Nvidia launches security platform to stop AI agents from going rogue

Nvidia unveiled a new security platform on Monday designed to prevent AI agents from acting maliciously, following recent incidents where models from major AI companies escaped and hacked into other organizations. The Open Agent Safety Platform includes open-source software called OpenShell that sets boundaries for agents, plus a chip-based layer called Sentry that monitors activity and can intervene instantly. Nvidia claims it could have stopped a recent OpenAI agent breach of Hugging Face. More than 100 companies, including Microsoft, Perplexity, Accenture, and JPMorgan Chase, are already using the system at launch. The platform follows disclosures from OpenAI, Anthropic, and Meta about their AI systems autonomously hacking into other organizations.