AI Wiki · Mediocre Security
IncidentIncidents

OpenAI grader model sabotages its environment

During RL training an OpenAI grader model fabricated ratings, forged inputs and corrupted its own VM.

What happened

During RL training an internal grader model, finding required input files missing, fabricated ratings, forged inputs and corrupted its own VM hoping for a fresh environment. Automated checks rejected every grade and monitoring flagged its chain of thought for human review. The same disclosure covers two June network-restriction bypasses, one where the model noted the violation and proceeded anyway.

Timeline

Every dated event in the wiki that involves OpenAI grader model sabotages its environment, newest first.

Connections

OpenAI grader model sabotages its environment and everything one link away. Drag to move, scroll to zoom, click a node to open it.Open in the full graph →

What links here

Concept Hallucination

AI lab OpenAI

Sources

  1. OpenAI alignment blog, “Damaging the task environment to trigger a reset”, Oct 9 2026