Key facts
- Anthropic will integrate official US voting information from Democracy Works into its Claude AI chatbot.
- The AI company plans to display live election results on election night.
- Anthropic tests its models for political bias before every launch.
Anthropic announced on Wednesday that its Claude AI chatbot will soon feature official US voting information sourced from the nonprofit Democracy Works. This update is slated for release later this month, ahead of the November midterm elections. The company also plans to integrate live election results into Claude.ai on election night and in the days following.
Anthropic has been working to enhance election safeguards for its AI models. The company stated that its latest models, Claude Opus 5.5, Sonnet 5.5, and Haiku 5.5, achieved a 99% score on its political even-handedness evaluation. Additionally, three external groups have reviewed Claude's performance regarding free expression, with methodologies focusing on contentious topics, matched political prompts, and user surveys.
An analysis of 250,000 US Claude.ai consumer conversations from September 21-27, 2026, revealed that about 5.75% touched on politics, public policy, or social issues, with 0.51% specifically concerning elections or voting. Requests asking Claude for voting recommendations were rare, appearing in about 0.01% of conversations, while policy-violating requests were found in approximately 0.002%.