Having high quality data sets, including training, validation and testing data sets, with appropriate data governance and management practices plays a vital role in ensuring that AI systems produce high quality outputs and reduce unfair bias or discrimination. High quality data sets are essential for the training, validation and testing of AI systems and help to ensure the accuracy and reliability of their outputs. High quality data should be free from errors and should be representative, relevant, up to date and complete in light of the intended purpose of the AI system. Depending on their intended purpose, the data sets should also be sufficiently statistically significant as well as free from individual or societal biases that might skew the outputs.