ChatGPT’s Computer History tracks your clicks and keystrokes

The landscape of AI assistants is shifting from passive tools to active collaborators, and OpenAI is taking a bold step in that direction. According to The Verge, ChatGPT’s desktop app for macOS now includes a feature called Computer History. This is not just a simple log of your actions. It is a system designed to learn how you work by tracking your clicks and keystrokes in real time.

The implications here are profound for anyone using AI in their daily workflow. The feature builds a timeline of your activity that ChatGPT and Codex can reference when you make a new request. This means the AI can pick up tasks you left half done or suggest automations based on patterns it has observed. It is essentially teaching the model your personal work style.

What makes this move particularly interesting is the default setting. The feature is opt-in rather than opt-out. This suggests OpenAI is confident in the value proposition but also aware of the sensitivity involved. Users can exclude specific apps and websites from the tracking. You also have the ability to delete entries if you want finer-grained control over your digital footprint.

Privacy advocates and enterprise security teams should pay close attention to this development. Ari Weinstein, a Product and Engineering manager at OpenAI, clarified on X that the system will automatically ignore content in incognito or private browser tabs. This is a crucial safeguard for sensitive research or confidential browsing sessions.

However, the line between helpful context and invasive surveillance is thin. For professionals handling proprietary code or sensitive client data, the idea of an AI recording every keystroke is a hard sell. Even with exclusions, the potential for data leakage or unintended model training remains a concern. The opt-in nature helps, but it places the burden of vigilance squarely on the user.

This trend reflects a broader industry move toward context-aware AI. Other companies are exploring similar features where assistants learn from your email, calendar, and document history. The goal is to reduce friction and make AI feel more like a natural extension of your mind. But it requires a level of trust that many organizations are still hesitant to give.

For AI enthusiasts, this is a fascinating experiment in human-computer interaction. It raises questions about how much data we are willing to trade for convenience. As these tools become more integrated into our operating systems, the definition of personal data will continue to evolve. We must remain vigilant about what we share and how it is used.

What this means for you If you use ChatGPT on macOS, review the Computer History settings immediately. Decide which apps are safe to include based on your work sensitivity. Try this prompt to test the feature safely: Ask ChatGPT to summarize your recent activity in a specific app to see what it has captured. This helps you understand the scope of data collection before you commit to using it for automation.