To address the question of ethics in any arena, including Data Science, first, we need to question ourselves what is the standard used to define what is “good” and “bad.” The importance of knowing such a standard is fundamental since choosing the wrong standard can generate false definitions for what is “good” and “bad” with a variety of consequences in society and, in this case, in the practice and use of Data Science. Hence, the standard must be absolute because if it changes, then the meaning of “good” and “bad” is lost, and we fall into a moral relativism situation.
Kreeft (2004) suggests that to talk about ethics, we must ask ourselves, which is the moral standard that we use in our daily lives. If we cannot answer such a question, we should embark on the search for the answer using logic and reason. Kreeft (2004) argued that to answer such a question we have two options: either our core moral values are objective or subjective, they are discovered as scientists discover the laws of physics, or they are created as the rules of a game or a piece of art. He also noticed that pre-modern cultures believe that core moral values were objective and it is only in recent times that society started to believe that those core moral values are subjective, human-made and can be changed over time. The latter scenario is what is called moral relativism, a very common and dangerous ideology in modern times.
Regardless of which option we believe, there are significant consequences to achieve a good set of moral rules for Data Science. For instance, If we believe moral values are objective, we should “find” them, but if we believe they are subjective, then we must “create” them.
For Data Science, the practical implications are that we should take a position about ethics regarding objective or subjective moral values. If we decide that moral values are objective, we should identify the unchanging core moral values and build our analytics ethical practices around them. In contrast, if we say that our moral values are subjective, then we need to create those moral values and agree on using them among the community.
Each of them has challenges but we know only one must be true. Subjective Moral values immediately move us into a dangerous moral relativism, which can be abused by interested groups and expose adoption issues since not all interested parties might agree on them. On the other hand, Objective Moral values present the challenge that, in order to not fall in a subjective approach, these moral principles need to be discovered and cannot be created by humans. They necessarily need to exist independent of us, and because of that, they present the benefit of being unquestionable and provide less resistance for adoption. Hence, the search should take us into a metaphysical research and inquiry.
I want to propose that it is in this metaphysical inquiry that we will not only find the necessary objective ethical standard for our Data Science practice, but a beautiful and fulfilling encounter. An encounter that will transform our lives and provide clarity on topics as complex as Ethics in Data Science. Let’s search for those core moral values and meet the One that provided them!!!
References
Kreeft, Peter. Ethics: A History of Moral Thought. Recorded books, LLC, 2004.