Well, the system would tag it with attributes like - political, Democratic, Biden, Flag.
Then, the system would see the pattern of image placement and the demographic profile of those who share and interact with it.
The engine likely has probabilistic scores for social desirability of content based upon this and a multitude of factors.
Now, you make a valid point - the engine will take a much longer time to demote this content allowing messages to reach critical mass before they censor it.
Well, the system would tag it with attributes like - political, Democratic, Biden, Flag.
Then, the system would see the pattern of image placement and the demographic profile of those who share and interact with it.
The engine likely has probabilistic scores for social desirability of content based upon this and a multitude of factors.
Now, you make a valid point - the engine will take a much longer time to demote this content allowing messages to reach critical mass before they censor it.