Anthropic plans to add an AI fluency score feature to Claude

Tech media TestingCatalog reports that Anthropic plans to add a personalized 'AI Fluency' score to Claude settings, measuring how well users use AI and giving tips to improve. It's based on Anthropic's AI Fluency Index study from February this year.

The study found users give clearer instructions when asking AI to write code or documents, but they question AI outputs less, missing wrong facts or weak arguments. Iterating and refining is the strongest sign of good AI use.

Now, Anthropic is turning this research into a feedback feature inside Claude. The scorecard scans your activity in Chat, CoWork, and Claude Code, grading 11 behaviors like goal clarity, communication style, and quality control.

It also tracks feature usage over the last 30 days, like 30 project chats, 27 web searches. The system combines behavior and feature use to suggest improvements, showing a score like 7.5/11 to help users see what works and what needs work.