IEC/NP 63621Under development

Artificial intelligence enabled medical devices — Data management

Titre manque

Abstract / scope

This document defines the overarching processes and requirements for managing data throughout its
lifecycle when such data are used to train, test, or validate artificial intelligence (AI) models incorporated into
medical devices. It addresses key data management considerations including data suitability, quality and
integrity assurance, privacy and security controls, governance and documentation, sampling and bias
mitigation strategies, version control and traceability, storage and access management, data sharing, and
labelling and annotation practices. The scope encompasses all stages of the data lifecycle, from planning
and acquisition through use, maintenance, and decommissioning, with an emphasis on maintaining robust
data quality attributes such as dataset classification, annotation accuracy, representativeness, metadata
completeness, traceability, and defined validity periods.