This. More precisely, AI Ethics is really about explicitly stating exchange rates between your assorted principles. E.g., if 1% more racism is as bad as 3 women raped, then instead of computing argmin(rapes), you compute argmin(rapes/3+racism/1%). argmin taken over policies.https://twitter.com/Plinz/status/1052245394627936256 …
I don't think that it can work like this. Racism is not an action, but a disposition, and can therefore not be compared to rape, and it only gets more complicated from there.
-
-
So by "racism", I mean "some particular disparity that activists describe as racist". Just hard to fit all that in a tweet. E.g., a racially correlated coefficient being non-zero, or a disparity in false positive rates, or something like that.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.