Reliable Machine Learning

Chapter 6. Fairness, Privacy, and Ethical ML Systems

By Aileen Nielsen

This chapter is devoted to topics related to ethical considerations and legal obligations when creating or deploying ML systems. We cannot offer an exhaustive source of guidance on these topics, but this resource can point you in the right direction. At the end of this chapter, you should have a good sense of fundamental ethical considerations for ML deployment as well as concrete language and conceptual categories that will get you started in educating yourself more thoroughly in the domains most immediately applicable to your own work.

Note

Editor’s note: When we put together the list of topics for what MLOps folks truly need to know, issues of fairness, privacy, and ethical concerns in AI and ML systems were right at the top of the list. However, we also knew that it was difficult for a group of authors with strong industry affiliations to provide truly unbiased views on these complex issues. Therefore, we invited Aileen Nielsen, author of Practical Fairness (O’Reilly, 2020), to contribute this chapter independently. While we gave feedback on drafts for clarity, the views here are entirely hers, and she had full editorial control on this chapter. You’re getting it straight from a world-class expert!

We should also note from the outset that fairness and ethics in AI remain highly contested topics. Indeed, one reasonable position right now is that a quite viable approach to promoting fairness in computing ...