AI RMF stands for NIST’s AI Risk Management Framework. It is a lifecycle framework for identifying, measuring, and governing risks from AI systems, from design and procurement through deployment and ongoing monitoring. The goal is not just to make an AI model accurate, but to make its use acceptable, explainable, and controllable in a business environment.
In cyber security, AI RMF matters because AI can introduce new attack paths and operational risks: prompt injection, data leakage, model poisoning, unsafe automated decisions, and weak vendor oversight. Defenders use the framework to inventory AI use cases, define approval gates, test for misuse, track drift, require human review where needed, and keep logs for audit and incident response. In practice, AI RMF helps organizations treat AI like any other high-risk system, with controls matched to the impact of failure.