Voters Use AI Chatbots for 2026 Midterm Advice Across Dozens of Races
Updated
Updated · The New York Times · Jul 4
Voters Use AI Chatbots for 2026 Midterm Advice Across Dozens of Races
2 articles · Updated · The New York Times · Jul 4
Summary
Los Angeles voter Mia Taylor used Anthropic’s Claude to sort through dozens of local contests, asking it which candidates could best block a Republican mayoral contender from advancing.
Claude first refused to tell her how to vote, reflecting guardrails shared by ChatGPT and Gemini, but it did provide voter-guide links, race summaries and strategic-voting analysis when she reframed the request.
Another voter, Robert Siebelink, used Claude to narrow a governor choice, with the chatbot posing a tiebreaker question and mapping candidates to different governing styles rather than issuing a direct endorsement.
The 2026 midterms appear to be the first U.S. elections in which voters are turning to AI in meaningful numbers to filter candidates, compare races and make last-minute ballot decisions.
Can AI truly empower voters, or does it risk becoming a subtle new tool for electoral manipulation?
As AI becomes a voter's guide, who is liable when it gives biased or false election information?
With deepfakes threatening elections, how can we safeguard democratic trust before it's too late?
The 2026 U.S. Elections in the Age of AI: Voter Engagement, Campaign Tactics, and Regulatory Battles
Overview
The 2026 U.S. midterm elections are being transformed by the widespread use of artificial intelligence, especially AI chatbots. Voters now rely on these tools for political information and guidance, changing how they access and understand election topics. At the same time, campaigns use AI to shape strategies and create content, making political operations more efficient but also raising concerns about accuracy and manipulation. While AI offers new ways for citizens and political entities to engage, it also introduces challenges, such as maintaining reliable information and addressing the risks of bias and misinformation.