I'd argue it's not even that. AI generated data all has malocious/hostile undertones that affect sanity of both users and viewers. The situation appears to be g

I'd argue it's not even that. AI generated data all has malocious/hostile undertones that affect sanity of both users and viewers. The situation appears to be getting better especially in Chinese open weight models, but it hasn't reached the point where the resultant data neutrally compare against human sanitized data("art").