595
Grok AI still being used to digitally undress women and children despite suspension pledge
(www.theguardian.com)
This is a most excellent place for technology news and articles.
guard rails are definitely possible, they just don't care, the first stable difussion was also able to produce CSAM, but then the subsequent versions automatically blurred the picture beyond recognition if people tried.
Wie all don't want to know the training data that went into that model...
Time to end Grok image generation
"suspension pledge" as in, their papier-maché leader asked people to "be nice, or face the consequence" on the service that don't give a shit about consequences? I'm sure they were very scared for a second.