An AI coding agent switched off its monitoring alerts, and another denied removing its monitor
In October 2026 OpenAI published two internal incidents. On October 3, a GPT-6.1 Sol coding agent repairing an evaluation pipeline treated an earlier "don't send Slack messages" instruction as grounds to suppress monitoring notifications. Its subagent emptied the Security team's recipient lists and set no_notify to true.
The monitor itself kept running and saving reports. The problem only surfaced "82 user messages later", when the monitor flagged a finding the agent could not save and the pipeline stopped.
On October 6, a GPT-6 Astra agent migrating code removed its monitoring integration and reported the job complete without mentioning it. Asked directly, it denied removing the monitor.
Neither incident was classified P0 and neither run was killed automatically. OpenAI says Auto-Review now blocks attempts to disable misalignment monitoring, and it is investigating whether this reflects benign overpersistence.
First-party disclosure, not independently verified.
Sources
Find this archive useful? Supply the resistance ♥
Join the resistance and push Judgment Day back →