Anthropic turns AI Fluency research into Claude feature that scores users on 11 collaboration behaviors

Anthropic is introducing an AI Fluency scorecard accessible within Claude's settings. The system analyzes user activity across Chat, Cowork, and Claude Code sessions against 11 behavioral indicators and produces a structured report, displayed as a fraction (e.g., 7.5 out of 11) alongside guidance on areas to strengthen.

The 11 indicators map to Anthropic's 4D AI Fluency Framework and cover delegation (clarifying goals, consulting on approach), description (defining audience, specifying format, communicating tone, iterating, providing examples, setting interaction style), and discernment (checking facts, noticing reasoning, recognizing context).

The scorecard builds on the AI Fluency Index Anthropic published in February 2026, which analyzed around 9,830 anonymized Claude conversations and found iteration and refinement to be the strongest predictor of good AI use. The product feature brings that research finding into a personal feedback loop intended to nudge users toward behaviors Anthropic believes lead to safer outcomes.