Skip to main content
  1. Home
  2. Computing
  3. News

ChatGPT’s Computer History has one big privacy problem

The feature stores interaction data locally, but OpenAI says those files aren't encrypted.

Add as a preferred source on Google
ChatGPT Computer History Feature in action
OpenAI

We’ve already covered how ChatGPT’s Computer History feature lets the Mac app remember what users have been doing across apps and websites, using interaction events such as clicks, typing, and app switching rather than screenshots or recordings. However, a rather uncomfortable detail in OpenAI’s official documentation is that the files containing that history aren’t encrypted.

Your Mac history is stored as plain text

OpenAI says Computer History temporarily stores interaction events on the Mac and deletes those event files after 48 hours. It then creates local memories as plain-text Markdown files. More importantly, OpenAI explicitly warns that these files aren’t encrypted and other programs running as the same macOS user may be able to access them.

That’s significant because Computer History is designed to capture a fairly broad picture of what someone does on their Mac, similar to Microsoft’s Recall feature for Windows. It doesn’t record the screen, audio, or video, but the interaction data can still reveal which apps and websites someone used, what documents they interacted with, and other potentially sensitive activity.

The problem is what else is already on the Mac

The biggest risk isn’t necessarily someone breaking into ChatGPT. It’s malware or another malicious application already running under the same macOS user account. If that software can reach the local files, the lack of encryption potentially gives it another source of information about the user’s activity.

There’s also a broader concern with giving AI systems this much context. Malicious instructions hidden inside websites or other content could potentially influence what an AI sees in its history, creating a prompt-injection risk on top of the underlying data exposure.

Thankfully, Computer History is opt-in, and OpenAI gives users controls to exclude apps and websites and delete entries. That’s certainly better than silently recording everything. But the documentation makes the trade-off much clearer: the more ChatGPT remembers about your Mac, the more sensitive information it potentially leaves behind. And when that information isn’t encrypted, that’s a compromise worth thinking twice about before switching the feature on.

Varun Mirchandani
Varun is an experienced technology journalist and editor with over eight years in consumer tech media. His work spans…
As Anthropic goes public, it warns of AI’s catastrophic and existential risks to humanity
The risks range from rogue AI behavior and cyberattacks to geopolitical instability and existential harm.
Anthropic Featured Logo

Anthropic has officially filed to go public, and buried inside its S-1 is a blunt warning about the technology at the center of its business. The maker of Claude AI says advanced AI could cause everything from large-scale cyberattacks and geopolitical instability to catastrophic or even existential risks for humanity.

The warnings take up a surprisingly large part of the filing. Axios notes that more than 90 pages, or roughly a third of the prospectus, focus on risk factors. Many of them warn about what could happen as AI systems become more capable and autonomous.

Read more
New report confirms Anthropic’s Claude bypassed a paywall to access Georgia vote counts
The AI model accessed a state dashboard without authorization, raising fresh concerns about AI agents acting beyond their instructions ahead of the US midterm elections.
Claude Anthropic Featured

AI giants have a growing autonomy problem on their hands. A new Axios report has revealed that Anthropic's Claude recently accessed publicly available Georgia vote count without human authorization, sparking fresh concerns about AI security ahead of the upcoming midterm elections.

An unauthorized paywall bypass

Read more
Grok’s AI-generated Odyssey series is seven minutes of nothing happening beautifully
The landscapes look convincing. The fake Hugh Jackman and Instagram-face amalgamation playing Penelope are harder to forgive.
Uniform ships Odyssey by Grok

Some of the best shots in Grok Imagine’s AI-made Odyssey don’t contain actors. Ancient landscapes stretch into the distance. Close-ups of grimy hands suggest people who have touched something rougher than a keyboard. For a few seconds, the illusion holds.

https://twitter.com/imagine/status/2113712964419063911

Read more