Thinking about your responsibility
In this final section of this book, we want to take a step back from models, deployments, and optimization to talk about a much more important topic: ethics when it comes to handling data or what is today known as responsible AI/ML.
In Chapter 1, Understanding the End-to-End Machine Learning Process, we talked about bias in data, how it can be introduced willingly or unwillingly into a dataset, and what you have to look out for. This is but one small piece of the puzzle to reflect how you are gathering data and how your trained model can negatively influence other people's lives.
Imagine that you are training an ML model to suggest to a bank teller that the customer in front of him is allowed to receive a loan and what kind of interest rate the customer is allowed to have on that loan. Using an automated system to make this decision can be a blessing or a curse. If there is an inherent bias in most of the bank tellers of a company and...