AI-Generated Political Ads in the US Midterms: What the Research Shows

With the US midterms just a week away, AI-generated political advertising has become one of the election's defining issues. An interview with Erika Franklin Fowler of the Wesleyan Media Project explains what her team's tracking of hundreds of ads reveals about how the technology is being used, who is behind it, and how it is being disclosed to voters.

■ Scale and Sources of AI Ad Spending
- The scope of the research and what the reported spending figure represents
- Which parties and races are driving AI-generated ad volume
- Whether the use is evenly split across the political divide

■ Types of AI Use in Ads
- AI use that closely resembles traditional advertising
- Hyper-realistic candidate deepfakes vs. clearly satirical or parody uses
- Creative examples of AI video used to make a political point

■ Disclosure and Regulation
- How often AI-generated content is labeled for viewers
- The absence of federal rules and the resulting state-by-state patchwork
- Research still underway on the effectiveness of AI ads, including satirical approaches

■ Effectiveness and Impact
- What advertising can and cannot shift at the margin
- Under what conditions an individual ad could affect an outcome

Viewers interested in election integrity, campaign strategy, political communication, or the real-world applications of AI video will gain a grounded understanding of how AI is reshaping political advertising, what distinguishes benign from concerning uses, and what disclosure gaps remain. Those planning to evaluate political ads ahead of elections may find it useful to check how AI content is labeled and sourced before drawing conclusions.