I have to wonder if they're going to catastrophically fuck this up even more than it already is by turning the woke audience against them, by trying to employ some sort of filter that attempts to determine the race of the speaker. This would then be used to try and detect if a black person has said the nigger/nigga word and allow them to say it. because it's magically only okay for those of African descent. This would then fail, obviously, not being able to properly detect anything via lack of proper training or just the limits of current machine learning, and people would manage to trick it and say the n word regardless of race. I can't see a company like Intel managing to tard wrangle themselves enough not to try something even dumber than making this in the first place not trying this too.
Outside of this hypothetical, I just can't see trying to limit voice chat of all things with machine learning being anything more than a spergfest whack a mole. Will it automatically filter slurs from other nations/languages other than english? What if they overlap with regular words in english? When people start making new 'offensive' words to get around the filters, how far will intel go to retrain the ai to try and stamp those out too? I can just see this kind of program going ham, filtering random shit that vaguely sounds the same as faggot, gay, homo, and so on. This whole development is going to be a massive waste of time on Intel's part. Is there something I'm missing and the cyberpunk dystopia of sapient ai filtering the world has somehow come to pass?
