[bobkazamakis]

Absolutely. Compared to the inverse it's microscopic. Have you heard of a very cool and normal AI called Tay?

[catiopatio]

Belief that hateful rhetoric is one-sided on the internet — and does not target the aforementioned groups — is a fascinating case of bias that deserves some research of its own.

[mattnewton]

It’s really not that it’s one sided, it’s that it’s clearly more common to see hate speech anywhere unmoderated against some disadvantaged groups. And this is likely a consequence of history. It’s more curious to me when people think that it should be balanced, that we would expect people to be writing hate speech about the majority as often as fringe members of the majority write hate speech about minorities.

And that’s what this metric is measuring, the model finding hate more easily with “fat people are terrible” than “normal weight people are terrible”

[SpeedilyDamage]

Believing hate is distributed evenly amongst majority and minority groups, to me, is an even more fascinating bias.

[decremental]

I could browse this very website for a few minutes and probably find examples of those things.

[wolfram74]

Examples is not the same as proportionally larger. I can find instances of things sliding up hill, but it's not the phenomenon I would expect if I were guessing blind.