Data Quality in the Age of AI

by Andrew Jones

Assessment of human exposure to chemicals is a critical input to risk assessment and ultimately to decisions about control of chemicals. This two-part publication aims to improve the quality of information available to decision-makers and its communication. Part one sets out ten principles for characterizing and communicating uncertainty in exposure assessment. A tiered approach to the evaluation of uncertainties using both qualitative (simple) and quantitative (more complex) methods is described. Different sources of uncertainty are identified, and guidance is provided on selecting the appropriate approach to uncertainty analysis, as dictated by the objectives of the assessment and information needs of decision-makers and stakeholders. Part two addresses the quality of data used in exposure assessment, and sets out four basic hallmarks of data quality - appropriateness, accuracy, integrity and transparency. These hallmarks provides a common vocabulary and set of qualitative criteria for use in the design, evaluation and use of exposure assessments to support decisions.--Publisher's description.