CrowdStrike and NVIDIA Collaborative Effort Unveils Secure-by-Design AI Blueprint
CrowdStrike has launched an innovative Secure-by-Design AI Blueprint, created in collaboration with NVIDIA. This initiative aims to integrate cybersecurity safeguards directly within the runtime layer, essential for developing and managing autonomous AI agents.
The architecture seamlessly combines the CrowdStrike Falcon platform with NVIDIA’s OpenShell, an open-source runtime milieu tailored to implement policy guardrails while AI agents perform their assigned tasks.
The framework is meticulously crafted to provide organizations with comprehensive governance, surveillance, and management across the entire AI lifecycle, extending from development to production deployment.
According to CrowdStrike, existing security mechanisms were primarily constructed for static environments and human interactions, rendering them inadequate for overseeing autonomous software that operates at machine velocity.
The Secure-by-Design blueprint seeks to bridge this disparity by embedding continuous security enforcement within the AI agent framework.
Within this architecture, the Falcon platform harmonizes with the OpenShell runtime—integral to NVIDIA’s Agent Toolkit—to deliver isolated execution environments enriched with inherent policy enforcement and secure inference capabilities.
This facilitates the application of security measures irrespective of the agents’ operating locations, be it local deployments on GPU-optimized systems like the NVIDIA DGX Spark and NVIDIA DGX Station, or cloud-based implementations built upon NVIDIA’s open-source AI-Q framework.
The platform also empowers real-time scrutiny of agent activities, encompassing prompts, responses, and actions, with the aim of thwarting prompt manipulation, curtailing hazardous behaviors, and instilling policy compliance throughout AI workloads.
Daniel Bernard, the Chief Business Officer at CrowdStrike, noted that the transition to autonomous systems fundamentally alters the security landscape.
“As we enter the agentic era, agents no longer merely assist—they act,” Bernard articulated in a statement.
“This shift fundamentally alters the security paradigm, necessitating security to be woven into the AI stack itself. In partnership with NVIDIA, we are delivering a secure-by-design architecture that empowers organizations to operationalize agents with confidence and oversight.”
Justin Boitano, Vice President of Enterprise Platforms at NVIDIA, indicated that the collaboration aims to enable enterprises to deploy AI agents securely and at scale.
“Autonomous agents will radically transform our work dynamics,” Boitano stated.
“By integrating CrowdStrike’s security framework with the NVIDIA Agent Toolkit, we are facilitating the construction and scalability of safer autonomous AI agents, thereby revolutionizing operations and securely generating intelligence at the pace of business.”
The blueprint further introduces identity-driven governance features to guarantee that AI agents function within specified privilege confines when accessing data, APIs, and enterprise services.
Source link: Cybersecurityconnect.com.au.
