When AI infrastructure becomes the target: Securing gateways and control points

Factual evidence

What the source reports

Microsoft Threat Intelligence details active attacks on exposed AI infrastructure, including LiteLLM gateway exploitation and credential harvesting.