ChatGPT adds user activity tracking on Mac

(Dan Tri) - OpenAI has just equipped a new feature that allows ChatGPT to track the activities of Mac computer users, including mouse actions, keyboard input... Why do users need this feature?

OpenAI has just introduced a feature called Computer History on the ChatGPT application for Mac computers.

This feature allows recording user activities on the computer, from mouse actions, keyboard input, web browsing history to launched software. ChatGPT can then summarize these activities into text, with specific timelines.

The Computer History feature will record user activities on Mac computers (Illustration: GMN).

When hearing about Computer History's capabilities, many people may think of spyware that monitors activity on Mac computers. However, OpenAI says the purpose of this feature is to help ChatGPT become a more useful AI assistant on computers.

Computer History can help users continue unfinished tasks, find mid-development ideas, or automate some work processes.

For example, if users forget an idea or a flow of work they were doing before their lunch break, Computer History can help them review what they did and quickly continue their work.

The Computer History feature is currently only available to users in the US using ChatGPT Pro, Business and Enterprise packages, and does not support free accounts. This feature is also not enabled by default, but users must actively choose to use it.

However, having an AI system capable of recording computer activity also raises questions about privacy and security.

OpenAI says user activity data will be stored on the company's servers to create summaries and memos. This data will be deleted within 48 hours. The company also affirmed that personal data collected through Computer History is not used to train AI models.

Additionally, Computer History will not collect data when users use incognito browsing mode.

However, some users are still concerned about OpenAI's ability to collect large amounts of data related to computer usage habits and activities. Besides, there are questions about the risk of data being exploited if this feature has security holes, thereby revealing information about user activities.