MIT Tracks How AI Chatbots Answer 2026 Midterm Voting Questions

As the 2026 U.S. midterm elections approach, MIT professors Adam Berinsky and Chara Podimata are launching a project to monitor how AI chatbots respond to voter questions. The initiative aims to shed light on the accuracy and neutrality of AI-generated election information, a growing concern as more Americans turn to tools like ChatGPT for guidance. Berinsky, a political scientist, and Podimata, a computer scientist, will track chatbot responses over time, analyzing whether they provide factual, unbiased answers or inadvertently mislead voters. Their work addresses the critical intersection of AI and democracy, where misinformation could influence electoral outcomes. With millions relying on AI for news, understanding these systems' limitations is essential. The project will offer real-time insights into how tech platforms shape political discourse, potentially informing regulatory efforts. Ultimately, it underscores the need for transparency as AI becomes a key player in civic life.