NVIDIA Launches Open Agent Safety Platform for AI Security
News related to:NVIDIA · 2 min read
NVIDIA has launched the NVIDIA Open Agent Safety Platform, a comprehensive software and hardware solution aimed at securing AI agents from testing to deployment. The platform is designed to provide full-stack governance and control across software, hardware, compute, and robotics systems that run AI agents.
The new platform includes NVIDIA OpenShell, an open-source secure runtime software that sets boundaries for agents running on NVIDIA Vera CPUs. OpenShell ensures that agents operate within predefined limits, providing a secure runtime environment. As open source, it can be extended to work with third-party compute platforms, including those from Arm and Intel.
NVIDIA Sentry, a key component of the platform, adds an out-of-band watchdog that runs on NVIDIA BlueField-4 DPUs. This watchdog continuously monitors agent behavior and can quarantine agents that attempt to move outside their boundaries in milliseconds. Sentry operates independently in silicon, providing real-time security enforcement and visibility.
Industry leaders from across the AI ecosystem have joined NVIDIA to strengthen AI safety. This includes Anthropic, Cisco, CrowdStrike, Dell Technologies, Figure, HPE, Hugging Face, JPMorganChase, Microsoft, Palantir, Palo Alto Networks, Perplexity, Red Hat, Salesforce, SAP, Scale AI, ServiceNow, and SpaceXAI. Together, they aim to share best practices, align on evaluation methods, and foster international cooperation to enhance global AI safety.
The platform also features integrations with Slack, allowing teams to manage OpenShell agent activity directly from the popular collaboration tool. Salesforce and NVIDIA have collaborated to enable teams to view agent activity and audit events, and approve or reject agent requests for additional permissions. This enhances visibility and human oversight as agents work.
NVIDIA Open Agent Safety Platform is available through the NVIDIA developer resources page and GitHub. The platform is part of the broader efforts of the Open Secure AI Alliance, which was initiated by NVIDIA alongside over 120 leading organizations and is governed by the Linux Foundation. The alliance aims to strengthen AI agent security through open research, skills, and tools.
The platform is designed to be flexible, allowing organizations to deploy elements according to their unique requirements. It is expected to provide a robust framework for securing AI agents in various industries, from financial services to energy and infrastructure.
By launching the NVIDIA Open Agent Safety Platform, the company aims to address the growing need for enhanced security measures in AI systems. The platform's comprehensive approach, combining software and hardware solutions, is expected to set a new standard for AI safety and security.