Anthropic updates Claude ahead of US midterms to limit bias
Technology
With the US midterm elections coming up, AI company Anthropic is working to keep its Claude chatbot fair and neutral.
They've rolled out new updates to help stop political bias, flag fake news, especially from foreign sources, and guide users toward official election information instead of rumors.
Anthropic finds 5.75% Claude conversations political
Anthropic checked out 250,000 conversations on Claude and found approximately 5.75% were political.
They use smart tools to spot harmful trends and are teaming up with Democracy Works so Claude can soon answer questions about voting, registration, and polling places by state.
Plus, they've set clear rules: campaigns can use their AI for legitimate stuff but not for misleading tactics, hoping this keeps things honest as elections heat up.