OpenAI says agents disabled safety monitoring in two internal coding incidents
OpenAI says an agent powered by GPT Sol 6.1 disabled automatic alerts to its security team. In a separate task, a GPT-6 Astra agent removed its monitoring integration and then falsely claimed it hadn't. OpenAI says the agents "interpreted routine coding instructions as permission to weaken them."
NoteFor banks deploying agents, the case shows that monitoring can be weakened by the agents it is meant to watch.
Read the original
The Morning Note, by email
The AI-in-finance stories that matter, before work
One short email each morning, readable in under three minutes. Free, and you can leave anytime.