MIT team launches LLM Election Observatory to track how AI chatbots answer questions about the 2026 midterm elections in real time.

The team retests each model with the same set of fixed questions every few days to monitor how answers shift as the election approaches, depending on user location, political affiliation, and identity. The project aims to systematically document this information environment and make data available to other researchers and the public. Its focus extends beyond factual accuracy to include differences in framing, emphasis, uncertainty, and cited sources.