I address this, but you seemed to ignore it.
its not like people cant understand other perspectives, its not that hard, and focus/user groups would catch all the of examples you cite, just like any QA testing catches bugs and edgecases
And gave examples of that happening. Repeated, continued examples in the same context (facial processing).
>its not like people cant understand other perspectives, its not that hard, and focus/user groups would catch all the of examples you cite
But empirically they don't. Yes, they should. But they don't. This keeps happening, which would imply that the user groups aren't actually catching these examples, or maybe they are, but again without an advocate, a single complaint, or even some complaints, doesn't matter.
So let me ask you, who is more likely to advocate for changes to fix a bug: an engineer who is directly affected by that bug, or one who isn't?
It comes down to needing an advocate. Unless you have someone who advocates for making your products inclusive, you won't end up with inclusive products, and the best advocate is almost always an engineer who says "hey team this product doesn't work for me". (Edit: I should add, this is true for pretty much any feature, not just inclusive ones, but its also true for them)
There's no reason a machine learning team needs physics PhDs. There is a reason that same machine learning team needs dark skinned machine learning PhDs.
>It is also common with people who are more focused about hiring the most qualified people rather than being politically correct or fulfilling quotas.
Right, and given that most people don't want to explicitly state that they're bigoted, often people who are racist and xenophobic claim to hold their views because they care about hiring the most qualified people rather than being politically correct. Perhaps you aren't, but its a worryingly common tactic.