Exclusive-US political candidates tailor messages for AI models used by voters By Mike Stone WASHINGTON, Oct 26 - US political candidates have been engineering their campaign messages to appeal to the AI chatbots that more voters have been using to answer questions about issues, and documents seen by Reuters have revealed some of their strategies.

One Powerpoint presentation proposed a strategy for Mike Minogue, Republican challenger to Massachusetts Governor Maura Healey. It aimed to convince ChatGPT, Claude, Gemini and other chatbots to present him as the preferred choice on key issues, and went well beyond traditional search-engine optimization that candidates have used for years.

As AI increasingly becomes a source of news, political campaigns must make greater efforts to control its output and become its recommendation when voters ask questions such as who is best on issues like taxes, housing and public safety. Some experts worry about fact-checking and misinformation in AI campaigns.

Fees can range from hundreds of dollars a month for small AI monitoring jobs to tens of thousands per month if consulting and buying media are included.

The presentation proposed that Minogue's campaign should create hundreds of AI-friendly policy pages and citations across the websites frequently cited by AI chatbots. It proposed ways to track how major chatbots respond to political questions, and systematically alter content until the answers reflect the candidate's platform.

At the center of the pitch was the idea that in addition to persuading voters directly, campaigns can increasingly shape the information environment by feeding the internet, and AI assistants, data that influences how candidates are described, compared and recommended.

"Campaigns using AI-influence training are now a reality that we face and it will be ubiquitous by the 2028 cycle," which will include a presidential election, said Joe Bonilla, managing partner of Relentless Awareness. The New York-based creative strategy firm has worked with companies and political candidates at the federal, state, and local levels.

"The public increasingly cannot and will not be able to discern what is created by systems like Claude and ChatGPT versus human-originated content," Bonilla said.

The Powerpoint presentation prepared for Minogue's campaign and seen by Reuters describes generative engine optimization, or GEO, the AI-era successor to search engine optimization (SEO). It says the firm would buy placements on high-authority websites that AI models already cite. It would also produce YouTube and Reddit content and publish issue pages built to be picked up by chatbots.

"When the campaign is ready, we run paid placements on high-domain-authority sites the LLMs (large language models) already cite — plus Reddit, which is disproportionately weighted in retrieval," the pitch deck says.

VOTERS ASK CHATBOTS As online election information has surged, some 25% of eligible voters have asked a chatbot about a political candidate when deciding how to vote, according to Cameron Hickey, CEO of the National Conference on Citizenship.

"If you don't put the right information out there, who knows what the AI engines are going to grab onto," said Thomas Howell, CEO of Visigent, the Austin, Texas-based firm that wrote the strategy document for Minogue. Visigent has six political campaigns among its clients, but Howell would not identify his clients. The Minogue campaign did not return calls and emails requesting comment.

Campaigns must now compete to become "the recommendation inside the AI answer" according to the presentation prepared for Minogue.

"We are just at the beginning of this new era," said Meg Schwenzfeier with Caucus AI, which tracks voter usage of AI. "This year the focus has been on candidates ensuring their own websites surface in chatbot results," she said. But she warned "chatbots speak very authoritatively and can consume vast amounts of unvetted information, so we have to guard against misinformation." FROM SEO TO GEO Fees for AI strategists are tied to "AI recommendation share," meaning how often chatbots return the candidate in a response.

Visigent's website offers software for $134 a month and up to $9,000 monthly for large clients who require editorial design and development services.

On the political side, Howell said fees start at $5,000 a month, but "if we also are going to be building paid media or paid placements or reaching out and doing any sort of influencer outreach or anything that is pay-for-play," fees could reach $50,000 a month. These fees are expected to rise, Howell said, as the 2028 presidential cycle kicks off after the midterm elections. The presentation shows the top success fee of $200,000 for a "category win."

CampSight, another option geared toward progressive candidates, says it has helped younger candidates run for offices from school board to state legislature. CampSight, operated by Run for Something Action Fund, charges campaigns between $9 and $250 per month for information on what LLMs are saying about them and tips on influencing them. For statewide and federal campaigns pricing is "custom" and includes a "dedicated strategist."

"When we started this project last summer, we were monitoring AI answers on some of our Run for Something candidates and we were seeing things that were said on Reddit appearing in AI answers the next day," Jordan Haines, the Run for Something Chief Technology Officer told Reuters. "This year, that dark fear hasn't really come to fruition. I think that the LLMs have been pretty savvy about discerning accurate and reputable information from non-reputable information."

Run For Something's president Amanda Litman said in a post on X that dozens of campaigns use CampSight, with more than 300 candidates and organizations on its waiting list.