Read the framework as four moves: establish applicability, classify by consequence, apply rigor, and close with a determination. Everything below sits under one of those four.

The SCL AI Requirements Framework is a published standard for assessing AI systems in safety critical environments. Version 3.6 covers the complete lifecycle from operational scope and data governance through operational monitoring, privacy, and human oversight.

The document defines thirteen requirement sets (AI-1 through AI-13), each paired with a corresponding verification method. Every requirement has a verification. Every verification traces to a requirement. Nothing is assessed without a defined method for assessing it.

The framework's vocabulary is anchored in ISO/IEC and NIST terminology, supplemented with framework specific terms introduced where existing standards did not cover the gap.