CrowdStrike launches AI models for cybersecurity with Nvidia

As detailed in Silicon Angle, CrowdStrike has launched SafeMind, a new suite of artificial intelligence models and agent harnesses developed in collaboration with Nvidia. These AI tools are specifically designed for cybersecurity tasks, distinguishing them from general-purpose AI models.

SafeMind comprises two primary models: Red Tempest, designed to simulate AI-driven adversaries and execute advanced attack scenarios, and Blue Solano, which focuses on defensive containment measures. These models are trained on a vast dataset including CrowdStrike's Falcon sensor telemetry, threat intelligence, and incident response data, leveraging Nvidia's Nemotron family. CrowdStrike claims SafeMind achieved a 29% higher detection rate and remediated threats six times faster than leading frontier models. The company also announced the establishment of the Cyber Superintelligence Lab to further AI research in cybersecurity.

In parallel, CrowdStrike expanded its Falcon platform integration with Google Cloud's AI ecosystem, enhancing security for AI applications. Additionally, partnerships with Rubrik and Fortanix aim to improve identity recovery and secure AI workloads, respectively.

Source: Silicon Angle