Anthropic Shows Claude Users How Dependent They've Become

On July 9, Anthropic launched Reflect, a panel inside Claude that gives users a detailed picture of how and when they use the AI assistant. The company presents it as a tool for building healthy habits, but tech journalists have already pointed out that the dashboard could just as easily work as a subtle nudge toward using Claude even more.

The dashboard is accessed from Settings in Claude's web version and desktop app. Once enabled, it shows a summary of the topics a user has engaged with most often, the times of day they turn to the assistant, and the types of tasks they hand off to it, from editing emails to planning strategy. A counter tracking time spent with Claude is expected to be added in the coming weeks.

Four pillars of reflection

At the core of the feature is the so-called 4D AI Fluency Framework, four skills Anthropic wants to promote among users: delegation, deciding when it's actually worth involving AI at all; description, formulating precise prompts; evaluation, taking a critical approach to what Claude produces; and accountability for the final result of working with the model.

The panel occasionally asks questions along the lines of what a user would still want to do themselves, even if Claude could do it faster, and lets them discuss the answer directly with the model. The idea is to nudge people toward consciously setting boundaries on their AI use, rather than delegating everything without a second thought.

Privacy and data scope

Anthropic says Reflect does not draw on conversations held in incognito mode, does not analyze files pulled from connected external tools, and entirely excludes conversations linked to health integrations. The company also states that sensitive topics appear in summaries only at a high level of generality, without detail, and that the panel's data is not used for any purpose other than displaying it to the user.

A critical view

TechCrunch pointed out that despite its stated wellness goal, Reflect effectively shows users how much of their daily work Claude has already taken over, which in itself builds a sense of dependence on the product. The outlet compared it to the strategy Google used in 2012 with Gmail Meter, a tool that visualized Gmail's central role in users' digital lives.

The feature subtly reinforces just how much your daily work now depends on Anthropic's chatbot - TechCrunch

The dashboard also suggests specific features, such as Projects, meant to weave Claude even more tightly into a user's daily workflow. Critics note that this is a classic mechanism for building platform loyalty and raising switching costs to competitors, regardless of whether Anthropic's intent was purely the user's benefit.

What it means for users in Poland

For Polish companies and freelancers using Claude at work, Reflect could be a practical tool for auditing how many tasks are already being handled by AI and where it's worth keeping direct control, especially for tasks carrying legal or substantive responsibility. At the same time, it's worth remembering that showing usage statistics isn't neutral in itself, and may push people toward using the tool even more intensively rather than cutting back.

The feature arrives as Anthropic is also running a broader image campaign called Hard Questions, through which the company is collecting tough questions from users about AI's impact on the labor market and society, pledging to publicly report on its response. Both moves fit a strategy of presenting Anthropic as a transparent company that is aware of the risks tied to its own product.

For now, Reflect only works in beta on Claude web and the desktop app; support for conversations in Cowork mode is expected to be added in the coming weeks. Anthropic has not yet said when the feature will leave beta or whether it will also reach customers using Claude through the business API.

Sources: Anthropic Newsroom (anthropic.com), TechCrunch (techcrunch.com), MacRumors (macrumors.com)