OpenAI Introduces Desktop Activity Tracking for ChatGPT Users
OpenAI has launched a desktop feature that logs user activity, raising significant security concerns regarding data encryption and corporate privacy standards.
Atlas Newsdesk ·

New Desktop Monitoring Capabilities
A fresh update to the ChatGPT desktop application introduces a tracking mechanism designed to monitor user interactions across various software and websites. By leveraging operating system accessibility permissions, the tool constructs a chronological log of digital behavior. This data collection aims to help the underlying artificial intelligence better understand individual workflows to provide more tailored assistance.
Security and Privacy Vulnerabilities
While the system avoids capturing direct audio or visual snapshots, it introduces notable privacy challenges. Reports indicate that the gathered information is stored without encryption, potentially allowing other applications on the same machine to access sensitive logs. Although users retain the ability to toggle tracking for specific programs, the burden of securing this locally stored data rests entirely on the individual or organization.
Regional Availability and Corporate Impact
Currently, this functionality is restricted to Mac users with Pro, Business, or Enterprise subscriptions. Due to stringent data protection frameworks, the feature remains unavailable within the United Kingdom, Switzerland, and the European Economic Area. Officials suggest that corporate entities must conduct thorough compliance reviews before enabling this tool to ensure it aligns with internal security protocols and data governance policies.
Operational Risks and Limitations
The primary uncertainty involves how effectively organizations can mitigate the risk of unauthorized access to these unencrypted logs. Because the data resides locally, it may conflict with existing enterprise security standards that mandate end-to-end encryption for all business-related activity logs. Future updates may address these vulnerabilities, but for now, the feature represents a trade-off between AI-driven productivity and traditional cybersecurity hygiene.