The author, Julia Evans, describes how numerical computations suffer from precision issues, "computers cannot calculate most decimal values exactly," and how these result in "small errors with large consequences." These concerns also relate to our prior discussion on utility calculus. How much confidence should we have in machines being able to perform simple arithmetic if they are subject to precision loss? Therefore, given how difficult it is to rely on computers for even the simplest of mathematical operations (the building blocks of data-driven ethics), quantifying something as abstractly subjective as human well-being appears to be an even greater challenge than previously thought.

The passage describes how utility calculus can be done effectively. However, utility cannot be quantitatively measured, as noted in the textbook, which defines utility as a “sort of combination of happiness and well-being.” Because there is no single unit for measuring utility, you cannot measure a person’s grief and another person’s satisfaction and sum them up into an effective or meaningful quantity. In a section on data-informed ethical decisions, the inability to measure the basic premise of the approach is a serious issue: because the basic premise of the approach cannot be measured, we are producing the appearance of rigorous morality rather than actual rigorous morality.

In Phaedrus, Plato foresaw this issue; He prophesized that writing was dangerous since it leaves the written word permanently disconnected from its author. A donkey is just an old-time representation of that; Bots currently represent the ultimate extreme of an ancient technology. Plato didn't have a problem with technology but rather losing accountability through its use.

This was an unsettling example for me. To realize that a message can travel around the globe without any connection to its creator, and that the carrier (whether it be a donkey or a bot) does not know the message, leads me to believe that the chances of holding someone accountable for their actions based on a message they have sent out into the world are statistically negligible. I also started to think about how much I have probably interacted with bot-generated information online without ever realizing it. At what point do I begin to feel deceived when interacting with a bot even when the content itself may be true?

It is certainly true that technological advancement is happening at a much greater speed than society and the law can adapt. Technologies such as deepfakes, bots, and recommendation algorithms are capable of affecting millions of individuals prior to regulatory bodies creating guidelines to govern their use. This has led me to believe that companies should be developing new technology with an eye toward the ethical implications it may have on its users, rather than relying upon subsequent laws or regulations to govern how they are used.

Utilitarianism is another theory associated with Consequentialism; a version of Consequentialism focused upon maximizing overall happiness or general well-being. Social Media and technology are common areas in which utilitarianism is applied due to the fact that many companies will defend their choices using claims such as "my platform supports/benefits millions." The problem with this form of reasoning, however, is that it may support causing harm to a small group of people while benefiting a large one. For example, features designed to create user engagement could lead to an increase in harassment and misinformation.