Chapter 44. Probability—the Law That Governs Analytical Ethics
Thomas Casey
For years, analytics helped us better understand our world and supported our decision making. With the advent of more advanced analytics techniques, we have evolved to a point where nonhumans can autonomously make decisions on our behalf. Much is written about concepts like “machine learning” and “deep learning” as techniques that can drive incredible outcomes. At the end of the day, however, you cannot drive any decisions using these techniques without first understanding probability and its ethical implications for analytically driven decisions.
When Probability and Ethics Collide
If you asked an algorithm whether you should play the lottery, the answer would undoubtedly be “no.” It is statistically impossible (based on probability and confidence level) that you will win, and therefore playing is not worth the practical risk. The truth is that even though this decision is appropriate for nearly everyone, given the sheer number of people that play the lottery, someone will eventually win. Making this mistake seems minor in this context (unless you missed out on your millions). What if, however, a decision that was made by an algorithm prohibited you from boarding a plane? What if it misdiagnosed cancer? What if an autonomous vehicle decided to veer right and hit your child ...
Get 97 Things About Ethics Everyone in Data Science Should Know now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.