A vulnerability has been detected in an OpenAI high-performance model provided to the Japanese government.

This is an ingested summary. Full reporting is available at the primary source below.