OpenAI has unveiled a significant update to its ChatGPT desktop application for macOS, introducing a feature that monitors how users interact with their machines. This new functionality, dubbed Computer History, aims to bridge the gap between a user’s daily digital tasks and the AI’s ability to assist with them. By observing actions in real-time, the tool creates a contextual record that ChatGPT can use to streamline workflows and automate repetitive processes.
What happened
At its core, Computer History acts as a persistent observer of a user’s inputs. Unlike standard AI interactions that rely solely on what a user types into a chat box, this feature records specific “events,” such as mouse clicks and keystrokes. This data is then used to construct a chronological timeline of activity that the AI can reference.
In a demonstration of the technology, OpenAI staff showcased the app’s ability to summarize a morning’s work. The AI was able to identify which documents had been edited and confirm whether those files had been shared with colleagues via communication platforms like Slack. By analyzing these patterns, Computer History allows ChatGPT and the Codex model to reference previous actions to answer complex queries about past tasks or to pick up work exactly where a user left off.
Context
The announcement has immediately drawn comparisons to Microsoft’s Windows Recall, a controversial feature that faced significant scrutiny for its practice of taking frequent screenshots of a user’s screen. OpenAI has been careful to distinguish its approach from Microsoft’s implementation. According to the company, Computer History does not capture images, video, or audio. Instead, it relies purely on event-based data to understand the user’s intent and progress throughout the day.
To address potential security concerns, OpenAI has implemented several privacy controls. The feature is strictly opt-in, meaning users must manually enable it before any tracking begins. Furthermore, Ari Weinstein, a Product and Engineering manager at OpenAI, confirmed that the tool is designed to automatically ignore activity occurring within incognito or private browser tabs. Users also maintain the ability to exclude specific applications or websites from being tracked and can delete individual entries if they desire more granular control over their data history.
Why it matters
This development signals a shift in the evolution of artificial intelligence from a reactive chatbot to a proactive digital assistant. By integrating deeper into the operating system, ChatGPT is moving toward becoming an “agent” capable of understanding the full context of a user’s professional life. For many, this offers a massive productivity boost, as the AI can handle task recaps and suggest automations without the user needing to provide manual context for every request.
However, the move also highlights the ongoing tension between AI utility and personal privacy. While OpenAI’s event-based tracking is less visually intrusive than taking screenshots, the recording of keystrokes and clicks still represents a significant level of digital surveillance. As AI companies continue to seek more training data and deeper integration into user environments, the balance between helpfulness and data security will remain a central debate in the tech industry. For now, the success of Computer History will likely depend on how much trust macOS users are willing to place in OpenAI’s data handling practices.