Key Takeaways

  • Jason Law funded and released a fully AI-generated political ad that depicts Steve Ehlmann and Bill Eigel.
  • The ad highlights accelerating use of synthetic media in local elections, with Missouri lacking specific regulations.
  • Researchers and election experts warn that deepfakes are outpacing governance frameworks and voter awareness.

With two weeks remaining in the St. Charles County executive race, a fully AI-generated political ad funded by Jason Law has become a flashpoint in a broader national debate over synthetic media in campaigns. The timing, so close to Election Day, injected the issue directly into a local contest already marked by ideological differences between County Executive Steve Ehlmann and Republican primary challenger Bill Eigel.

The ad uses generated likenesses of Ehlmann, Eigel, and Law himself. It landed with tens of thousands of views in under 24 hours, amplifying its impact well beyond typical local race advertising. Law framed the effort as a clever approach rather than a direct attack, although that distinction appears to have landed inconsistently with voters.

Immediate backlash followed the ad's release. Ehlmann rejected the ad outright, arguing that the fictional elements undermine public trust and urging residents not to be swayed by the fabricated video. Eigel called the portrayal dishonest and suggested Missouri may need to consider state-level rules for AI in political ads.

The incident in St. Charles County reflects a broader national governance issue. According to the Brennan Center for Justice, many existing deepfake regulations in the United States focus on disclosure requirements or bans on misleading synthetic media during sensitive voting periods. The challenge is that laws lag behind technology. Missouri and Illinois currently have no AI-specific rules governing political deepfakes even as tools to create them have become inexpensive and accessible.

At the federal level, the Federal Communications Commission has moved to ban AI-generated robocalls and is considering rules for AI-manipulated broadcast political ads. Guidance from the NIST AI Risk Management Framework has also shaped how agencies and campaigns govern generative AI, provenance, and synthetic content. Still, the regulatory patchwork remains uneven and reactive.

The local angle in St. Charles County reflects this broader pattern. A political science researcher at Harris-Stowe State University noted that AI-generated content is already widespread in the current election cycle, appearing in forms as varied as animated characters and synthetic audio. If voters cannot reliably distinguish real from synthetic media, campaigns face concrete risks such as widespread voter manipulation and suppressed turnout.

Researchers tracking political deepfakes have raised similar concerns. A 2025 analysis of the Canadian federal election found that 5.86% of election-related images circulating online were deepfakes, indicating synthetic media is present but not yet dominant. The figure suggests synthetic content can move quickly from experimental to normalized, especially in charged environments like primaries or contested local offices.

Industry analysts note that emerging provenance efforts may eventually help verify digital media. One such effort, supported by various tech companies and media groups, is the Coalition for Content Provenance and Authenticity (C2PA) standard, which focuses on authenticating the origin and edit history of digital content. Other experts, including those writing for MIT Technology Review, emphasize that political deepfakes create information gaps that voters must navigate with heightened skepticism. The gap between technical capability and public understanding shows no sign of shrinking soon.

Even tools designed to contextualize political media, such as fact-checking operations from organizations like PolitiFact or FactCheck.org, can only intervene after content spreads. As platforms like YouTube test political ad detection methods trained to spot AI-generated elements, the arms race between synthetic creation and automated detection continues to evolve.

Public expectations also shift when campaigns normalize AI-driven messaging. If even lighthearted or satirical synthetic content becomes routine, voter fatigue could set in, making it harder to discern manipulation. Regulators and political strategists are openly questioning whether this leads to broad distrust of all political messaging.

For business and technology leaders, the dynamics in St. Charles County offer a view into a near-term environment where synthetic media intersects governance, ethics, and communications. State lawmakers will likely revisit the issue, given that 26 states already have political deepfake laws according to data cited by NBC News and the National Conference of State Legislatures. Local pressure often accelerates these policy debates.

The immediate burden still falls on voters. Experts urge residents to cross-reference information before casting a ballot, a reminder that media literacy remains a frontline defense. The Missouri primary serves as a case study in how rapidly deepfake tools can reshape campaigning norms and voter expectations. The technology continues to advance, requiring legal and educational frameworks that support election integrity to adapt quickly.

Whether the recent ad changes the trajectory of the race remains to be seen. What is already clear is that political AI has arrived in local elections, transitioning from a theoretical concern to a practical governance challenge with immediate implications for trust, regulation, and democratic engagement.