be managed, but an actual stakeholder with legitimate concerns that governance exists to address — becomes not a philosophical nicety but an operational constraint.

We are at that point now.

• The polling data cited in this piece draws on surveys from Pew Research Center, Edelman Trust Barometer, YouGov, Navigator Research, Data for Progress, and the University of Melbourne/KPMG Global AI Trust Study (2025). Data on data center opposition comes from Data Center Watch and Heatmap Pro.