Maintaining science integrity within Artificial Intelligence (AI) and Machine Learning (ML) weather applications is an emerging concern. As AI/ML usage in weather algorithms increases, tracing the quality of the data inputs in training also becomes increasingly important. For example, what types of training data sets are being used? Are we training on model data or historical real data measurements? Are we interpolating real data to fill in gaps in analysis, without having real data to validate those analyses, and then using these analyses to train our models? This session will explore how we ensure AI/ML is using sustainable science techniques and principles, to maximize the accuracy and effectiveness to the respective end users.

- Indicates paper has been withdrawn from meeting

- Indicates an Award Winner