Microsoft’s new guidance on threat modeling for AI applications arrives at a moment when enterprises are scrambling to put generative and agentic systems into production — and it does something important: it forces security teams to stop treating AI as “just another component” and start modeling...
