Artificial-intelligence-generated political ads have become a regular feature of the 2026 U.S. Midterm campaign, appearing on television and social media as candidates and political groups use synthetic video to attack opponents or reshape their public image. The material ranges from realistic impersonations of politicians to visibly altered scenes and AI-enhanced footage that may not look artificial to viewers.
One Texas Senate ad created by the National Republican Senatorial Committee showed Democratic candidate James Talarico appearing to deliver inflammatory remarks he never recorded. The video included a small “AI generated” label. Mike Collins defended its use of new tactics, while Ossoff’s campaign said it would not use deepfakes that fabricate an opponent’s words or actions to mislead voters.
The technology has also entered lower-profile races. In Virginia, Republican officials in Loudoun County released AI-generated attacks depicting Democratic Gov. Abigail Spanberger making statements she did not make. In Texas, an ad in the Republican attorney general primary showed an artificial version of Sen. Cornyn’s campaign answered with its own AI-generated attack on Attorney General Ken Paxton.
Tracking efforts show the scale of the shift, although estimates differ by method and timing. A separate count through mid-October found more than $137 million spent on at least 220 AI-generated ads, compared with less than $50,000 identified in the 2024 cycle.
Federal rules do not specifically regulate AI-generated political messaging, leaving states with a patchwork of laws that mostly require disclosures rather than banning deceptive content.
Security specialists have warned that even viewers who suspect an ad is fake may still remember its imagery, while election officials and campaigns face a widening gap between the speed of synthetic content and the uneven rules governing its disclosure.