You work For Amazon Web Services. Part of your job is to provide HPC facilities to your customers. You notice that some customers use the resources extensively for some "illegal" purposes, and you want t o stop them. So you build a model that can detect, based on the activity of a user, if the user is using the resources properly or not. Which one is worst for such a model, FN or FP or both, considering these facts: AWS is the number one platform of its kind in the world. It is famous enough. They don't need to acquire new customers. At the same time, they have customers they never want to lose, as those customers bring millions of dollars to the platform. So, they really don't want to tell a valuable customer, mistakenly, they are doing illegal work.

Assume a "nice" user as "negative" and an "illegal" user as "positive"

a) FP is more important as we don't want to tell a user you are doing fraud while they are not;

b) FN is more important as we don't want not to predict a user is a nice one while this is not true;

c) Both;