This a pointless nearly unenforceable rule to make people feel better. Sure, if you generate something that seems like a real event that is provably false you can be caught, but anything mundane is not enforceable. Once models reach something like Sora 1.5 level of ability, we are kind of doomed on knowing whats real in video.
>This a pointless nearly unenforceable rule to make people feel better.
Pretty much. If Google says "Swiper no swiping" they can point at their policy when lobbying against regulations or pushing back against criticism.
Before surveillance capitalism became the norm, web services told users to not share personal information, and to not trust other users they had not met in real life.
naah, there still will be certain patterns and they will be recognisable.
once something sora 1.5 level of ability is there – definitely reverse-sora model which can recognise ai-made videos should be possible to train as well
