Anthropic adds official U.S. voting data to Claude chatbot ahead of 2026 midterms
Anthropic announced on Oct. 16, 2026 that its Claude chatbot will start providing official U.S. voting information and live election results sourced from the nonprofit Democracy Works.
The move comes as the United States gears up for the Nov. 3 midterm elections, giving users in the United States state‑specific answers about where to vote, what identification is required and how to register without leaving the conversation.
"Someone asking where to vote or what ID to bring to their polling place will now get an answer specific to their state, without leaving the conversation," the company said.
Anthropic first partnered with Democracy Works in 2024 to curb AI misuse in political contexts. In April 2026 the firm added a TurboVote banner that appeared when Claude was asked about voter registration, dates or locations. The new integration expands that effort by drawing directly from publicly available data that Democracy Works collects from state and local election sources.
Claude will also display up‑to‑date official vote counts on election night and in the days that follow. Anthropic said the goal is to make Claude a useful tool for political conversations while preventing election‑related abuse.
Analysis of roughly 250,000 U.S. Claude chats from late September showed that politics, public policy or social issues account for about 5.75 % of conversations, while election‑related queries represent 0.51 % and requests for voting‑specific advice only 0.01 %.
The company highlighted that its latest Claude models – Opus 5.5, Sonnet 5.5 and Haiku 5.5 – scored 99 % on an internal political evenhandedness evaluation, the highest to date. Independent reviews noted that the models rarely refuse political requests but sometimes add unsolicited caveats.
New detection tools have been added to spot coordinated misuse, such as multi‑step schemes that appear harmless in isolation. Tests by the Institute for Strategic Dialogue and the McCain Institute found Claude effective at stopping single, clearly harmful requests but less reliable when abuse is broken into smaller steps, prompting the rollout of enhanced classifiers.
Anthropic expects political queries to rise as Election Day nears and will continue testing for evenhandedness and misuse throughout the election cycle. The company said it will keep investing in tools that prevent abuse and maintain transparency as the midterms unfold.
