MIT researchers launch dashboard to track AI chatbot political bias

As voters increasingly turn to artificial intelligence for political information, researchers at the Massachusetts Institute of Technology have built a new tool called the LLM Election Observatory. The dashboard is designed to monitor how popular AI chatbots answer questions about the upcoming 2026 U.S. midterm elections. Led by a team of political science and computer science experts, the project aims to simulate how citizens nationwide interact with automated systems.

Tracking Changing Responses Over Time

The research team regularly feeds a fixed set of questions into various AI models. These queries cover topics such as election certification, result predictions, and major political issues. By running these tests every few days, the team can observe how chatbot answers shift as election day approaches. The system also accounts for variables like user location and stated political affiliation to see how personalized the outputs become.

Invisible Information Environments

Unlike traditional media or social platforms where public posts are easily viewed, interactions with AI chatbots remain largely private. Two people asking the exact same question might receive different answers based on background details or model updates. The MIT project seeks to make these hidden differences visible by documenting the information environment and sharing the data with the public and other researchers.