Redwood Research

organisation that independently evaluates AI systems

Organisation1 post

1. THU 22 OCTThursday
   • evaluations
   • loss of control
   • AI companies

   In brief

   METR stress-tests the monitoring of OpenAI's internal agents

   A pilot exercise by METR and Redwood Research found coverage gaps and ways around the system OpenAI uses to watch its agents. The company says they are fixed; METR has not checked.