How U.S. political campaigns have used generative AI

Summary

This Marketplace Tech episode has [[MeganMcCartyCorino|Megan McCarty-Carino]] interview Tim Harper of the [[CenterForDemocracyAndTechnology|Center for Democracy and Technology]] about how generative AI appeared in the 2024 U.S. presidential election. The report’s core finding is that the feared wave of spectacular deepfakes did not dominate the cycle, but campaigns did use AI for speed, scale, targeting, writing, data analysis, and strategy work.

The episode’s main wiki contribution is AI Political Campaign Operations: campaign AI risk is not only about fake videos, but also about faster persuasion infrastructure. It also adds AI Political Ad Disclosure Patchwork and AI Election Misinformation Risk, because state-level disclaimer rules vary and future campaigns could use personalized voting misinformation or search poisoning rather than obvious synthetic spectacle.

Key Claims

  • The source frames the 2024 U.S. presidential election as the first major election in the generative-AI era.
  • Tim Harper says the major campaign use of AI was “speed and scale” rather than a flood of visible deepfakes.
  • Campaigns and consultants used AI for faster messaging, email drafting, data analysis, outreach, persuasion, and strategy.
  • Voluntary norms, guardrails, and fear of voter backlash helped limit manipulated-video use in 2024.
  • Harper warns that restraint may weaken if campaigns normalize manipulated media or fear that opponents will use it first.
  • More than half of U.S. states had passed AI political-advertising laws by the time of the episode, but their disclosure size, timing, duration, and scope differ.
  • The lack of a federal standard creates a patchwork that may be easier for large campaigns with legal and operational resources to navigate.
  • Future risks include AI-generated misinformation about the time, place, and manner of voting, especially when public voter-file details make false messages feel personally credible.
  • Harper also flags search poisoning: AI-generated content could push incorrect, outdated, biased, or misleading voting information into search results.
  • Public resilience against AI election threats needs renewal every election cycle, even if the previous cycle avoided the most visible catastrophe.

Key Quotes

“speed and scale” - Harper’s summary of the practical campaign use case.

“time, place, and manner of voting” - the misinformation category Harper treats as especially concrete.

“public resilience” - Harper’s frame for repeated voter education around AI threats.

Connections

Contradictions

  • No direct contradiction found with existing wiki content.
  • The source qualifies AI Information Pollution by showing that election AI risk may be mundane and operational, not only vivid fake evidence: targeted texts, search results, disclaimers, and public education can matter more than viral deepfake spectacle.