From what I understand, this is an example of the issues of algorithms not being able to understand context. We would obviously want to discourage animal cruelty, but graphic depictions of animal cruelty from an anti-cruelty perspective also get the crackdown. It could lead to a scenario where bad things continue to happen just because any depiction of them is filtered as advocating for it.

At the risk of drumming up unneeded controversy, it reminds me of when Marine Le Pen in France tweeted graphic images from ISIS actions as a rallying cry against them, only to be charged by the state with circulating violent content inciting terrorism ( https://www.theguardian.com/world/2018/mar/01/marine-le-pen-... ). Without proper application of context and judgment, a terrorist spreading propaganda and an anti-terrorist illustrating the terrorist's propaganda would be seen as the same sort of thing.

Of course, there's no simple solution -- we wouldn't want a world in which someone could get out of responsibility for spreading needlessly graphic content just by saying "oh I was just spreading it for educational purposes!".

Can automated processes ever hope to achieve the nuance we need in our societies?