News for hackers. Stuff that matters.
Re:Identifying and Eliminating CSAM in Generative ML Trainin... by Footnote7341 on Wednesday December 20, 2023 @05:26PM (#38714469)
Such a vanishingly small percentage of the images its not even worth calculating. Of course search engines also contain these links, LAION-5B only contains links not images as well....

protect the kids!!! or something. Can we do a scandal about how many 'extremist' images are in the data-set next too? anti-vaxer, climate denier, nazi, religious extremist propaganda, scientific misinformation. Maybe we're all safer off using corporate models with closed data sets so no one gets any of the wrong ideas.

Re:Identifying and Eliminating CSAM in Generative ML Trainin... by WheatMillington on Wednesday December 20, 2023 @05:58PM (#38714783)
For the children who have been victimised, I doubt the small percentage provides any comfort.
Re:Identifying and Eliminating CSAM in Generative ML Trainin... by ipaddr on Wednesday December 20, 2023 @11:57PM (#38717152)
The victims have already been victimized many years ago. Finding this content in a training set doesn't re-victimize them.
Re:Identifying and Eliminating CSAM in Generative ML Trainin... by 127361 on Thursday December 21, 2023 @09:00AM (#38720447)
Yes, and it's better we spend the effort on preventing the physical abuse from happening in the first place.
Re:Identifying and Eliminating CSAM in Generative ML Trainin... by nulld3v on Wednesday December 20, 2023 @06:15PM (#38714915)
Wtf, if you can get rid of the images why wouldn't you? Clearly it's not an exorbitant amount of work as shown in the paper, they use very conventional techniques.
Re:Identifying and Eliminating CSAM in Generative ML Trainin... by colechristensen on Wednesday December 20, 2023 @06:10PM (#38714885)
CSAM does not fall under “freedom of expression”, everything else you list does. It’s not a slippery slope.

Resist anyone’s attempt to make everything else you list illegal to express or possess.