i'm working on a chapter about synthetic data and democratic processes this week and reading some truly distressing horseshit from people who should know better, so i'm taking to opportunity to plug some of the previous stuff i've written on this topic:
in this 2017 piece i discuss why big data extrapolation techniques encourage the exclusion of actual people from representative processes like elections and constituent consultations, a thing i am seeing people actually recommend using LLMs for, the motherfuckers web.archive.org/web/20240121...
i feel you. thanks for linking your work!
Thanks for these articles, doing some civic assembly workshops and only just encountered AI and democracy groups who seem to think that listening to people is somehow irrelevant or unnecessary now