> "The more data we have on what happens over the next few days is going to really accelerate our ability to retrain our AI systems and then help all of you accurately understand tenant risk moving forward into this new world,"

This is actually dumb in terms of ML, and really dumb in terms of discrimination.

Given someone missed rent April 1st 2020, what does that tell you going forward about their future? Does it imply they're likely to miss rent again someday? Far less than the same data point any other time before. So it's not useful that way. So why get these data points about people?

Perhaps what it does imply is that they may have poor health? Or a disability? Or that they don't have a university degree? Or (data still coming as to this relationship) that they are more likely to be a specific race?

I feel that what these folks are really doing is trying to have their model learn to predict "won't pay rent" based on a data point that implies all the things they're not allowed to train their model to predict with due to obvious discrimination laws.

Edit 5 minutes later: It could be innocent stupidity too. Hanlon's razor is important.