The power of having 350k Twitter followers

I just wonder why this escalates - there is a CreditScore System in the US that is based on how much credit you use. We are here in the US for four month and my score is much higher than my wife’s because I use the card more for travel etc

If you read the whole thread, that seems to be unlikely to be the correct explanation in this case.

Don’t see anything against it. He only says that his wife has a higher TransUnion score but this is only one part. My TransUnion info shows me the score, my card limits, payment history, # of accounts-cards and card utilization.

So if I would program a algorithm that creates a card limit for a person I only have the SSN and the TransUnion info, I would use exactly all the information to generate a limit. Especially the existing card limits and tx history.

By the way - it was absolutely normal in Germany to take the gender as THE central criteria for financial products until 2013 when EU regulated this with Unisex Tarifs. Before 2013 women had much *better* tarifs, as they statistically had less accidents and live longer.

But overall it shows that statistics and ML based algorithms can have very bad side effects without clear regulation. And can definitely imagine them turn into a discriminating problem for groups.

E.g. my wife did not get a credit card at Bank of America where I got one. As she got declined, her credit score/history is even worth now. We both were only a month in the US with basically no history.

So if two people have the same CreditScore of e.g. 850 but person A has a overall credit card limit of $200k and person B a overall credit card limit of $20k reported by TransUnion, I can see why person A would get 10x the credit limit by Apple than person B.

Not entirely sure whether we agree or disagree here; my guess is it’s the former. Gender should not be allowed to play a role because that would be discriminatory, and companies need to be able to explain and justify their algorithms’ decisions.