As more healthcare organizations leverage machine learning and AI to support improvements in clinical outcomes, cost, and experience, they have a responsibility to avoid unintended consequences and “do no harm”.  Using a real-live example of a healthcare predictive model, this project presents a practical approach to how organizations can deliver data science products and services that are transparent, fair and ethical.  These include technical and operational methods that will uncover potential issues and demonstrate mitigation techniques.  Finally, the project will dig into ethically questionable uses of data and models to identify opportunities for organizations to have open and courageous dialogues and come to a decision on whether they should just because they could.