Anthropic AI Fluency Scorecard Spotted in Claude Settings

AI Fluency scorecard spotted in Claude settings — not yet officially launched

References to the feature were found inside Claude's settings UI. Users would be able to open a dedicated screen, trigger Claude to generate a personal scorecard, and view results from within the settings panel. This is a spotted-in-settings discovery, not an official product announcement.

Scorecard scans Chat, Cowork, and Claude Code sessions against eleven behavioral indicators

The system is designed to analyze activity across all three Claude session types, score each against a defined behavioral rubric, and produce a structured report once analysis completes. Results appear to be presented as a fraction — e.g., 7.5 out of 11 — alongside targeted improvement guidance.

Scoring maps to the 4D AI Fluency Framework built with academics Rick Dakan and Joseph Feller

The eleven behaviors are grouped around competencies from Anthropic's 4D AI Fluency Framework. At least three of the four dimensions are identifiable in the source material — delegation, description, and discernment — covering themes like setting goals, framing conversations, and applying quality control. The fourth dimension is not fully named in available reporting.

Anthropic's February 2026 research found iteration and refinement is the strongest predictor of effective AI use

Anthropic's research analyzed approximately 9,830 anonymized Claude conversations to baseline how people collaborate with AI. The study also found that polished outputs like artifacts and code tended to lower critical checking behavior among users — a key behavioral insight the scorecard appears designed to address.

Feature converts a one-time research baseline into an ongoing personal feedback loop for Claude users

By embedding the same scoring system used in the February study directly into the product, Anthropic would give individual users a concrete, repeatable view of their AI habits — nudging behavior toward patterns the company associates with safer and more effective AI collaboration.

Scorecard targets newcomers with concrete habit benchmarking, not just power users

The fraction-based score format and accompanying guidance on which areas to strengthen suggest the feature is designed to be accessible to users still developing their AI collaboration habits, giving them a structured entry point rather than abstract advice.