2 ms·

I mostly agree with the first bit. Re: advice, well, there could be, but the people who put these things in place aren't necessarily thinking in those terms. T

I mostly agree with the first bit. Re: advice, well, there could be, but the people who put these things in place aren't necessarily thinking in those terms. They're thinking about a statistical edge and acceptable negative outcomes on their end and no one else's. They're not maximizing what's good and helpful for you unless it helps them also, they're probably maximizing short to medium term profit seeking. Computers are an amplifier of human bad behavior. See also, "computer says no": https://en.wikipedia.org/wiki/Computer_says_no https://en.wikipedia.org/wiki/Computer_says_no

mjburgess 2y ago
Yes, and I dont think this paper or those who talk about "Algorithmic Aversion" are at all engaged with this. Perhaps unsurprisingly author's i've read on this topic are all from the humanities, or adjacent (here law) and seem to not really to have a grasp of technology as a social and political phenomenon the way modern (aware) tech workers do. Here I think software engineers feel intuitively and immediately that replacing any of their decision making with "algorithms" is a high risk. I suspect papers like this would benefit a lot from assessing the "algorithm aversion" amongst "algorithm experts".