Anthropic, an artificial intelligence safety and research company, announces a new feature called Reflect for its Claude AI chatbot. The beta dashboard helps users track their chatbot usage patterns and digital habits. Reflect functions similarly to end-of-year review tools seen on streaming apps. It gives free and paid users insights into how they spend their time interacting with the artificial intelligence model. Users can access the tool from the settings menu on the web and desktop applications, provided they turn on the memory option.
Key takeaways
- Reflect tracks Claude chatbot usage across one, three, six, or 12 months.
- The dashboard excludes incognito chats and connected external applications for privacy.
- The tool prompts users to take breaks and set quiet hours.
- The feature analyzes interactions using the 4D AI Fluency Framework.
The new dashboard categorizes user interactions to show frequent topics and the types of tasks users perform with Claude. Anthropic plans to add a specific metric to track the total time spent using the chatbot soon. Support for tracking usage in Claude Cowork conversations will also launch in the near future.
To develop the wellbeing elements, Anthropic partnered with the MIT Media Lab Advancing Humans with AI program, the Digital Wellness Lab at Boston Children’s Hospital, and the Family Online Safety Institute. The dashboard periodically asks users questions to gauge their reliance on artificial intelligence. It prompts users by asking what tasks they want to keep doing themselves even if Claude could do them faster. Users can also configure quiet hours within the settings and schedule reminders to step away from the screen and take breaks from the tool. The company designed these features to prevent digital fatigue.
Anthropic explicitly designed privacy rules to govern the new feature. The dashboard ignores data from incognito chats, external calendar or email connections, and health applications. For example, if a user asks Claude to summarize a connected email inbox, the Reflect tool will note the summary task but will completely ignore the source emails. The system processes sensitive topics, such as mental health discussions, at a high level without exposing raw conversation details. Anthropic states that the data stays entirely within the user dashboard and the company never uses it for other purposes.
The Reflect feature grades user habits based on the 4D AI Fluency Framework. This system assesses four specific skills. Delegation involves deciding when to hand tasks to the chatbot and knowing platform capabilities. Description covers how clearly a user writes prompts and defines formatting rules. Discernment relates to evaluating the accuracy of the output and checking the logical steps the AI took. Diligence requires taking responsibility for the final result and verifying facts before sharing the output. The dashboard offers practical tips to help users write better prompts and manage their tasks.
FAQs
Q1. What is the Claude Reflect feature?
A1. Reflect is a new dashboard inside the Claude AI app that shows users their usage habits, frequent topics, and interaction patterns over recent months.
Q2. How do I access the Reflect dashboard in Claude?
A2. Open the settings menu in the Claude web or desktop app and look for the Reflect tab. You must enable the memory feature first to generate a report.
Q3. Does Reflect read my personal emails or health data?
A3. No. Anthropic excludes incognito chats, health tracking apps, and files from connected tools like your inbox from the data analysis.
Q4. What is the 4D AI Fluency Framework?
A4. It is a system Anthropic uses to evaluate how well users collaborate with artificial intelligence. The four categories are Delegation, Description, Discernment, and Diligence.
Q5. Who can use the Reflect tool?
A5. The beta version is currently available to Free, Pro, and Max tier users.
