Anthropic launches the "AI Mirror" function: let your AI assistant write a "use physical examination report" for you Anthropic officially announced the launch of the beta version of the function Reflect, which turns Claude into a "mirror" to help users review and review their AI usage habits

Anthropic launches the "AI Mirror" function: let your AI assistant write a "use physical examination report" for you Anthropic officially announced the launch of the beta version of the function Reflect, which turns Claude into a "mirror" to help users review and review their AI usage habits. This feature is open to users of the free version, Pro version and Max version who have turned on the memory function. The entrance is located in the settings page of the web and desktop applications. Users can choose to review chat activities in the past 1 month, 3 months, 6 months or 12 months, and the system will automatically generate a personalized usage summary. The report content covers the distribution of key topics, high-frequency usage patterns, common task types, as well as the most active time periods for users and the corresponding work content. Anthropic also revealed that a visual view of usage time will be added in the future, and it will support setting silent periods and usage time reminders. The four-dimensional "physical examination", from using AI to being able to use AI, is the core design of Reflect, which is to conduct a comprehensive assessment of users' AI collaboration habits around four dimensions. The four dimensions are task delegation, goal description, result identification and responsibility prudence. Each item will give specific collaboration examples and targeted improvement suggestions. This is equivalent to issuing an "AI usage physical examination report" to each user, telling you which tasks were handed over correctly, which instructions were written vaguely, and which results should be read more before believing. The intention of Anthropic is very clear: the improvement of AI capabilities is not just a matter of the model, but the user's own "question literacy" also determines the final effect. While most people only focus on what AI can do, Anthropic chooses to instead focus on how well users are doing. This is a rare "introverted perspective" in the current AI industry. The privacy red line is clearly drawn, and health data is completely excluded. In terms of privacy protection, Anthropic has drawn a clear red line. Reflect does not read private chats or extract underlying files in connected tools. For example, when a user asks Claude to summarize their inbox, the report may mention the behavioral pattern of "users often summarize emails" but will never include the original email content. Additionally, all conversations related to health integration tools were completely excluded from the analysis. From the perspective of product logic, Reflect tries to find a balance between "helping users reflect" and "protecting user privacy". As AI assistants become more and more deeply embedded in personal workflows, this kind of boundary design that "can look in the mirror without peeking at your diary" may become the new standard for AI privacy experience. via AI News (author: AI Base)