ALG Blog 1: Exception to Data Driven Rules

The Exception Exists: Data-Driven Rules in Our Society

News Article:
The Right to Be an Exception to a Data-Driven Rule

Purpose

I believe the purpose of this case study was to help us see the pros and cons of data-driven systems and how exceptions will always be present. I think we are also meant to understand the severity of some of the areas we apply these data-driven rules to, like in our justice system, and how exceptions are handled in such cases.

Discussion Questions

1. What is a data-driven rule, and what does it mean to be a data-driven exception? Is an exception the same as an error?
   • A data-driven rule, from my understanding, is a rule created using data to predict outcomes. A data-driven exception would be anything that doesn’t necessarily fit into that rule, meaning you are outside the reach of the predicted outcome. For example, I grew up in North Minneapolis, which is a neighborhood many would say is very dangerous and full of crime. However, I have never committed a crime. In the eyes of the police, I would be a data-driven exception because I am outside the assumption of a risky pattern of crime. However, I wouldn’t say an exception is the same as an error because it is simply OUTSIDE the rule, not necessarily proving that it’s wrong.
2. In addition to those listed above, what other factors differentiate data-driven decisions from human ones?
   • Data-driven decisions aren’t very flexiable and open-minded. Humans apply empathy, reasoning and fairness towards decision making. The case study discussed the death penalty as an example. It used to be a mandatory consequence to some crimes. It was a black and white decision, no grey areas. However, many people resorted to empathy and reasoning, taking into account outside circumstances data-driven decisions wouldn’t acknowledge. Life isn’t just black and white, although sometimes we wish things were that simple.
3. Beyond what is discussed above, what are some of the benefits and downsides of individualization?
   • A benefit to individualization would be personalization and getting what is best for YOU. For example, your “For You” page on TikTok. I’m glad I don’t get recommended the same videos as my sister because her humor is very millennial… However, a downside to this would be invasion of privacy. Like maybe I’m okay with my “For You” page not knowing about the conversation I was having with my boyfriend. It’s very scary to see an ad for something we were discussing over dinner, like what else do you know about me???
4. Why is uncertainty so critical to the right to be an exception? When the stakes are high (e.g., in criminal sentencing), is there any evaluation metric (e.g., accuracy) that can justify the use of a data-driven rule without the consideration of uncertainty?
   • Uncertainty is critical because it accepts that these data-driven systems cannot be perfect in the presence of an exception. In high-stakes areas like sentencing a criminal, we should not use metrics dependent on data-driven systems because of these exceptions. Like the case study mentioned, an algorithm that is 95% accurate does not account for the missing 5%. When discussing matters like human lives, that 5% is crucial. We shouldn’t use data-driven metrics to make final decisions, rather we should use them to reinforcement or give insight to our decisions.

New Discussion Questions

• Since exceptions will always exist, how can we design data-driven systems that acknowledge them while still doing what is best for everyone?

Reflection

This case study really gave me a new perspective on how our society works and how flawed some of our systems are. I listen and watch a lot of true crime, so the topic sounding the criminal justice system made me think. Lots of old cases end with the death penalty, and although my stance on that punishment varies, I’m glad it is no longer being used as a mandatory minimum. The discussion on accuracy was also eye-opening. Many tend to dismiss the 5% when a system is 95% accurate, but the 5% still exists. Like yes, duhh, but are you ever really thinking about it? Anyway, thank you for reading my blog! Keep thinking about that 5% of air that’s in your 95% full chip bag. It matters!