Political campaigns are entering an unfamiliar and uncanny era, and AI may be changing the rules faster than candidates, regulators and voters can keep up. The very important part of politics that once depended heavily on television, polling and broad voter targeting is moving toward something far more personalized.
At Harvard Kennedy School, strategists from both parties described a future in which AI could reshape how campaigns find voters, create political advertising and even figure out what might persuade someone to cast a ballot. The report is based on findings from the Harvard Gazette and a recent report by NPR.
Although, the reports strongly supports a claims about AI, digital advertising and changing campaign strategy in the 2026 midterms,it does not really specifically establish that TikTok and podcasts are replacing traditional TV ads.
The biggest question may no longer be simply what a candidate says. It could be how precisely that message can be designed for each individual voter.
For far too long, political campaigns have relied excessively on television advertising, polling and increasingly sophisticated voter data to reach the public. However, in 2026, the old practices seem to have gone for a toss, as AI is the new normal and it is now opening another door, one that could make political messaging faster, cheaper and far more personalized.
At a Harvard Kennedy School event, political strategists Doug Sosnik and Sara Fagen discussed how AI could transform campaigns, from identifying supporters to producing advertisements and testing messages.
“It’ll be the first AI campaign, so the stakes are going to be unbelievably high,” Sosnik said.
Sosnik, a former White House political director and longtime adviser to Democratic candidates, including President Bill Clinton, compared the coming change with the transformation brought by television during the 1960 presidential election, as per a report by The Harvard Gazette.
The technology is developing quickly, but what could it actually mean for the 2026 midterms and the 2028 presidential race?
Could every voter eventually see a different campaign?
Fagen, a strategic data and technology consultant who served as White House director of political affairs under President George W. Bush, said AI could eventually allow campaigns to tailor messages to individual voters using information about their activities and motivations.
“Eventually, probably not in ’28, but perhaps in the next presidential [election], you’ll hit a button and your target audience of 20 million people will get 20 million different campaigns,” said Fagen.
That possibility builds on a political technology shift that has been developing for more than two decades.
Before 2004, campaigns were largely organized and analyzed around precincts. But Republicans began using early forms of data analytics during George W. Bush’s 2004 presidential campaign against John Kerry to identify likely supporters and voters who might be persuaded.
“We were able to figure out a different way to do this and sort of pinpoint who would be a supporter, who would need to be persuaded. It didn’t matter where they lived; it just mattered based on our algorithm and how to make those predictions,” Fagen said. “And we did that, and we did it very successfully.”
The Obama campaign pushed microtargeting and data analytics further in 2008, making them an important part of its campaign structure.
Now, AI could take that approach considerably further.
What happens to political advertising?
Advertising is yet another area expected to change. Sosnik said campaigns will be able to produce new content more quickly and at lower cost. AI could also end up influencing advertising decisions, including where and when advertisements are placed and how campaigns approach different groups of voters. This paradigm shift is already visible in the 2026 election cycle.
According to the Wesleyan Media Project, at least 164 political advertisements had been generated or enhanced with AI by September 2026, with identified spending approaching $80 million. The project said its figures were an undercount because AI use is not always easy to identify.
The examples are not limited to major statewide contests.
NPR reported on AI-generated political advertising in races such as the North Carolina state Senate contest involving Jessica Bichler. An advertisement showed an AI-generated version of Bichler performing physically impossible yoga movements while making claims about her positions.
Bichler said the use of her face and body had not been something she consented to or participated in creating.
Another example came from Texas, where Republican Senate candidate Ken Paxton released an advertisement featuring an AI-generated version of Democratic opponent James Talarico supposedly reading the Bible. NPR reported that the voice was not Talarico's real voice.
These examples illustrate why AI advertising is becoming such a complicated issue.
Is AI creating a new kind of political advertising?
Not every AI-generated political advertisement is designed to fool voters. Scott Babwah Brennen, director of the Center on Technology Policy at NYU, told NPR that widespread deceptive deepfakes had not emerged in the way some experts feared during the 2024 election cycle.
Instead, many AI advertisements have moved toward satire and parody.
One advertisement depicted a candidate refusing to leave a birthday party, while another showed a Republican candidate driving a school bus filled with migrants.
Brennen described this emerging style as “AI-generated slopaganda.”
Isabel Linzer of the Center for Democracy and Technology compared the development to an updated form of political cartooning. Politicians have long used exaggerated imagery to portray opponents negatively, but AI can now create highly realistic video, making it harder for viewers to know how to interpret what appears on their screens.
Bichler said some of her supporters initially believed the AI-generated yoga advertisement was real.
The Lee campaign later added a larger disclaimer stating, “Warning, the video in this advertisement is artificially generated.”
Could AI change political polling too?
Advertising is only one part of the transformation. Fagen said polling could also change as AI becomes better at processing enormous amounts of information. Pollsters have continued to struggle to reach younger people and adults without college degrees by phone. That can create an incomplete picture of the electorate.
AI could allow polling organizations and news organizations to create “synthetic” groups to test political messages, including among people who are difficult to reach through conventional polling.
That could provide another way to examine what voters might support without depending entirely on traditional polling methods. There is also a possible benefit. Fagen said fact-checking could become faster and draw from a broader range of sources, giving voters more information with which to assess political claims.
But the concerns remain substantial. AI-generated images, deepfakes and highly realistic voice manipulation can put words into someone's mouth that they never actually said. And while lawmakers and courts may attempt to regulate the technology, Fagen said those efforts may struggle to keep pace.
Sosnik agreed. By the time Congress finally takes action on AI, “It’s not going to matter functionally,” he said.
The 2026 midterms elections are therefore becoming part of a much larger experiment in political communication. The technology is already appearing in campaign advertising, while strategists are looking toward a future in which AI could influence targeting, messaging, polling, advertising and voter engagement.
As of now, the shift is still unfolding. But the political campaigns of the next presidential cycle may look very different from those that came before them.
FAQs
What is AI changing in political campaigns?It is changing how campaigns target voters, create advertisements, analyze data and test political messages.
Are AI political ads already being used?
Yes, the Wesleyan Media Project has identified at least 164 AI-generated or AI-enhanced political ads in the 2026 cycle.
