OpenAI has released Computer History, a new ChatGPT tool that tracks mouse clicks and keyboard activity to help the AI learn from user behavior, sparking cybersecurity concerns over unencrypted data storage.
Feature availability and functionality
Computer History is now available for some Apple Mac users, according to Ari Weinstein, OpenAI's product and engineering manager, who announced the release on X last week. The feature is optional, off by default, and available only for Pro plan users with desktop memory enabled.
It does not take screenshots or recordings, nor does it listen to microphones or system audio. Instead, ChatGPT converts input events—such as mouse clicks, keystrokes, keyboard shortcuts, and app switches—into a timeline. In a demo, Dominik Kundel of OpenAI's Developer Experiences team showed the bot locating his last edited document and offering a shareable link when it confirmed he hadn't sent it via Slack.
Data handling and user controls
OpenAI says users can exclude apps and websites, and the feature does not track activity in private browsers. Users can inspect and delete memorized data at any time. The timeline is stored in local memory files, which are deleted after 48 hours but are not encrypted. OpenAI warns that other programs running as the macOS user may access these files, advising users to protect their Mac accounts and exclude unwanted sources.
OpenAI does process the files into 'memories' but states it does not retain event files after processing unless required by law, and does not use them for training.
Expert concerns and comparisons
Cybersecurity experts interviewed by Metro highlighted risks. Stefanie Schappert of Cybernews said unencrypted summaries could be stolen by malware, exposing health, financial, and personal information for phishing or identity theft. Wade Woolwine of Rapid7 argued that opting in does not significantly increase risk, as attackers would already need to compromise the computer.
OpenAI acknowledges a prompt injection risk, noting an increased chance ChatGPT may follow malicious prompts from dodgy apps or websites. Alex Goller of Illumio compared the feature to Microsoft's Recall, which faced criticism for lacking a disable option. Goller noted OpenAI has learned from that situation, making Computer History opt-in and excluding private browsing, but stressed the need for high security standards as AI adoption grows.
