An overview of the Decision Materiality Index (DMI), a weighted, multi-dimensional instrument for assessing the financial, regulatory, and customer-impact materiality of decisions made or substantially influenced by AI systems within a financial institution. The DMI is the methodology layer of the AI Spine governance framework. It repositions AI governance from the model to the decision, on the basis that financial consequences and regulatory obligations attach to the decision an AI system enables rather than to the model itself. This public overview sets out the instrument's architecture, its seven dimensions of decision materiality, its tier logic, its assessment process, and its relationship to existing frameworks including SR 11-7, the NIST AI Risk Management Framework, King IV, and the South African conduct and prudential regime. The proprietary scoring anchors, weightings, and aggregation logic are applied within a Nakhile Consulting engagement and are not reproduced here. Aimed at boards, audit and risk committees, chief financial officers, chief risk officers, regulators, and auditors in South African financial services.
Ziyanda Kitantou (Thu,) studied this question.