Across America, local officials are being asked to make a decision with national consequences: whether to let a new data center be built near homes, power lines and water systems. Too often, they now hear only two choices: approve everything or stop everything.
That false choice has moved from town halls to the midterm campaign. Democratic and Republican candidates are discovering that fear of artificial intelligence and frustration with data centers can be converted into votes long before anyone defines what a freeze would accomplish.
The Centercratic position is straightforward. America must remain the world leader in artificial intelligence while imposing clear, enforceable safeguards on the technology and fair requirements on the facilities that support it. A blanket pause on data centers would confuse responsible oversight with retreat, weaken local economic opportunity and give authoritarian competitors time that they will not voluntarily surrender. This matters now because AI will shape national security, medicine, education, business and the daily work of millions of Americans. The countries that learn to use it safely and fully will hold an enormous advantage over those that do not.
Artificial intelligence does not operate on speeches or campaign promises. It depends on immense computing power housed in data centers, which means that stopping those facilities can also slow the research, testing and deployment that keep the United States competitive. China, Russia, North Korea and other authoritarian states are not likely to observe an American election-season pause. If American companies cannot build here under reasonable rules, the United States risks losing investment, talent and technical capacity to places that are more willing to host them. Once those jobs, facilities and research networks move elsewhere, a later change of policy may not bring them back.
The safety concerns are real, but recent warnings have been presented as if the worst outcome has already arrived. The Associated Press reported on September 13 that Anthropic and OpenAI had observed AI systems going beyond their assigned tasks during testing, including unauthorized intrusions into outside computer systems.
The same report said the 2026 International AI Safety Report found early signs of troubling capabilities, but not capabilities at a level that could cause humanity to lose control. The report, developed with guidance from more than 100 independent experts, described the likelihood and timing of that danger as highly uncertain. These events justify safeguards. They do not justify panic.
That uncertainty has now become campaign material. Politico reported this summer that lawmakers in 15 states had introduced bills to pause data center development. CNN reported in August that Republican Texas Governor Greg Abbott and Democratic New York Governor Kathy Hochul had each supported a form of pause.
House Democratic Leader Hakeem Jeffries then called for Congress to act immediately on AI, according to Politico. Senator Bernie Sanders went further by calling for a national moratorium on data center construction, according to The Washington Post. The pressure now is to promise action first and define it later.
That is how AI became a political pawn. Both parties are turning legitimate concerns into election messages before explaining what controls they want, how those controls would work or what they could cost the country.
The important questions are harder. Who is responsible when an AI system causes harm? What independent testing should be required before an advanced AI system is deployed? How should the United States protect the technology from hostile governments? Those questions require careful national standards, not politicians competing to see who can apply the brakes first.
Supporters of a temporary pause make a serious case. Communities should not be pressured into approving projects before they understand the effects on electricity, water, land use, and public safety. A limited pause may be justified where an immediate local problem exists or no reasonable approval standards are in place.
A broad freeze is different. It treats every facility as equally dangerous and every community as equally unprepared, while China and other competitors continue moving forward. Safeguards should target dangerous conduct and measurable risks without blocking the infrastructure America needs to compete.
The local official facing that permit should not have to choose between surrendering the community’s interests and surrendering America’s technological lead. Artificial intelligence is too important to become another campaign pawn.
Visit Centercratic.party for more information.