I would say the loudest voices and it doesn't have to be very many of them at that.

actually it's the loudest voice with the best appeal to emotion that gets to define what things mean.

I would also add that sometimes the voice that "sounds the most reasonable" while requiring low cognitive reasoning.(Which in turn, turns into your 'loudest voice'). Notice that this means that that voice might not be correct, it's just that often people don't bother with good explanations of what/when/etc., they just whats something that validates their own view(which sadly often times leads to a negative outcome, especially when that view has been influenced before).

I'd contend that BigTech commissars currently get to define it, as they also define volume of a voice. They are more powerful than governments, and due to their downstream effects, they can bully even BigMedia.

There seems to be a recursive aspect to this. Accusations of racism are used to silence (censor) voices.

From the article:

Charges of “racism,” for instance, are primarily deployed against the political opponents of upwardly-mobile, highly-educated progressive white people. Even to the point of branding prominent black or brown dissenters as race-traitors (despite the reality that, on average, blacks and Hispanics tend to be significantly more socially conservative and religious than whites).