Story perspectives

OpenAI patches critical ChatGPT macOS vulnerability after report

By Drooid · Oct 2, 2026, 3:09 PM UTC · How we work

Summary

Story summary

OpenAI patched a critical vulnerability in its ChatGPT macOS app.
The flaw could let attackers seize the app, read chat logs, and hijack browser sessions.
Objective-See Foundation researchers discovered the vulnerability and reported it privately to OpenAI.
Analyst Patrick Wardle warned that corrupted AI agents could grant unprivileged code broad access.
OpenAI logged the fix on September 25 and released the patch.

1 of 1

Story summary

OpenAI patched a critical vulnerability in its ChatGPT macOS app.
The flaw could let attackers seize the app, read chat logs, and hijack browser sessions.
Objective-See Foundation researchers discovered the vulnerability and reported it privately to OpenAI.
Analyst Patrick Wardle warned that corrupted AI agents could grant unprivileged code broad access.
OpenAI logged the fix on September 25 and released the patch.