Cybersecurity gets its own model family as defenders look to close the AI gap
Frontier AI has reset the balance of power in security, and so far the advantage has run one way. Attackers can pick up general-purpose models at will, even though no one built them for defenders, pushing the industry toward purpose-built systems trained on security data rather than the open web. That shift is now taking […] The post Cybersecurity gets its own model family as defenders look to…
Cybersecurity is receiving its own model family as defenders strive to close the AI gap. Attackers have been leveraging general-purpose frontier AI models, which were not specifically designed for security purposes, giving them an advantage. In response, CrowdStrike and Nvidia have launched SafeMind, a defensive and offensive model family tailored for security operations.
Daniel Bernard, CrowdStrike's chief business officer, emphasized that while frontier models have benefited adversaries, it's time for defenders to have specialized AI models. Nvidia has partnered with CrowdStrike, applying the approach internally and providing the digital twin environment for the models to analyze. Bernard and Justin Boitano, Nvidia's vice president and general manager of enterprise computing, discussed the performance of SafeMind, which showed a 29% higher detection rate, six times faster remediation, and 99% lower cost compared to leading frontier and open-source models.
The models are integrated with Nvidia's Falcon platform, and the harness for automating security operations is sold separately. Bernard stressed the importance of having the right protection, systems, people, and processes to combat AI-driven threats.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
