My feelings about ROOST keep changing and I'm still not sure where I land: using LLMs to categorise CSAM instead of human moderators is far better ethically and much easier to implement at scale, but I am always sceptical when Big Tech comes together to offer a solution to a problem they created
Comments