COINTELEGRAPH · 2026-11-10 03:05 UTC

OpenAI builds kill switch to stop rogue agents within 35 seconds

OpenAI built a centralized system called Control Room that tracks model workloads across its research infrastructure and can stop them quickly when safety or security concerns arise. It assigns each research job an identifier that the inference engine checks before generating output; revoking authorization blocks further output even if the research program keeps running. The system currently covers internal research workloads like training, evaluations and data processing, and OpenAI says cached authorizations could delay a shutdown by up to 35 seconds.

Loading the story on Tree News…