US Political Campaigns Use AI Optimization to Influence Voters

U.S. political candidates are adopting "generative engine optimization" (GEO) to influence the responses of AI chatbots such as ChatGPT, Claude, and Gemini. Strategy documents reveal that campaigns are engineering content to ensure candidates are recommended as the preferred choice on key issues like taxes and public safety. This process involves creating AI-friendly policy pages and purchasing placements on high-authority platforms, including Reddit and YouTube, which are frequently cited by large language models.

Visigent, an Austin-based consulting firm, and the Run for Something Action Fund are among the entities offering these services. Visigent proposed a GEO strategy for Republican candidate Mike Minogue in Massachusetts, with consulting packages costing up to $50,000 per month and potential success fees of $200,000 for a "category win."

This shift is driven by increasing voter reliance on AI, with the National Conference on Citizenship estimating that 25% of eligible voters have used chatbots for candidate research, though Pew Research Center reports a lower figure of 8%. In response, AI developers have implemented safeguards. OpenAI integrated Democracy Works voting information, Anthropic connected users to official sources, and Google added pop-up messages directing users to verify election data.

Critics warn that the authoritative nature of chatbots could facilitate the spread of unvetted misinformation or "poisoned" search results. Conversely, some strategists argue these tactics are necessary to ensure AI engines retrieve accurate information about their candidates.