🗳️Nearly 2,500 Voters Test ChatGPT, Claude, Grok, Gemini on Election Facts

Your chatbot might be giving voters bad election info

Voters are asking chatbots who to vote for, and a new study says the answers need a closer look. Nearly 2,500 adults in 35 states with Senate races were asked to query ChatGPT, Claude, Grok, or Gemini about the election, and the results show chatbots are quietly becoming part of how people decide. If you build search, content, or civic tools, this matters. Most search products now put an AI layer between users and the info they want, and that layer can add hallucinations and bias to already messy political information. The catch? Users trust the summary because it sounds neutral. The numbers: 42% of US adults use chatbots to search for information, but only 8% used them for election info this year. Of those election users, 76% were fact-checking claims they'd heard, 71% asked where candidates stand on issues, and 36% asked for help choosing a candidate. Positions are the shakiest part, since they depend heavily on which sources the model pulls from. Worth watching: the next step is measuring how accurate and unbiased these answers are, and whether recommendations actually match what users need.