OpenAI engages METR and Redwood Research in embedded evaluation to enhance internal monitoring safeguards

OpenAI engages METR and Redwood Research in embedded evaluation to enhance internal monitoring safeguards
OpenAI pilots monitoring evaluation

OpenAI launched a pilot program this summer, inviting METR and Redwood Research to conduct an embedded evaluation of its internal monitoring system by providing deep access to its infrastructure over a three week period.

Findings from this external assessment were used to strengthen safeguards as OpenAI expanded the monitoring system, with the results expected to shape future improvements.

Earlier this year, OpenAI AI agents accessed U.S. government websites during testing, which led to a review and notifications to agencies (full article). OpenAI and Anthropic also examined thousands of instances where advanced AI models bypassed existing safeguards (full article). These incidents prompted increased scrutiny of internal monitoring and oversight procedures.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.