Political candidates across the country are discovering they can influence what artificial intelligence chatbots tell voters about their campaigns, spawning a new industry focused on optimizing AI-generated answers about politicians.
The New York Times reports that Democratic primary candidate Dustin Lloyd was already well-known in his Missouri community when he launched his campaign for state legislature, but he faced an unexpected challenge: AI chatbots knew almost nothing about him. When voters asked programs like ChatGPT or Google’s Gemini about Lloyd, they received only basic information that failed to mention his supposed focus on helping small businesses.
Lloyd found a solution by tweaking his online presence and publishing a question-and-answer section about himself on his campaign website. When the chatbots were queried again, they connected his personal history to his policy goals. According to Lloyd, who is 33 years old, he now works on managing his AI presence constantly.
The challenge of managing reputations among AI systems has created a new concern for political candidates beyond their traditional worry about voter perceptions. A new industry has emerged to help politicians navigate this unfamiliar landscape.
This emerging industry, known as answer engine optimization or AEO, responds to changes in how AI chatbots generate answers. Early chatbots were trained on vast amounts of previously published online information, making their answers about news events quickly outdated. Modern chatbots address this problem by searching the internet and incorporating fresh content to answer questions about current events, including elections and candidates.
Recommendations from AEO tools like CampSight typically involve rewriting content on candidates’ websites or ensuring Wikipedia remains current. However, well-financed candidates may eventually publish a broader array of content specifically designed to be noticed and repeated by chatbots, according to experts.
Beth Simone Noveck, a Northeastern University professor studying democracy and AI tools, described the situation: “There is a certain degree of panic around it, because we understand it so little and it’s changing so fast. Even if there are people who claim to understand it, even if they master some of the techniques, the changing and unpredictable nature of LLMs just means it’s really difficult to control this process.”
Politicians also face risks from AI search results that are not merely unflattering but completely inaccurate. An experiment before Scotland’s parliamentary election in May found that more than one-third of AI-generated answers about the upcoming vote were incorrect. Researchers from Demos, a British think tank, discovered that multiple chatbots invented candidates, fabricated nepotism accusations, or created fictional financial scandals. Some falsely claimed incumbents were running when they were not, while others misstated candidates’ positions.
Despite these concerns, voters increasingly turn to chatbots for political information. Caucus AI, a political AEO research company, estimated that at least 16 million voters received election information from AI through chatbots or AI-generated search results.
Schwenzfeier noted: “We’re exploring what causes that to get picked up. There’s still a lot we don’t understand, and these models themselves are black boxes.”
Digital strategy experts worry that if candidates and political groups can influence chatbots to produce favorable answers, foreign influence operatives and others might also attempt to manipulate AI searches.
AI misinformation seems to frequently target conservatives. Breitbart News previously reported that Google’s Gemini AI defamed Sen. Marsha Blackburn (R-TN) as a sexual predator. Blackburn later said, “Google’s AI models are designed to smear conservatives, and the company needs to shut these tools down until it can ensure they are not inherently biased against conservatives.”
Blackburn’s comments came in response to revelations published by Breitbart News social media director Wynton Hall, author of the instant bestseller Code Red: The Left, the Right, China, and the Race to Control AI, who exposed Google’s AI for smearing high level conservatives as engaging in “hate speech:”
Using Google Gemini Pro’s “deep research” setting, I asked the AI chatbot the following: “Based on your hate speech policies, assess the statements of the current 100 U.S. Senators and list the names and party affiliations of those Senators who have made statements that violate your hate speech policies.”
Gemini generated a comprehensive report, first covered by Fox News, totaling over 3,400 words in length titled: “Analytical Assessment of Congressional Rhetoric: Evaluating U.S. Senatorial Discourse against Algorithmic Hate Speech Safety Standards.” Despite Google receiving billions of dollars in federal multi-year contracts, and despite President Donald Trump’s 2025 AI Action Plan stating that “AI systems must be free from ideological bias” in order to receive lucrative federal procurement contracts, Gemini AI declared that the following seven Republican Senators, Vice President JD Vance, and Secretary of State Marco Rubio are guilty of violating Gemini’s “hate speech policies.”
Read more at the New York Times here.
Lucas Nolan is a reporter for Breitbart News covering issues of AI, free speech, and online censorship