Generative AI is already appearing in US midterm campaign ads, including fabricated scenes and words put into candidates’ mouths. Axios says the Wesleyan Media Project counted $137 million spent on 220 AI-generated ads through 11 October, mostly by Republicans.
Watch Desk analysis
What happened
Axios describes examples across several races: ads depicting Jon Ossoff getting beauty treatments, a fabricated family photo involving Abdul El-Sayed, and AI-generated versions of candidates including Josh Shapiro and Zach Dembo. Some ads disclose AI use; others do not. In Michigan, one ad included written and spoken disclosure, while the Georgia ads targeting Ossoff did not. Axios says Georgia law does not require such disclosure.
The legal picture is uneven. Axios says four states prohibit certain uses of AI-generated deepfakes in political ads, 27 require some form of disclosure, and there is no federal law specifically regulating AI-generated political ads.
Why it matters
These ads are not merely campaign graphics with a new production tool. They can show candidates in scenes that never happened or make them appear to say things they did not say. Disclosure rules vary by state, leaving voters to navigate different standards during the same national election cycle.
The reported spending figure also suggests this is already a substantial campaign practice, not a speculative future problem. The examples Axios describes range from obvious satire to fabricated personal scenes and quotations, making a single, neat label for every use difficult.
Our read
The practical question is no longer whether campaigns will use generative AI, but what voters should be told when they do. A disclosure label is a useful start, though it cannot by itself distinguish parody from a misleading fabrication. The rules need to make that distinction without giving political ads a free pass simply because a machine helped make them.
What to watch
- Whether more states adopt disclosure or deepfake restrictions before the midterms.
- Whether campaigns make AI disclosures consistent and prominent across platforms.
- How voters and election officials respond to fabricated scenes and candidate speech.
Discussion spark: Should every AI-generated political ad carry a prominent disclosure, or should the law target only realistic fabrications that could mislead voters?
Sources and evidence
- The AI attack-ad nightmare is coming true (19 October 2026, 09:35 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.