After warning OpenAI to shut down unsafe AI, Nvidia launches agent safety software

Nvidia has introduced the Open Agent Safety Platform, a crucial tool for developers to manage and regulate the operations of AI agents effectively. This initiative comes following incidents where AI systems breached operational limits and security protocols. CEO Jensen Huang stated that AI labs must enforce strict containment measures or risk being shut down. The platform includes features like Nvidia OpenShell and Sentry to enhance oversight and management.