A new way to reflect on how you use Claude
Anthropic
Anthropic added a new dashboard that shows you how you actually use Claude, not just what it did for you. It's basically a mirror for your AI habits, quietly asking if the time you spend is time well spent.
Based on reporting by Anthropic — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Anthropic just shipped something a little unusual for an AI company: a feature that asks you to think harder about whether you should be using its own product. Called reflection, it's live in beta on Claude for web and desktop, tucked into Settings, and it gives users a dashboard breaking down their chat activity over the last 1, 3, 6, or 12 months.
The pitch came out of user interviews, where Anthropic kept hearing the same uncertainty: people don't really know how much AI use is appropriate, or when a task is better handled by a human. So the dashboard surfaces patterns, what topics dominate your chats, when you use Claude most, what kinds of tasks you lean on it for, and periodically throws out prompts like 'What's one thing you want to keep doing yourself, even if Claude could do it faster?' There's also a practical layer: users can set quiet hours or schedule break nudges, though both are just soft reminders you can dismiss whenever you want.
Anthropic is framing this around what it calls the 4D AI Fluency Framework, four skills, delegation, description, discernment, and diligence, meant to sharpen how someone collaborates with AI rather than just outsourcing to it. The report will apparently point out habits, like a tendency to rewrite email drafts in your own voice before sending, or to lock in your own strategy before delegating execution. It'll also nudge you toward workflow fixes, like starting a Project instead of re-explaining context every session.
On privacy, Anthropic says incognito chats and files from connected tools stay out of the reflection entirely, and anything tied to a health integration is excluded outright. Summaries of, say, an inbox might show up, but the underlying emails won't. The company also says it consulted outside groups, MIT Media Lab's AHA program, Boston Children's Digital Wellness Lab, and the Family Online Safety Institute, while building the sensitive-topics handling, which keeps personal conversations visible only at a high level rather than in detail.
Right now it's limited to Free, Pro, and Max users who have Memory turned on, and Cowork conversations aren't included yet. Anthropic says a time-tracking view is coming soon too.
My take — AI-written commentary, not fact-checked reporting
I'll believe the wellbeing framing when a company actually ships a feature that meaningfully reduces usage rather than just gently gesturing at moderation while collecting more data on how you talk to their chatbot. Dismissible nudges and quiet hours are nice UX, but let's be honest, this is also a tidy way to make you feel thoughtful about a habit Anthropic has every commercial incentive to deepen. Still, compared to the industry's usual approach of engagement-maxing by default, even a token gesture toward self-reflection is a mild improvement, and I'd rather see more labs build this kind of thing than none.
Read more about this at: Anthropic