it’s been a while since i looked into any of this but my understanding is, it isn’t nearly as bad as it’s made out to be when ppl say “AI is scanning your images”

there have been ways to scan media for illegal material since long before AI hit the market which are also usually cheaper than using AI

iirc they aren’t using AI for that, but they use images’ hashes, comparing those to a database of known illegal material. so nobody would be seeing the images being sent, just a hash generated from the data in the image. if the hash matches a known bad image, it gets flagged, if i understand right

my understanding is that the main instance (fluxer.app) uses that because they legally have to have some kind of moderation for illegal stuff like CSAM (discord does something similar)… but i think a self-hosted instance wouldn’t have that same setup, at least by default? don’t quote me on that tho

it is so stable these days