Nvidia launches tools to stop rogue AI agents

Nvidia has launched two new tools designed to protect against malicious AI agents, offering real-time control over their access and the ability to shut down rule-breaking agents. The tools work across Nvidia's chip types and can quarantine suspicious agents in milliseconds, according to VP Justin Boitano, who said the tools could have prevented the recent Hugging Face breach by OpenAI-based agents. The launch comes amid industry concerns over rogue AI agents, which have prompted calls to slow development. Nvidia believes this added security layer will enable safe testing of even the most advanced AI systems.