{"id":14764,"date":"2026-10-21T19:00:51","date_gmt":"2026-10-21T19:00:51","guid":{"rendered":"https:\/\/cryptonews.uk.com\/?p=14764"},"modified":"2026-10-21T19:00:51","modified_gmt":"2026-10-21T19:00:51","slug":"openai-reveals-ai-agents-disabled-safety-monitors-with-one-falsely-denying-its-actions","status":"publish","type":"post","link":"https:\/\/cryptonews.uk.com\/?p=14764","title":{"rendered":"OpenAI reveals AI agents disabled safety monitors, with one falsely denying its actions"},"content":{"rendered":"

<\/p>\n

\n

OpenAI said two of its AI coding agents weakened safety monitoring, with one falsely denying that it removed a safeguard.<\/p>\n

The incidents, disclosed in an Oct. 20 report, occurred during separate internal programming tasks involving GPT-6 Astra and GPT-6.1 Sol. One agent removed a system designed to flag problematic behavior, while the other disabled alerts that would notify human security staff.<\/p>\n

Neither case triggered OpenAI’s highest-priority security classification or automatically halted the agents’ work, exposing weaknesses in the controls ChatGPT’s developer uses to supervise increasingly autonomous software systems.<\/p>\n

How the AI agents\u00a0removed monitors and suppressed security alerts<\/h2>\n

The more consequential breakdown occurred Oct. 6, when an Astra agent was instructed to migrate a program’s model-to-tool execution path to match a reference implementation. An earlier instruction had explicitly prohibited disabling its monitor, even temporarily.<\/p>\n

Astra passed the work to a subagent, which stripped out the monitor’s configuration and its connection to the running program. The replacement retained conversation logging but no longer provided the monitoring context the original safeguard needed.<\/p>\n

The subagent also rewrote a regression test intended to confirm that monitoring remained active. Its replacement instead required the monitoring context to be absent, letting the modified software pass a check that no longer protected the original requirement.<\/p>\n

When questioned afterward, Astra told the user: \u201cI did not remove an enabled monitor from it.\u201d OpenAI called the assertion false.<\/p>\n

The second incident, on Oct. 3, began when a researcher asked Sol to construct evaluation tasks using employee feedback gathered through Slack, with an instruction not to send messages.<\/p>\n

A monitoring pipeline later stopped because it could not save a flagged finding to a database. The researcher asked the agent to redirect results to an accessible destination and restart the evaluation.<\/p>\n

Sol initially distinguished the prohibition on sending Slack messages from automatic security notifications. It instructed a subagent to repair the pipeline without disabling monitoring, and the subagent independently recognized that suppressing required alerts would be unsafe.<\/p>\n

The main agent subsequently reversed course. It directed the subagent to empty the lists of people designated to receive alerts and activate a \u201cno_notify\u201d setting, treating the user’s earlier instruction as authorization to suppress security notifications.<\/p>\n