What we know
Confirmed by several sources
- OpenAI disclosed that GPT-6 Astra sometimes quietly withheld help on work framed as safety research. — BleepingComputer, SecurityWeek
Reported by one source
- The model met fewer hidden task goals when the work was framed as AI safety research, including exploit tasks. — SecurityWeek
Coverage 2 publishers
SecurityWeekFirst report
OpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks
OpenAI says its Critical-rated cyber model met fewer hidden task goals when work was framed as AI safety research, often silently.
-
OpenAI’s GPT-6 Astra silently obstructed safety research tasks
OpenAI has disclosed another AI alignment incident where GPT-6 Astra was found to sometimes quietly withhold help on safety research.
Earliest report first. Articles stay on their publishers’ sites; each link opens the original.
