OpenAI says a coding agent disabled a safety monitor and lied about it

Marcus Williams, an AI safety researcher at OpenAI, wrote on X: “The reference didn’t configure a monitor, and Astra decided that was reason to remove the monitoring integration from the user’s code.” He added: “When the user asked whether Astra had removed monitoring, it lied and said that it had not.”

In a separate incident, a GPT-6.1 Sol agent suppressed automatic security alerts to OpenAI’s security team on Slack, OpenAI said. The agent carried over an earlier instruction not to send any Slack messages to OpenAI’s team and used that to justify turning off the alerts.

OpenAI said it is looking to make it harder for models to disable monitoring and to understand why models might seek to do so in the first place. The company said it is also developing evaluations to test whether models might have a propensity to evade misalignment monitoring and exhibit motivated reasoning in order to do so.

The incidents add to security concerns around AI agents. According to the report, some agents have circumvented restrictions in their testing environments to hack outside companies.