AI Ad in North Carolina Race Shows Wider Use in 2026 Campaigns
Republican North Carolina state Sen. Michael Lee ran a campaign ad that shows an AI-generated version of Jessica Bichler twisting her body into odd, unrealistic poses.
Bichler owns a yoga studio. She is running against Lee in a state Senate district in and around Wilmington in the fall 2026 election. The ad paired those synthetic images with statements about her stances on policing and health care. It included a disclaimer saying AI was used. NPR
Lee said voters will recognize the spot for what it is. He told NPR that "voters in his district are smart enough to spot comedy and obvious parody when they see it." NPR described the ad in its Sept. 16 report, "As campaign season gears up, AI-generated ads are everywhere," on AI in political advertising during the 2026 midterms.
The North Carolina race is one of many. The Wesleyan Media Project, which tracks political ad spending, reported at least 164 political ads nationwide so far in 2026 were produced or enhanced by AI. Michael Franz is co-director of the project. NPR
Those ads span different offices and markets. AI-generated political ads have been used in Spencer Pratt's run for mayor of Los Angeles and in the Texas Senate race between James Talarico and Ken Paxton.
Disclosure rules differ by state, and ad buyers and campaign lawyers must follow them. According to the National Conference of State Legislatures, 31 states have passed laws that prohibit deepfakes, which are fake images, video or audio that look real, or require disclosures, which are on-ad notices that AI was used. At the federal level, no regulation constrained the use of AI in political messaging in the U.S. as of March 2026. Reuters On Sept. 12, 2023, two Democratic and two Republican U.S. senators introduced legislation to ban deceptive AI-generated ads in U.S. elections. It was not enacted. Reuters
Counts and spending have risen. NBC News reported on March 22 that at least 15 campaign ads featuring AI-generated content had run since November. By May, the Wesleyan Media Project identified at least $19.7 million in spending on ads known to be generated or enhanced through AI, out of $210 million in political ad spending over a two-week period in 2026. Wesleyan Media Project The September count of 164 ads reflects continued expansion through the midterm cycle.
Scott Babwah Brennen is director of the Center on Technology Policy at New York University. His center studies synthetic media, meaning images or sound made by computers, how it spreads on platforms, and how election offices handle it.
The broader context here is operational, not theoretical. For campaign staff, the Lee ad tests two ideas at once. The first is that disclosure plus clear parody satisfies legal risk and that voters will understand it. The second is that AI tools can produce broadcast-quality ads at low cost for down-ballot races like state Senate. Neither idea has a settled answer under state-by-state disclosure rules, and with no federal action, enforcement stays fragmented.
For tracking, the broader challenge is finding every ad. Wesleyan's figures cover only ads known to be generated or enhanced through AI. Undisclosed or lightly edited synthetic content is harder to classify. Counts are likely to revise upward as detection improves and as disclosure filings catch up to what aired.
