... Videos speak a million words to a million people where a jury speaks to only 12...
So you do advocate American Idol mob-style voting instead of a jury trial? Fair enough, but I'd personally stick with the time-proven concept.
As for the frequency of videos... it is entirely possible that the public display of those videos creates a false impression about the frequency of such events, due to an availability bias.
I remember a quiz question in a class on decision making: "Which cause of death is more frequent, car accidents or heart attack?" Typically, a dominant answer in a class of 80 people is "car accidents" (wrong). The reason being that we all read in the media about almost every car accident that happens in our neighborhood, but only about heart attacks when it concerns someone famous. In other words, we are biased by the information available to us.
Similar thing with those videos. Only those that portray something shocking become viral. When a man bites a dog is shocking, not the other way around. When a policeman shoots someone unarmed and black, that is shocking and goes viral. When a policeman shoots a white unarmed guy, nobody notices, and it doesn't go viral. When a policeman is shot on duty, it is just "normal," that is what we paid them for, and it doesn't go viral. When police risk their lives to save somebody, it doesn't become viral.
So, as a result, the dominance of that one type of viral videos creates the impression that that type of behavior is the norm. While in reality, it is just the fact of life that shit happens sometimes, and in the grand scheme of things, it isn't a proof of "systemic racism" but simply a proof that noise happens no matter what (to return to the photographic lingo).