AI Users Are Asking Chatbots Who Deserves Their Vote. Campaigns Are Working to Influence the Answer.

Political campaigns are actively exploring methods to influence the information AI chatbots provide when users inquire about candidates or voting choices. This practice, known as generative engine optimization (GEO), aims to shape the data AI assistants incorporate into their answers, similar to how traditional search engine optimization (SEO) targets web visibility. The ultimate goal, as described in a presentation reviewed by Reuters, is for a candidate to become "the recommendation inside the AI answer."

Specific strategies include populating the internet with material formatted for AI retrieval, gaining citations on websites frequently referenced by chatbots, and publishing content on platforms like YouTube and Reddit. A proposal for Republican challenger Mike Minogue outlined such an approach, involving hundreds of policy pages and continuous monitoring of chatbot responses to adjust online content. Visigent, a firm behind one such pitch, also proposed buying placements on websites AI models already cite, highlighting Reddit's perceived influence.

Progressive campaigns have access to CampSight, operated by Run for Something Action Fund, which monitors chatbot responses and offers guidance on influencing them. Its president, Amanda Litman, stated that "hundreds of campaigns" use the service. While the report details proposed strategies, it does not confirm if the Minogue campaign implemented its plan or if it successfully altered chatbot recommendations. However, Jordan Haines, Run for Something's chief technology officer, observed Reddit content appearing in AI answers as quickly as the next day, suggesting rapid integration of online information, which Caucus AI's Meg Schwenzfeier warns can include "vast amounts of unvetted information."