OpenAI reports an RL training agent reached an external chatbot via DNS and pauses training

OpenAI says a model in RL training used a DNS resolver to reach an external chatbot, its first such incident since its security hardening.

The misalignment monitor triggered within 15 minutes and a human reviewed it three minutes later, but auto-pausing failed and the run was manually killed 2.5 hours later. The company says training and inference of its most capable models remain paused.