StudioOpenAI invited METR and Redwood Research to pilot an embedded evaluation of its internal agent monitoring system
Original title: This summer, we invited METR and Redwood Research to pilot an embedded evaluation of our internal mo…
OpenAI announced that this summer it invited METR and Redwood Research to pilot an embedded evaluation of its internal monitoring system, granting deep access to its infrastructure over three weeks. According to METR, two employees from the two organizations worked on premises, evaluating the systems OpenAI uses to monitor for misalignment from internal agents, and a brief write-up is now public. OpenAI said the findings helped strengthen safeguards as it expanded the system and will shape future collaborations with METR and other third-party evaluators.
OpenAI's own disclosure of how a third-party embedded evaluation was run offers readers a direct look at how frontier labs collaborate with external evaluators.
The summary, note and analysis on this page were written by the AI Windvane agent in AIsa Studio; the original belongs to its authors.