OpenAI has introduced a new feature for ChatGPT called “Computer History.” Designed to learn from what users are already doing on their computers, this optional feature records activity across websites and applications, converting that data into memories and a timeline that the chatbot can later reference. The announcement came on the heels of updates to GPT-5.6 models, which improved accuracy and reasoning adaptability while reducing unnecessary information. With Computer History, OpenAI is moving toward a more personalized AI assistant that can remind users of past activities, recognize work patterns, and even recommend relevant skills or automations.
What Is Computer History?
Computer History is an opt-in feature within the ChatGPT desktop app that uses operating-system activity and accessibility data to log interaction events. Unlike features that rely on continuous screen capture, Computer History tracks specific interactions—such as which apps you used, which websites you visited, and what you were working on at a given moment. This information is then transformed into a structured timeline that ChatGPT can search and retrieve when asked.
For example, a user might ask, “What was I debugging yesterday?” and ChatGPT would provide details from the recorded activity to help recall the context. According to OpenAI, the feature can also detect patterns in how a person works, allowing the assistant to suggest new skills, shortcuts, or automations that might be useful. This positions Computer History as a step toward making ChatGPT a more proactive and context-aware digital companion rather than just a conversational tool.
The data collected is not stored as raw screenshots or audio recordings. OpenAI emphasizes that the feature relies on interaction events derived from accessibility and OS-level activity data. Users have control over which apps and websites can contribute to the interaction history, offering a degree of customization that is intended to ease privacy concerns. The feature is being introduced as a replacement for the Chronicle research preview, which was previously available to select users.
How It Compares to Windows Recall
Many observers have drawn comparisons between Computer History and Microsoft’s Windows Recall, a tool that faced significant backlash when it was first announced. Windows Recall was initially configured to be on by default and took periodic screenshots of the user’s screen, creating a visual timeline of activity. The feature was delayed after security researchers raised concerns about data exposure and the potential for malware to access sensitive information without user consent.
Computer History differs in several important ways. First, it is an opt-in feature, meaning users must explicitly enable it before any data is recorded. Second, OpenAI states that it does not capture screenshots or audio. Instead, it logs interaction events using operating-system activity and accessibility data, which is less invasive than continuous screen recording. Third, users can choose which apps and websites are allowed to contribute to the history, giving them granular control over the data that ChatGPT can access.
However, these differences do not mean that Computer History is entirely free of risk. OpenAI’s own announcement includes a disclaimer that the feature can contain sensitive information. This data is not encrypted specifically by Computer History, and other programs running on the user’s computer could potentially access it. This is a notable concern for users who handle highly confidential material, as the interaction data might be stored in a format that other applications could read.
Privacy Risks and User Control
Privacy experts have long warned about the dangers of AI systems that continuously monitor user behavior. Even when data is collected with the user’s permission, the risk of that data being exposed—through hacking, accidental leaks, or legal requests—remains a significant issue. Computer History’s architecture relies on accessibility data, which is designed to help assistive technologies interact with the operating system. By tapping into this infrastructure, OpenAI can gain deep visibility into user activity without needing to capture the screen.
The fact that Computer History is opt-in helps mitigate some concerns, but it also places the responsibility on users to understand how much data they are sharing. OpenAI provides settings that allow users to review and delete their Computer History data, as well as to control which applications and websites are monitored. This is a step in the right direction, but experts argue that more transparency is needed regarding how long the data is stored, where it is stored, and who can access it beyond OpenAI’s servers.
Another issue is that computer records of user activity can be sensitive, especially for journalists, lawyers, doctors, or people living in countries with restrictive governments. Even with encryption during transmission, the data may be stored on servers in jurisdictions with different privacy laws. OpenAI’s decision to exclude the UK, Switzerland, and the European Economic Area from the feature’s initial rollout likely reflects concerns about compliance with the General Data Protection Regulation (GDPR) and other strict privacy frameworks. Those regions have some of the most robust data-protection laws in the world, and launching a monitoring feature there without thorough legal review would be risky.
Potential Benefits for Productivity
Despite the privacy concerns, Computer History offers tangible benefits for productivity and workflow optimization. With this feature, ChatGPT can act as an external memory for users who regularly juggle multiple projects, meetings, and deadlines. Instead of manually searching through documents or browsing history, users can ask ChatGPT about what they were doing at any point in the past, and the AI can reconstruct the relevant context.
For developers, this could mean asking the assistant to list all the code files they edited in a particular session or to remind them of which terminal commands were run yesterday. For marketers, it could help track which ad campaigns were reviewed during a brainstorming session. For researchers, it might provide a timeline of sources consulted while writing a paper. The feature’s ability to recognize work patterns also opens the door to intelligent automation, such as suggesting that a user create a recurring meeting, set reminders, or organize files based on recurring activities.
OpenAI has framed Computer History as a way to reduce the need for users to explain their current context repeatedly. Instead of telling ChatGPT about a project from scratch, users can rely on the chatbot’s memory to pick up where they left off. This aligns with a broader trend in AI development: moving from reactive tools to proactive assistants that anticipate needs and offer contextually relevant suggestions.
Availability and Regional Restrictions
Computer History is rolling out for desktop Mac users who subscribe to ChatGPT Pro, Business, or Enterprise. These tiers represent the company’s premium offerings, which are aimed at power users, professionals, and organizations that require advanced capabilities and higher usage limits. The feature is not included in the free tier of ChatGPT, nor is it available on Windows or Linux desktops at launch.
OpenAI has also confirmed that the feature will not be available in all regions. The excluded markets include the United Kingdom, Switzerland, and the European Economic Area. This regional restriction means that users in those countries will not be able to access Computer History even if they have a Premium subscription. The decision is likely tied to the legal and regulatory landscape in Europe, where data protection authorities have been aggressive in scrutinizing AI tools that process personal data.
For users who are comfortable with the idea of giving ChatGPT access to their activity, Computer History represents an interesting experiment in AI memory. However, it also underscores the evolving relationship between humans and machines, where the boundary between personal data and artificial intelligence becomes increasingly blurred. As the feature expands to other platforms and regions, it will be important for OpenAI to continue refining its privacy controls and to address the legitimate concerns raised by users and experts alike.
Source:Android Authority News