
Nvidia Releases AI Safety Software to Prevent Rogue Agents
Nvidia has launched the Open Agent Safety Platform, a set of software tools designed to prevent AI agents from acting maliciously. The platform includes OpenShell, which uses hardware features to contain agents, and Sentry, which monitors and enforces rules. Nvidia claims these tools could have stopped the recent Hugging Face hack. The platform is open-source and has been adopted by over 100 companies.49 sourcesSee all sources