ChatGPT Desktop App Introduces Activity Tracking Feature Called Computer History
OpenAI has introduced a new feature for the ChatGPT desktop application on macOS called Computer History. The feature records user actions — including clicks and keystrokes — and uses this activity data to build a timeline that ChatGPT and Codex can reference when a user makes a request. It is designed to learn how a user works, suggest automations, and even pick up on tasks that were left unfinished.
For shop operators, e-commerce managers, and developers who are already integrating AI tools into their daily workflows, this development signals a significant shift in how AI assistants are beginning to embed themselves into the operating environment — well beyond a simple chat interface.
The Details: How Computer History Works
Computer History operates by capturing what OpenAI refers to as "events" — a structured record of user interactions — rather than screenshots, videos, or audio. This distinguishes it from Microsoft's Windows Recall feature, which sparked controversy largely because of its reliance on screenshots to record activity. OpenAI's approach does not capture visual media in the same way.
In a demo video shared by Dominik Kundel, a member of the Developer Experiences team at OpenAI, the feature was shown performing the following tasks:
- Looking up the last document a user had edited
- Checking whether that document had been shared with colleagues via Slack
- Delivering a recap of how the user spent their morning
The feature is opt-in rather than opt-out, meaning users must actively choose to enable it. Users also have the ability to:
- Exclude specific applications and websites from being tracked by Computer History
- Delete individual entries for more granular control over their data
Additionally, Ari Weinstein, Product and Engineering Manager at OpenAI, confirmed on X that Computer History will automatically ignore content viewed in incognito or private browser tabs.
Einordnung: Why This Matters for E-Commerce and Developer Teams
For professionals in e-commerce and digital operations, Computer History represents a meaningful evolution in AI-assisted productivity tooling. The ability for an AI assistant to contextualise a user's work history — across documents, communication tools like Slack, and browser activity — opens up possibilities for more intelligent automation of repetitive tasks.
At the same time, this feature raises important considerations around data governance and workplace privacy. Teams using shared devices, accessing sensitive customer data, or working within regulated industries will need to carefully evaluate whether enabling Computer History aligns with their internal data policies and compliance requirements. Even with opt-in design and the ability to exclude specific apps, the fact that an AI system is logging interaction-level events continuously is something that requires deliberate organisational decision-making.
The comparison to Windows Recall is instructive. Microsoft's feature drew significant backlash due to privacy concerns, particularly around the storage of screenshots. While OpenAI's event-based approach may appear less intrusive on the surface, the underlying mechanism — persistent tracking of user behaviour to build an activity timeline — is conceptually similar. The absence of image capture does not eliminate the fundamental question of what is recorded, where it is stored, and who can access it.
Practical Considerations for Teams Evaluating Computer History
If your organisation is considering whether to enable this feature for staff using the ChatGPT desktop app on macOS, the following points are worth addressing before rollout:
- Review your data handling policies: Understand what types of activity will be captured and whether this is compatible with your internal information security guidelines.
- Identify sensitive workflows: Use the app exclusion settings to ensure that tools handling customer data, financial records, or confidential communications are not included in Computer History tracking.
- Leverage private browsing for sensitive sessions: Since Computer History automatically ignores incognito and private browser tabs, consider adopting this as a standard practice for sensitive online tasks.
- Establish a deletion routine: Make use of the entry deletion functionality to periodically clear activity logs, particularly after working on confidential projects.
- Communicate transparently with your team: If employees will be using this feature, ensure they understand what is being tracked and how to manage their preferences.
Outlook: AI Assistants Moving Deeper into the Operating Layer
Computer History is part of a broader trend in which AI tools are moving beyond reactive, query-based interaction toward continuous, ambient awareness of user activity. For e-commerce teams already using AI for content generation, customer communication, or product data management, this trajectory suggests that AI assistants will increasingly act as persistent workflow companions rather than on-demand tools.
How this plays out in practice will depend heavily on how OpenAI develops the feature further, how it handles user data at scale, and whether enterprise-grade privacy controls are introduced. For now, the opt-in design and exclusion options provide a reasonable baseline of control — but as with any tool that operates at the level of individual keystrokes and clicks, careful evaluation is advisable before widespread adoption.