WASHINGTON — Political campaigns are increasingly turning to free artificial-intelligence tools to produce ads that elevate their candidates and attack opponents, fueling concerns among election analysts and lawmakers that the wave of synthetic content will undermine the 2026 midterm elections, researchers and officials said.
The number of deepfakes shared on social media grew from about 500,000 in 2023 to roughly eight million last year, according to DeepStrike, a cybersecurity firm. About 85% of Americans say political content created by AI will spread misleading information about the midterms, according to a poll from NPR, PBS News and Marist University.
Researchers tracking the trend say Republicans are using AI-generated content more frequently, with President Trump among the most prolific posters of such material, according to the Wall Street Journal. Some posts have depicted Trump as a Christ-like figure. One video showed former President Barack Obama and Michelle Obama as apes, drawing backlash; Trump later took down that post.
Top Democrats, including California Gov. Gavin Newsom, also are using the technology, researchers and campaign officials said.
“Through engaging posts and banger memes, we are successfully communicating the President’s extremely popular agenda,” White House spokeswoman Abigail Jackson said in a statement. “There’s a reason so many people try to copy our style—our message resonates.”
In recent weeks, an AI-generated video portrayed Texas Senate candidate James Talarico, a Democrat, wearing a dress and apron similar to the costume worn by Julie Andrews in “The Sound of Music” while singing about transgender children. The ad was purchased by a political group aligned with Trump.
The National Republican Senatorial Committee separately targeted Talarico with an AI-generated video that featured a fake version of the candidate reading provocative old social-media posts from his past, with a deep voice repeating phrases like “white men” and “pronouns” for emphasis. The video included a small disclaimer in the corner labeling it as AI-generated. The NRSC produced a similar ad targeting Maine Democratic Senate candidate Graham Platner.
“If Graham Platner didn’t want rural Mainers to know he called them racist and stupid, and James Talarico didn’t want white men to know he called them the ‘greatest domestic terrorist threat in our country,’ they shouldn’t have said it,” said Bernadette Breslin, an NRSC spokeswoman.
Republicans have also been targets of AI-enhanced content. Mike Rogers, a GOP Senate candidate in Michigan, went viral after a photo posted on X showed him with bulging muscles at a parade. The image, which included a label noting it was made with AI, spawned additional AI-generated images and videos of a buff Rogers.
An animated raccoon appeared in a video by Alabama Republican lieutenant governor hopeful Wes Allen ahead of a primary runoff, making allegations about his opponent John Wahl, including that the raccoon lived with Wahl. The Allen campaign said the video was based on a public Instagram page connected to Wahl that showed a raccoon inside a house and on a bed. Wahl said the raccoon was his fiancée’s before they started dating and that he has never owned a raccoon.
“This is one of the concerns with AI because it makes it easier for political candidates to lie and spread false information,” Wahl said.
Minnesota Lt. Gov. Peggy Flanagan, a Democrat, was depicted in an AI-generated ad smiling and extending her hand while accepting campaign contributions from big companies. The ad was paid for by a super PAC supporting her Senate primary opponent, Rep. Angie Craig.
“It just feels creepy to see this image of you on a screen that looks like you but isn’t exactly you,” Flanagan said.
Minnesota is among about 30 states with a law prohibiting election-related deepfakes, but digital content researchers said enforcing such rules is difficult because of First Amendment protections. Flanagan said her team is evaluating whether the ad violates the law.
A Craig campaign spokesman said Craig does not support the use of AI in political ads.
University of Virginia law professor Danielle Citron, who studies the technology, described the current environment as “a perfect storm.” She said the combination of improved AI tools, weaker content moderation, and consumers’ tendency to believe deepfakes that reinforce their existing views is deepening political polarization.
A recent rollback in content moderation on social-media platforms including Instagram is exacerbating the trend, researchers said. Calling out AI-generated deception often ends up amplifying the original content, they added.
Andrew Jones, chief product officer of cybersecurity firm Adaptive Security, said that “what’s in production right now is already scary enough for a lot of election-related misinformation.”
OpenAI recently said it would work with the Associated Press and the nonprofit Democracy Works to provide election information to users.