Candidates Turn Data Center and AI Safety Worries Into a New Campaign Pledge

More than 15 candidates running in state and federal races have signed a new AI Pact that ties data center regulation to AI safety, reflecting how quickly energy use, secrecy, labor concerns and artificial intelligence risks have become campaign issues in the 2026 midterm cycle. The pledge matters because it gives politicians a ready-made framework to promise action on both the physical infrastructure powering AI and the harms the technology can cause.

Dan Osborn, an independent Senate candidate in Nebraska, says the issue cuts through partisan noise. At a recent town hall, he said a self-described hardline MAGA attendee who disagreed with him on most subjects immediately aligned with him on one question: what to do about AI and data centers.

The new pact is designed to capitalize on that overlap. It asks candidates to endorse a set of policy commitments that go beyond broad concern and toward specific guardrails, including safety reviews for AI models, worker protections, and the ability for consumers and businesses to sue when models cause harm. It also calls for tighter scrutiny of data centers, which critics say are being built too quickly, too quietly and with too little accountability.

What is the AI Pact?

The AI Pact is a candidate pledge created by organizers connected to the Political Integrity Project, an anti-corruption group that has previously rolled out commitments for politicians on money-in-politics issues. Its authors say the new pledge is meant to create a baseline for candidates who want to speak about AI without sounding either alarmist or overly deferential to the tech industry.

According to the organizers, the pact is not backed directly by AI safety organizations or technology companies. Instead, it is funded through a progressive political action committee, a detail meant to distinguish it from industry-backed policy efforts that often shape tech regulation debates.

Daniel Lobo-Lewis, one of the pact’s founders, said the pledge was developed through conversations with AI safety advocates and other groups working on legislation, and that the goal was to offer a minimum standard rather than the most aggressive possible platform.

The current signers are mostly Democrats, though organizers say they welcome Republican candidates as well. That matters in a year when data center politics are spilling into races across ideological lines, especially in places where communities are divided over whether new facilities bring jobs, strain utilities or threaten local land use.

Why are data centers suddenly a campaign issue?

Data centers have become a flashpoint because they sit at the intersection of several voter concerns: electricity bills, land use, environmental impacts, corporate secrecy and the rapid expansion of AI services. For many communities, the debate is no longer about whether tech companies are building facilities nearby, but about how much power they will consume, what incentives they receive and whether local residents get a meaningful say.

In some places, opposition has been driven by fears about water use, noise and strain on the grid. In others, residents worry about deals negotiated behind closed doors, tax abatements, or the prospect of farmland and open space giving way to industrial development.

That tension has given candidates a political opening. They can criticize the excesses of Big Tech while still acknowledging that some projects bring construction work, tax revenue and long-term jobs. The AI Pact is built around that balancing act.

How the pledge frames data center regulation

The pledge avoids calling for a blanket halt to data center construction. Instead, it asks signers to support policies that would block tax giveaways, secret arrangements and excessive energy consumption by data centers. That language is deliberate: it gives candidates room to support some development while signaling resistance to what organizers see as the worst practices in the industry.

For labor-friendly candidates, that flexibility is important. Several unions have continued to support building data centers, seeing them as sources of employment and infrastructure spending. The pact’s authors appear to be trying to avoid forcing politicians into an all-or-nothing position that could alienate workers or local governments looking for investment.

Who is signing the pact?

The list includes more than 15 candidates from across the country, though the strongest early uptake has come from Democrats and left-leaning independents. The signers include people who have made opposition to data center development part of their broader campaign message, as well as others who are focusing on AI regulation as a stand-alone issue.

Among the more prominent names is Justin Pearson, the Tennessee Democratic representative who gained national attention through his resistance to pollution associated with SpaceX-linked operations. Another signer, Will Lawrence, won a competitive Michigan primary in a district where data center proposals had already stirred public backlash.

Osborn, who is also a Navy veteran and labor leader, represents another version of the same political argument: data centers may create work, but that does not mean Congress should let the industry expand without rules.

Osborn said the facilities can bring substantial infrastructure and jobs, but argued that lawmakers still need to step in and make sure the country “gets this right.” He suggested that pausing construction briefly would not be unreasonable if it gives regulators time to catch up.

Across the aisle, the issue is beginning to draw attention as well. The pact’s organizers say they would welcome Republican signers, though so far the list has skewed heavily toward Democrats.

How does AI safety fit into a data center fight?

AI safety enters the picture because data centers are the physical foundation of modern AI systems. The same facilities that power cloud computing and model training also make the scale of AI expansion visible to ordinary voters. That makes data centers a political symbol for a broader debate about whether AI is being deployed faster than governments can regulate it.

The AI Pact connects those two debates directly. It urges candidates to support mandatory safety evaluations for AI models, a policy idea that would force companies to face external review before releasing or expanding powerful systems. It also backs the concept of an AI dividend, or a mechanism that would return some of the gains from AI-driven productivity to workers and the public.

Perhaps most notably, it wants consumers and companies to retain the right to sue over harms caused by AI systems. That provision would go further than many current proposals, which often focus on transparency or voluntary commitments rather than legal liability.

What the five promises are trying to do

The pact’s five commitments are meant to serve as a political floor. In practical terms, they are designed to help candidates say they support AI innovation while still drawing a line around safety and public accountability.

• Support mandatory safety reviews for AI models.
• Back an AI dividend or similar worker-benefit policy.
• Ensure people and businesses can sue over AI-related harms.
• Oppose tax breaks and secret deals for data centers.
• Push for limits on excessive energy use by data center facilities.

For voters, that combination may sound less abstract than the typical AI policy debate. Rather than speaking in terms of model architecture or technical benchmarks, the pledge translates AI governance into local consequences: utility bills, employment, safety and the balance of power between residents and large companies.

Why are politicians shifting on data centers now?

Politicians are shifting because the political costs of ignoring data centers are rising. Officials who once welcomed every new facility as a sign of economic growth now face organized opposition from residents who believe the projects are driving up costs, consuming too much electricity or being approved without sufficient transparency.

That has pushed leaders from both parties to respond more aggressively. In Pennsylvania, Democratic Governor Josh Shapiro has sought to reassure skeptical voters that his administration is paying attention to the issue. In Texas, Republican Governor Greg Abbott has recently taken a tougher line, arguing that the industry has overplayed its hand in his state.

Abbott’s comments and recent temporary moratorium on data center construction show how much the politics have changed. A governor who once might have framed the facilities as a growth opportunity is now talking about them as a problem that needs restraint.

The same pressure is showing up in Washington. Some policy proposals coming from the White House have focused on voluntary safety review systems for AI models and on voluntary commitments to shield ratepayers from rising electricity bills tied to data center growth. That approach has drawn criticism from organizers like those behind the AI Pact, who argue that voluntary promises do not amount to real enforcement.

Lobo-Lewis dismissed the federal approach as insufficient, saying the administration’s gestures do not look like the kind of binding reform the moment requires.

What are candidates saying about the risks?

Candidates signing the pledge are emphasizing different parts of the same worry. Some focus on environmental and utility costs, others on labor displacement, and others on the broader question of whether AI systems are being deployed in ways that weaken public protections.

In Texas, Democratic gubernatorial candidate Gina Hinojosa has argued that the industry is moving faster than lawmakers and is threatening jobs, health and safety. She says the state needs clear ground rules and rejects the idea that schools, hospitals or public services should simply replace human workers with AI tools wherever companies find it convenient.

Hinojosa said AI can be useful, but insisted it must be governed responsibly. In her view, current policy has not yet caught up to the pace of deployment, leaving Texans without adequate safeguards.

Her opponent, Abbott, has gone in the opposite direction by accusing the industry of bringing problems on itself. That contrast captures the emerging split in how elected officials are trying to win on the issue: one side frames data centers as a threat needing guardrails, the other as a development opportunity that companies must earn by playing by the rules.

How are national AI politics influencing local races?

National AI politics are increasingly shaping local races because the same voters who worry about a utility bill or a nearby industrial project also see headlines about chatbots, automation and job loss. The issues now reinforce one another. A resident who is skeptical of a data center may also be uneasy about AI products that seem to threaten employment, privacy or children’s safety.

That overlap has encouraged candidates to speak more explicitly about the downstream effects of AI. In March, Sen. Bernie Sanders proposed a national data center moratorium until Congress passed broader AI protections, linking the physical buildout directly to the technological risk. Since then, a small but notable group of politicians has adopted similarly integrated messaging.

Texas Attorney General Ken Paxton recently added his own version of that strategy by releasing a data center plan that would make operators legally responsible for chatbots that could undermine children’s safety. It is a different political style from the AI Pact, but it points in the same direction: candidates increasingly want to show they understand that AI is not only a software story but also an infrastructure and accountability story.

What voters seem to care about most

Polling suggests that the public is still more worried about data center environmental impacts than about AI itself. That is one reason some strategists believe candidates cannot assume that every data center critique will automatically translate into broad AI skepticism.

For many voters, the concern is practical rather than philosophical. They may not be deeply engaged in the technical side of machine learning, but they do understand what a power-intensive facility might do to their electric bill, land values or water supply. That makes data centers easier to campaign on than AI alignment or model evaluation.

Still, the public conversation is evolving. As more people connect data centers to the AI products they use every day, the political line between infrastructure and software is starting to blur.

Comparison of the main policy approaches

The table below shows how the AI Pact compares with other recent political responses to data centers and AI safety.

Why the pact could matter beyond this election

The pact may matter most not because every candidate who signs it wins, but because it helps normalize a new category of tech policy politics. Candidates are no longer being asked only whether they like or dislike AI; they are being asked how much control the public should have over the companies building the systems and the infrastructure that powers them.

If the issue keeps gaining traction, future campaigns may increasingly treat data centers the way they treat factories, pipelines or large-scale energy projects: as assets that can create jobs, but only under conditions that protect communities. That shift would be significant because it would move the debate from abstract concern to concrete bargaining over rules, costs and liability.

It could also widen the field for candidates who want to distinguish themselves from both Silicon Valley boosters and purely anti-tech critics. The AI Pact offers one possible middle path: not a full stop, but a demand for enforceable limits.

How local politics could shape the national AI debate

Local politics may end up determining how far the national AI debate goes. If state and congressional candidates continue to talk about land use, electricity, worker displacement and legal accountability in the same breath, lawmakers in Washington will face growing pressure to turn scattered concerns into legislation.

That pressure is already visible in the way campaign messages are changing. Candidates in rural states, industrial districts and fast-growing suburbs are all finding ways to say that the AI boom should not come at the expense of ratepayers or local communities.

Osborn’s town hall anecdote shows how quickly the issue can bridge partisan divides. Even voters who may disagree sharply on immigration, guns or climate can still agree that big tech companies should not be allowed to build quietly, consume massive amounts of power and leave everyone else with the bill.

Osborn argued, with a touch of irony, that Washington often underestimates how clearly ordinary people can see the tradeoffs in front of them.

The political challenge now is whether that common sense can be turned into law.

Timeline of the issue

For now, the AI Pact remains a relatively small initiative. But in a campaign year defined by anxiety about technology, utilities and the concentration of power, even a modest pledge can signal where the conversation is heading next.

And if the rise of the pledge says anything, it is that candidates are no longer treating data centers as a niche development issue. They are treating them as a proxy for the future of AI itself.