The Viral Internal Google Memo

[J Boogie]

Seeing racism where it doesn't exist isn't a thing, unless you are talking about "reverse racism." On the contrary, white people tend to NOT SEE racism where it exists, because it doesn't impact us. And because we don't see it, we think it doesn't exist, even when folks who do experience it tell us about it. It's called White Fragility, look it up. The same thing happens around sexism.

Clearly, you must acknowledge it is possible that someone might mistakenly chalk something up to racism when another factor is actually in play.  Many of these tragic police shootings are a good example.  Many of them involve a black man who is clearly not a threat to the officer, and many of them involve a black man who is.  I can't take anyone's views seriously if they say it's always the cop's fault or it's always the dead black man's fault.  If a person is not interested in the facts of a given situation, then I'm not interested in hearing their exercise in confirmation bias. 

Regarding Trump supporter's popularity in the media, yes they were featured prior to his victory as well.  Not because we are racist and would rather see white people express their opinions on TV, but because they are a spectacle and make entertaining TV.  John Kasich supporters were pretty white as well, but generally pretty reasonable boring people in comparison - not good TV. 

Do you really, truly think Trump supporters get airtime because most viewers are unconsciously racist and networks have explicitly or implicitly figured this out, or do you think it could be because most viewers just want sensationalism?

[boy_bye]

There are so many unexamined assumptions in this that I hardly know where to begin.

"Bias is in the data" has been used as a justification for oppression throughout history. I suggest reading through some beginners' texts on the topic of unconscious bias -- clearly it's a topic you have no understanding of.

Look, I did my PhD in AI/machine learning and I work personally with a team in a big tech company focused on bias in machine learning at the forefront of research, and I'm telling you that bias occurs naturally when training a model for high accuracy, and that bias is difficult to define mathematically in the first place. Whereas you're probably an old woman with an art history degree and minor in women studies working at the local library, who's just cherrypicking old texts because it makes her feel better. And you seriously think you're right?

I'll set aside your goofy ad hominem attacks and just point out that scientists are human beings, operating out of unconscious bias as much as anyone else, not infallible fact machines. That means that even with all your scientific training, you, too, are not an infallible fact machine. You, too, are operating out of unconscious bias.

Your use of the word "naturally" in particular gives me a shiver up my spine, because throughout history up to and including NOW, "nature" has been used as a reason why women shouldn't be able to vote, why black folks should be enslaved, why Jews should be exterminated, and why gay people should be shocked into heterosexuality.

What constitutes "naturally" when we are dealing with AI anyway? It's ARTIFICIAL intelligence. And the limitations of the people creating anything show up in their creations. I don't think a PhD is required to understand that.

And I can feel you wanting to reduce my statements down to an anti-science point of view, and that's not it at all. I'm only saying that science is a human artifact, and since humans operate out of unconscious bias, it's probably a good idea for scientists to be aware of this and have a least a drop of humility about it.

[Tass]

I'm only saying that science is a human artifact, and since humans operate out of unconscious bias, it's probably a good idea for scientists to be aware of this and have a least a drop of humility about it.

Scientist +1

[FrugalToque]

Even though I try not to wear my moderator hat where I've commented, this is getting nuts and way off-topic.

Locked.