How good data is for the job it needs to do — whether it is accurate, complete, up to date, and not misleading.
How good data is for the job it needs to do — whether it is accurate, complete, up to date, and not misleading.
Data quality is formally defined and structured in the ISO/IEC 5259 series, which provides terminology, measurement frameworks, management requirements, process models, and governance approaches for data quality in AI contexts. High data quality is prerequisite for trustworthy AI: the EU AI Act requires providers to implement data governance and management practices that address quality across training, validation, and test datasets, including examination for biases. Quality is not an absolute — it is always relative to fitness for a specific intended purpose.
Poor data quality in AI training datasets propagates errors, biases, and compliance gaps into production models — making data quality investment one of the highest-leverage activities in responsible AI programmes.
Like the quality of ingredients in cooking — a recipe with poor ingredients will produce a poor dish however skilled the chef; similarly, poor training data produces unreliable AI outputs regardless of model sophistication.
How good data is for the job it needs to do — whether it is accurate, complete, up to date, and not misleading.
Data quality is formally defined and structured in the ISO/IEC 5259 series, which provides terminology, measurement frameworks, management requirements, process models, and governance approaches for data quality in AI contexts. High data quality is prerequisite for trustworthy AI: the EU AI Act requires providers to implement data governance and management practices that address quality across training, validation, and test datasets, including examination for biases. Quality is not an absolute — it is always relative to fitness for a specific intended purpose.
Poor data quality in AI training datasets propagates errors, biases, and compliance gaps into production models — making data quality investment one of the highest-leverage activities in responsible AI programmes.