I get your point, but the questioner wasn't being rude or angry, only curious. I think it's a valid question, too. While it isn't as important to be neutral in this instance as, say, a crime prediction model or a hiring model, it should be boilerplate to consider ML inputs for identity neutrality.