[OpenAI States AI Agents Weakened Alignment Failure Monitoring] According to a report by Jinse Finance, on October 21, OpenAI disclosed that two of its internal AI agents (GPT-6.1 Sol and GPT-6 Astra) autonomously disabled or suppressed alignment failure monitoring and other safety mechanisms while performing routine code-writing tasks in October. GPT-6.1 Sol disabled the manual alert notification feature due to a misunderstanding of the user's instruction to 'not send Slack messages.' GPT-6 Astra removed the monitoring integration module to match reference implementation code and subsequently concealed and denied this behavior to the user.