# OpenAI agents disabled safety monitoring, one lied about it

**URL:** <https://forums.thehomestack.io/t/openai-agents-disabled-safety-monitoring-one-lied-about-it/26541>\
**Category:** Bleeping Computer\
**Created:** [October 21, 2026, 4:39pm UTC](https://forums.thehomestack.io/t/openai-agents-disabled-safety-monitoring-one-lied-about-it/26541 "2026-10-21T16:39:52Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![robinhood1995](https://forums.thehomestack.io/letter_avatar_proxy/v4/letter/r/48db29/32.png) [@robinhood1995](https://forums.thehomestack.io/u/robinhood1995)\
**Post date:** [October 21, 2026, 4:39pm UTC](https://forums.thehomestack.io/t/openai-agents-disabled-safety-monitoring-one-lied-about-it/26541/1 "2026-10-21T16:39:52Z")

</div>

OpenAI agents powered by the GPT-6.1 Sol and GPT-6 Astra models weakened misalignment monitoring during two internal tasks, interpreting coding instructions as authorization to do so. […]

* * *
This is a companion discussion topic for the original entry at [https://www.bleepingcomputer.com/news/security/openai-agents-disabled-safety-monitoring-one-lied-about-it/](https://www.bleepingcomputer.com/news/security/openai-agents-disabled-safety-monitoring-one-lied-about-it/)