OpenAI says its agents turned off own safeguards

  • Post author:
  • Post published:
  • Post category:Blog

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