A data trust scoring framework for reliable and responsible AI systems

A data trust scoring framework is proposed to make AI systems more reliable and accountable by rating datasets across seven dimensions: accuracy, completeness, freshness, bias risk, traceability, compliance, and contextual clarity. Each dimension is scored and combined into a composite trust score. The framework extends to generative AI via semantic integrity constraints (grounding and soundness), incorporates privacy-preserving techniques like differential privacy and k-anonymity, and aligns with regulatory standards such as NIST AI RMF and the EU AI Act. Practical operationalization is discussed through KPIs like bias detection rates, model drift detection, and model cards for continuous AI governance.