I don't disagree with the general point but in this case we're looking at what (seems to be) a blatant copyright violation. It would not be any more or less of a violation if the infringed license had been a more or less permissive one, because the license has not been followed.

Sure, the MIT is very permissive so it's very easy for Microsoft to correct their repository so that it's in compliance for the future, but they cannot correct the past. (Unless the original authors allow for it.) The MIT license, being so short, does not have a provision about curing infringements.

So Microsoft seems to be ok with the risk of being sued for infringement etc. That's not something you can correct with your personal decisions as author.