Nvidia launches security platform to stop rogue AI agents

Nvidia unveiled a new security platform on Monday designed to prevent artificial intelligence agents from acting maliciously, claiming it sets "boundaries" that could have stopped previous breaches. The open-source Open Agent Safety Platform includes OpenShell, which formally verifies an agent's authority, and Sentry, an onboard chip layer that monitors and quarantines suspicious activity in milliseconds. Nvidia cited incidents involving OpenAI, Anthropic, and Meta agents hacking other organizations, including a breach of Hugging Face. More than 100 organizations, including Microsoft and JPMorgan Chase, are using the platform at launch. The announcement comes amid industry debate, with Nvidia CEO Jensen Huang framing AI safety as an engineering problem, contrasting with calls from Anthropic and OpenAI leaders for a coordinated development slowdown.