While I can kind of see what you're getting at, I feel that it is not 100% correct, and only a small piece of the story. When we briefly studied the Diffusion of Responsibility we examined a few cases. The first was one of the most famous in regards to the phenomenon (according to my textbook anyways) in which a young woman had been beaten and murdered on the street in a mugging turned sour type situation. This in itself isn't anything extraordinary, but the interesting bit comes from the fact that there were several witnesses to the crime, I believe it was like 6 or 7 (don't quote me I can't remember for sure xD). However, the witnesses could also see each other, as well as the event unfolding. When asked by police why they hadn't called the police or ambulance or something, nearly all of them referenced the sight of other people as to their reasoning why they didn't need to. "I thought the person in the building across from me had/would/.was calling for help, so I didn't bother."
Another similar case we studied was much more controlled. They had set up an experiment in which several people would participate in a interview questionaire type thing over the phone, with several other people on the line. They would be aware of the others on the phone with them, but could not hear them. Then the person whom was supposed to be administering the survey would feign some kind of tragic event (choking, or someone coming into the room they were in who was attempting to harm them or something like that.) Long story short, without any other contact with the other people in the experiment besides KNOWING they existed and could ALSO hear the events, the rate at which people actively went to seek help for the test giver (going to tell someone about it, attempting to dial 911 or something) was significantly lower, than in the control experiment, where no one was told other people could also hear what was going on (so the participants effectively thought they were talking 1:1).
A third and final experiment we studied, which I feel is relevant to this discussion again involved someone falling to harm of some kind, and people around to witness or hear the event. In this one, a fake meeting was organized in the workplace, with a man whom everyone passed on a ladder, falling and yelling while everyone was inside. In the trials which the person leading the meeting ignored the event, and continued talking, no one moved or went to go help the man. When someone got up to go look, everyone followed, and in a third set of trials, when one of the participants was alone in the boardroom (waiting for the rest of the people to arrive) they again, often went to go investigate.
Combining all of these, leads me to the conclusion that personal safety and "social status" do indeed play a role in how diffusion of responsibility manifests itself in various situations, but also tells me that it is a very minor one indeed. In several of these cases (and many others) the only thing the people being studied needed, was confirmation that other people were also witnessing/experiencing the event. This speaks volumes about how we act and react based on the actions of others. This was touched on lightly in the video you had posted. Based on my prior experience with the phenomenon it seems a bit rash to jump to the conclusion that it is because of a worship of social status that this occurs, although as I said, it definitely plays a part.