NVIDIA, Microsoft, CrowdStrike and more than 30 industry participants have backed a new coalition focused on open-source tools for AI safety and security, with the technical challenge centered on making complex AI systems easier to inspect, test, and govern.
A new 37-member alliance points to a wider shift in AI defense: agents, logs, harnesses, and remediation are becoming as important as the model itself.
NVIDIA, Microsoft, and CrowdStrike have joined an open coalition aimed at defending AI-enabled infrastructure with shared tools, models, agent harnesses, and security techniques.