Artificial intelligence is barging into the workplace and transforming the battlefield. Now, it’s coming to the ballot box, promising to rewrite the DNA of politics.
Driven by a flood of Silicon Valley money, AI has emerged as one of the biggest financial forces in this year’s US elections, fueled by hundreds of millions of dollars from tech billionaires led by venture capitalist Marc Andreessen and OpenAI co-founder Greg Brockman. Their goal is simple: elect a slate of candidates who will heed their calls for looser government oversight of the nascent technology.
All that spending has run headlong into deepening bipartisan resentment of AI itself, touched off by the nationwide proliferation of data centers and intensified by companies embracing the technology in a bid to reduce labor costs. In polls, voters increasingly blame AI for electricity bills that have surged as much as 267% and say they fear that it will eventually take their jobs, an extension of the lasting economic unease that has defined the last several US elections.
On the campaign trail, AI is reworking the playbook for how to run for office, just as it’s changing the way business is done across the economy. There’s now a mad dash by candidates and consultants to deploy the technology for an edge in producing ads and targeting voters. At the same time, the AI models used as a tool by cash-strapped campaigns can be wielded by opponents and outside agitators to spread disinformation and doubt among the electorate.
“This is the first AI election from a substantive perspective,” said Beth Simone Noveck, director of the Burnes Center for Social Change at Northeastern University. “It’s front and center in the substance of the conversation.”
November’s congressional contests promise to be the most consequential in years, serving as a referendum on President Donald Trump’s stewardship of the economy — especially his bet on the AI boom as an engine of growth. For Trump, there’s more on the line than just his economic legacy: losing Republican control of the US House and Senate threatens to unleash an effort to impeach him for a third time and to initiate a wide range of investigations into his administration.
AI’s influence has permeated nearly every aspect of the midterms, from financing to core issues facing voters, and it’s poised to play an even bigger role in the 2028 race to succeed Trump. The campaign is unfolding against an AI-fueled market rally that points to risks that the technology could widen the gap between rich and poor, a concern highlighted by Pope Leo XIV in a 43,000-word missive on AI that urged governments and industry to keep it from “dominating humanity.”
Younger voters in the US increasingly see AI wiping out entry-level opportunities, and polls show their rising hostility toward the technology. Graduates at the University of Arizona booed former Google Chief Executive Officer Eric Schmidt when he brought up AI in a commencement speech, signaling that a crucial bloc remains unconvinced about the benefits promised by the tech industry.
Races across the country show signs of AI’s pervasive presence. In Louisiana, a super political action committee promoted the first AI-generated ad in the state, depicting a blue-haired liberal protester and mustachioed barista horrified over a Republican congressional contender. In Texas, the conservative PAC Citizens for Sanity in June began running an AI-enabled ad against Democratic Senate candidate James Talarico depicting him wearing a dress and singing lovingly about transgender children.
“The AI genie is out of the bottle,” Virginia Democratic Senator Mark Warner, who’s made AI issues a central plank in his reelection campaign, said in an interview. “You can almost not make a better bogeyman than, ‘Oh my gosh, are the robots going to take over? Are we going to have jobs?’”
Silicon Valley’s biggest AI developers sense the risk that a voter backlash poses to the hundreds of billions of dollars in spending they’ve planned on new infrastructure, and they’re shielding those bets with a wall of money in campaigns across the country.
In total, the industry has pledged $275 million toward the midterms at the federal and state level, and it has already spent $44.5 million on federal primaries. Those commitments, if made good, would eclipse the $133 million that the cryptocurrency industry spent on dozens of congressional candidates in 2024, when it helped to topple multiple Democratic incumbents, including Senate Banking Committee chair Sherrod Brown.
Perhaps no other race sums up the stakes for AI better than one for a House district in central Manhattan. There, Alex Bores, a Democratic state assemblyman who’s spearheaded tougher AI safety legislation, became a target for millions of dollars in spending by industry-backed groups. The contest quickly became a broader proxy battle between industry rivals OpenAI and Anthropic PBC over how firm a hand the government should take with the technology.
Representing one side is Leading the Future, a super PAC funded by Andreessen, Brockman and other tech industry figures that plans to spend at least $125 million across the US on candidates favoring a softer approach to AI regulation. It opposes states taking the lead on AI rules, arguing the federal government should pass a lighter-touch national standard.
The other side is led by Public First Action, a political nonprofit that funds a super PAC backing candidates who favor tougher AI rules, including requirements that companies disclose safety risks to the government. The nonprofit has received $20 million from Anthropic, whose insistence on extra guardrails for its technology touched off a bitter dispute with the Pentagon. The group’s super PACs have received donations from current and former former engineers at OpenAI, Anthropic and Meta Platforms Inc.
In New York, Leading the Future’s Democratic arm spent over $8 million opposing Bores, yet that opposition helped boost his profile, spurring significant contributions from AI safety advocates and tech workers across the country.
Public First’s Democratic group spent over $15 million on his behalf, citing Leading the Future’s involvement as proof that he would crack down on corporate power. It was joined by super PACs including one funded by San Francisco billionaire Chris Larsen that began running pro-Bores ads blaming AI for “giant data centers, draining our water supply, relentless noise,” as well as the erasure of “millions of jobs.”
Bores ultimately lost in this week’s primary to fellow New York Assemblyman Micah Lasher, a longtime political operative who was favored to win the race and had previously served as an aide to Michael Bloomberg, the founder of Bloomberg LP, who contributed $10 million toward Lasher’s candidacy.
All that money gushing in from both sides is injecting the topic into other races across the country and spurring AI safety advocates to donate elsewhere. Larsen's group, which was founded to boost Bores, has started spending in favor of Colorado State Representative Manny Rutinel, who’s running for Congress and has previously backed state-level rules to keep AI from discriminating against minorities.
Other advocacy groups have emerged with a goal of swaying midterm voters. An organization known as the Innovation Council, backed by venture capitalist and former Trump adviser David Sacks, intends to spend up to $100 million defending the president’s AI agenda. Set up as a political nonprofit, the council doesn’t have to disclose its donors, and it hasn’t started spending on races.
“If you think about AI, everything feels like it’s out of peoples’ control, it’s in the hands of these tech titans,” said Sarah Kreps, a professor focused on tech and politics at Cornell University. “That’s creating backlash.”
Outside New York, ads funded by AI-backed super PACs largely steer clear of the technology and focus instead on other topics that resonate with voters, a tactic deployed by the crypto industry in 2024. Even so, the sheer volume of money pouring into races across the country has put AI on the agenda — including in districts where it would typically be less of a priority.
In Texas, Republican congressional candidate Jace Yarbrough published an op-ed in February hailing the virtues of AI, calling for the US to “own the future of AI by investing in American innovation.” Leading the Future spent $130,000 backing him, and Yarbrough won his runoff election.
Meanwhile, in Illinois, Jesse Jackson Jr., son of the late civil rights icon, published an op-ed in January hailing AI as potentially life-saving for caregivers, raising an issue that was new for his majority-Black working class district on the south side of Chicago. When he last held office in 2013, AI policy didn’t even exist.
Leading the Future’s Democratic arm spent over $1.4 million on Jackson’s campaign after he published the op-ed, but he still lost the race. Jackson’s opponent, progressive Illinois State Senator Robert Peters, said he believes he was targeted because he’d introduced legislation to regulate the industry.
“This creates a chilling effect down-ballot,” Peters said. “This impacts how people at the state legislature or municipal level think about these policies because they see these rich billionaire corporate leaders who are flooding money into the political system and willing to lie to the public.”
But there’s a flip side: some legislators have emerged from races with AI spending more motivated than ever to regulate the industry. In May, Peters joined 51 Illinois state senators to vote for an AI safety bill that would require developers to submit to third-party audits, a provision long opposed by the industry.
While the nitty-gritty of tech regulation ranks lower as an election issue, voters are lumping AI into their broader misgivings about the economy. The industry’s breakneck plan to spend as much as $725 billion this year on data centers is tied to a central theme of the campaign: affordability.
Public ire has zeroed in on energy-guzzling AI facilities that are driving up wholesale power prices more than three-fold in some areas. Many voters say they’ve had enough, with seven in 10 Americans now opposing new AI data centers, according to a Gallup survey released last month. Concerns over energy and water consumption ranked as the top factors, the results showed.
“The data center debates are galvanizing discussion about the role of this industry and what its impact will be,” said Noveck, who served as deputy US chief technology officer in the Obama administration and has published a new book called “Reboot: AI and the Race to Save Democracy.”
“We are coming to recognize these things are not just these far-off abstractions but they are being powered by infrastructure in our backyard,” she said.
Progressives have homed in on data centers as a potential winning issue in their bid to wrest House and Senate control from Republicans. Democratic Senator Bernie Sanders and Representative Alexandria Ocasio-Cortez in March proposed a nationwide moratorium on new centers. While their bill stands little chance of passing, it nonetheless reflects growing voter animosity for new AI infrastructure seen more visibly at the local level.
In April, Compass Datacenters cited public resistance in abandoning a years-long effort to build part of a 2,100-acre tech corridor in Northern Virginia — an area known as the world’s top data-center market. Earlier this month, New York lawmakers approved a one-year moratorium on new data centers that would be the first statewide pause in the US if signed by Governor Kathy Hochul.
That sentiment is surfacing in campaign ads in New Jersey, Wyoming, Virginia and Montana. “Rob Menendez has worked to make life more affordable,” says a voiceover in a spot supporting the New Jersey Democratic representative. “Now, he’s taking on big corporations whose data centers are causing energy costs to skyrocket.”
Because data center decisions including permitting often fall to local officials, AI companies are spending to sway those races, too. In Texas, Meta’s PAC spent over $1.3 million on Texas legislature races in early March, including $174,000 on behalf of state senate candidate, David Cook.
One of Cook’s opponents, Rena Schroeder, kicked off her campaign by urging an end to data center construction. She learned weeks later that Meta was backing her opponent and now believes the company’s funding made it impossible for her to win. “Texas was sold to the highest bidder,” Schroeder said.
Some Republicans are turning against data centers over energy concerns and reports that the sites fail to create many new jobs. In Pennsylvania, Republican gubernatorial candidate Stacy Garrity said she supports a moratorium on new data centers. Her opponent, Democratic Governor Josh Shapiro, has been a major supporter of the data-center buildout in the state.
“There’s frustration, there’s confusion,” Garrity said during a campaign stop in June. “People think that it’s going to go in their backyard. They’re afraid. These concerns have to be addressed.”
For all the misgivings about AI, candidates have been quick to embrace it, just as they adopted past innovations like television, text messaging and social media in their never-ending quest for an edge.
AI’s presence is most evident in the wave of computer-generated videos permeating high-profile contests in California, Texas and Kentucky. So-called deepfakes have shifted from a novelty — like clips showing Joe Biden and Trump eating ice cream on a park bench two years ago — to an increasingly sophisticated and ubiquitous element of campaigning.
Former reality television star Spencer Pratt’s long-shot quest to become the next Los Angeles mayor benefited from an animated series of AI videos that went viral last month, spurring an increase in contributions to his campaign. One of the spots portrayed Pratt as a Batman-like superhero battling incumbent Mayor Karen Bass, who’s depicted as a Joker-like villain.
While Pratt ultimately fell short of qualifying for a runoff against Bass, the videos supercharged his candidacy as a political newcomer. “The primary impact these tools can have is they do help upstarts who have a fundraising disadvantage over incumbents to get their message out in ways that were difficult or more expensive,” said Neil Chilson, head of AI policy at the Abundance Institute.
Clips like those have stirred questions about where to draw the line. Banning deepfakes outright remains off the table, owing to free-speech safeguards under the US Constitution. Still, 30 states have rules governing their use in campaign messaging, with most requiring disclosures, including whether AI was used and who funded it. Minnesota and Texas have gone as far as prohibiting political deepfakes a certain number of days before an election.
Candidates targeted by unflattering deepfakes have little recourse to get them pulled. A clip of Representative Sean Casten depicted the Illinois Democrat calling for jailing Trump supporters, eliminating the police and providing fentanyl to the homeless.
Circulated on Facebook by an anonymous constituent using nothing more than Casten’s official congressional portrait, it showed him grinning, hectoring and making suspect salutes. He notified Facebook’s parent Meta of the video, and the company agreed to tag that it was made with AI. But Meta didn’t remove the post, reflecting the tech industry’s largely hands-off approach to AI.
“It would not surprise me to see in every competitive race AI-generated ads of the Democrat hanging out on Jeffrey Epstein’s island and partying with a bunch of young girls,” Casten said in an interview. “We need to be prepared for a lot of that.”
AI is acquiring a permanence in politics in other ways. Americans now look to the technology to inform Election Day decisions, with 55% of US voters under the age of 45 likely to use chatbots to learn about candidates and elections, according to a survey by Data for Progress, a progressive think tank.
In response, strategists are racing to figure out how to game political queries, just as they learned to to optimize search engine results decades ago. The trend is prompting political and tech transparency groups to seek ways to ensure that chatbots offer credible and unbiased responses — a much harder task than for SEO because chatbots respond differently to each user.
Publishing more content may be the best way to ensure that chatbots give favorable information on candidates, said Mark Meckler, who runs Convention of States, a grassroots advocacy organization. AI tools are designed to find answers, and if campaigns don’t provide them, the tools will find them elsewhere, he said.
“Candidates have to become the authoritative source on themselves,” said Meckler, who’s advising campaigns to publicize their positions on a wider range of issues and in greater depth.
It will take years and several election cycles for a clearer picture to emerge of AI’s economic impact. So far, it’s most directly affecting industries like marketing, customer service and tech, resulting in a loss of about 11,000 jobs per month, according to a recent report by Goldman Sachs. But it’s yet to be seen how widespread job loss could become as AI use accelerates.
For now, a number of politicians in both parties still see benefits in the data center rollout, in part as generator of high-paying blue-collar jobs. AI data center construction has helped add 212,000 jobs since 2022, according to Goldman Sachs. Construction of tech infrastructure helped fuel a 10.4% increase in business spending that lifted first-quarter US growth to a 2% annualized rate.
Republican Representative Byron Donalds, who’s running for Florida governor and received a pledge from pro-AI Leading the Future to spend $5 million on his behalf, derided data center opposition as “performative” during a February event. “We should lead, not play cute on social media, not just say no without planning,” he said. But when it comes to AI safety legislation, Donalds in June split from Leading the Future and Trump, saying he believes that states should take the lead in regulating the AI companies.
Warner, the senator from Virginia, describes himself as “pro-AI” and supports data centers. While he has received money from all of the top AI firms, he said AI companies need to be clear about how they will avoid burdening people. Looking ahead, he said the midterms would be “shaped by AI” with the real inflection point coming in two years.
“This is the warm-up round,” Warner said. “I’ll make a big wager this will be the defining issue of the presidential in 2028.”