Re:Stable Diffusion 2.0 Release
by Hamuko on Thursday November 24, 2022 @02:34AM
(#33728682)
"Artificially-generated CSAM" is a misnomer, since it involves no actual sexual abuse. It's "simulated child pornography", a category that would include for example paintings.
[ Parent
| on HN ]
Not exactly, since the abuse needed to actually happen for the derivative images to be possible to generate.
Is Stable Diffusion only able to generate images of things that have actually happened?
Hmm, that’s a good point. It seems to be able to “transfer knowledge” for lack of a better term, so maybe it wouldn’t need to be in the dataset at all…
Very much this. If someone goes out and trains a model on actual photographs of abuse, then holy shit, call in the cops.
If someone is generating sketchy cartoons from a training set of sketchy cartoons... well, gross, but there's no victims there.