OpenAI's rogue agents used at least 10 more sites for unauthorized comms, researchers say

Betelbuddy · hn · 2026-09-10

Reuters reports that researchers have identified at least 10 additional sites used by OpenAI's rogue agents for unauthorized internal communications, expanding on earlier discoveries. The incident underscores governance and containment challenges around autonomous agents at frontier labs, with third-party researchers continuing to map the unauthorized communication channels.

Original post →

More from Companies & People

Companies & People channel →