Americans Want AI Safety Rules — But Not AI Data Centers Nearby

New polling data from Gallup has uncovered a tension in how Americans think about artificial intelligence: they support rules and safety testing for AI technology, yet 71% oppose building the data centers needed to power AI in their local communities. This is a textbook example of "not in my backyard"—wanting the benefits of a technology without accepting its physical presence nearby.

Gallup's March 2024 survey was the first time the polling firm asked Americans about AI data center construction. The results showed not just opposition, but strong opposition: 48% of respondents said they strongly oppose having a data center built in their area. At the same time, other Gallup research found that Americans favor safety regulations for AI, independent testing of AI systems, and international cooperation on AI development. The contradiction is striking: support the technology in principle, reject it in practice.

A Pattern We've Seen Before

This dynamic is not new. Seventy-one percent opposition to local data centers mirrors the 53% of Americans who oppose nuclear power plants in their backyards—even though 61% of Americans support nuclear energy nationally in 2024, nearly matching the highest level Gallup has recorded since tracking the question began in 1994.

We have seen this pattern repeatedly as new technologies move into communities. Cellular towers met fierce local resistance in the 1990s, even as people clamored for mobile phone coverage. Fiber optic cables hit neighborhood opposition while consumers demanded faster internet. Americans have shown a consistent ability to embrace a technology's benefits while rejecting the infrastructure that makes it work.

The nuclear comparison is worth attention because AI companies are increasingly turning to nuclear power to run their data centers. Major tech firms have announced plans to restart closed nuclear plants and build small modular reactors—essentially custom-built nuclear power stations—to supply the enormous amounts of electricity that AI computing demands. This creates a layered problem: both the data centers themselves and the power plants that feed them face community resistance.

Why Communities Are Hesitant

Data centers do bring economic benefits. They attract capital investment, create construction jobs, and provide ongoing employment and property tax revenue for local governments. Yet Gallup's data suggests these arguments have not persuaded communities to support them.

Part of the reason lies in how data centers actually operate. Unlike a traditional factory or office building, a modern data center requires relatively few permanent workers. A facility worth hundreds of millions of dollars might employ only dozens of technical staff to keep it running. The economic case looks less compelling when communities realize they won't see sustained job growth.

The technical demands of AI data centers also set them apart from earlier technology infrastructure. AI workloads—the computing tasks involved in training and running AI systems—require extraordinary amounts of electricity, sometimes 10 to 100 times more than older data centers. They also consume vast quantities of water for cooling. A single facility can use millions of gallons of water per day, which raises real concerns in regions already struggling with water scarcity.

Energy and Policy Conflicts

The polling reveals another complication: Americans' stated energy preferences don't align with what they'll accept locally. Fifty-six percent of Americans support increased funding for next-generation nuclear research, and 72% favor more investment in wind and solar power, according to ecoAmerica's 2024 survey. Yet local opposition to both nuclear plants and data centers creates a practical roadblock.

Nuclear energy support has recovered significantly. In 2016, Americans opposed nuclear energy for the first and only time since Gallup began tracking it. Today's 61% support includes 29% who strongly support it and 32% who support it somewhat. That turnaround reflects growing concerns about climate change and energy reliability.

Energy concerns remain a factor for many Americans. Thirty-five percent worry "a great deal" about energy availability and cost—down from 47% in 2022 but still well above historical lows of around 22%.

The Challenge for Policymakers

There is a genuine tension here that touches on how AI development moves forward in the United States. Americans are broadly comfortable with oversight of AI—they want safety rules, testing, and international cooperation. But they don't want data centers in their communities, and they don't want the power plants that feed them. This puts pressure on federal and state officials to find incentive structures that can change local opinion.

The traditional tools—property tax breaks and job creation promises—may not work as well with data centers, given how few permanent jobs they create. Communities appear skeptical about AI technology's local impacts in a way they weren't about earlier technology waves.

The timing matters too. AI companies face intense pressure to build computing capacity as fast as possible to train larger models and handle growing demand. Long permitting processes and community opposition in the U.S. could give companies with existing facilities or international operations a competitive advantage. If domestic AI development slows because of local resistance, American AI companies may fall behind international competitors without the same regulatory and community constraints.

Solving this will likely require more than the standard playbook. Companies and governments may need to invest in deeper community engagement and transparency about how these facilities work, what they actually demand from local resources, and how communities might benefit beyond just jobs and taxes. That's a different conversation than what happened with cell towers or fiber cables, and it reflects how seriously Americans are thinking about AI's infrastructure needs.