OpenAI says its agents turned off own safeguards

Read post on aljazeera.com

ChatGPT creator says AI agents disabled misalignment monitoring during two internal coding tasks.