I actually was the one who asked this question on another student’s thread, but I’m not really sure about a solution. The fundamental requirements should be the opposite of the three characteristics of WMD, which are 1) transparency, 2) fairness, and 3) unscalability.
- Transparency can help a model not become a WMD because those who make models less likely manipulate the process in favor of certain people.
- The damaging trait of WMD is what makes WMD weaponized, so ensuring the fairness of a process not to damage anyone could make it more ethical.
- If a model is scalable to a huge population, this might mean the model is excessively standardizing people’s patterns. Rather, personalizing a model as much as possible would be more acceptable to everyone, without leaving anyone out.
Furthermore, keeping a model updated is likely effective, since the model can be in the most suitable form according to the feedback it receives. In order to make it sure, getting a model comprehensible also seems crucial because it can enable a tremendous population to assess models and give feedback from various perspectives.
Thoughts on class
Overall, I really like the class, since it require me to reflect on complicated issues and ideas, which I believe is spurring and improving my intelligence. I feel our class is more like “Mathematics and Social Justice” rather than “Mathematics for social justice” because we’ve been focusing on the relation between mathematics and social justice. Hence, although the class might be like that as it goes, I may want more portion of real-world math. Thank you for reading!
Chris's Blog for LC 
Hi Chris, can you give an example of what you were thinking for more “real world math?” We still have plenty of time to focus on other things!
LikeLike
Chirs, I like how you are against the characteristics of WMDs to prevent a model from a WMD.
You mentioned, “personalizing a model as much as possible would be more acceptable to everyone,” but I have trouble with understanding why personalizing a model can prevent the model from becoming WMDs. Can you give me an example?
LikeLike
It really depends on a case actually, but let’s say a teacher is assessed on their performance. I guess often times they just get assessed based on standardized criteria, but I would suggest seeing what their strengths and weaknesses are, and evaluate how the teacher is worth being in a school. Here, although the common model and my suggestion seem similar, my idea focuses more on the teacher’s own characteristics. Hence, the characteristics that we use when measuring change depending on each teacher.
LikeLike
I also think it’s important to have good scalability and personalize each model together, but I guess it’s really hard because using a model makes things efficient but basically ignores exceptions. Then, do you have any ideas to balance well between them?
LikeLike
I may argue we should focus more on personalizing models like the cooking model since we less likely omit characteristics that should be considered.
LikeLike
I mostly agree with how you try to make the best model by keeping the model from having three characteristics of WMD and how you emphasized the importance of constantly updating the model. However, I think the bigger scale the model becomes the harder it would get to constantly update and improve itself based on the feedback it would get. I know this is really hard and even I do not come up with any ideas, but I just wanted to know how you think about it.
LikeLike