I don't understand how the author of this article thinks this could be applied to an audience or protestors. For this to work, the audio output to each individ

I don't understand how the author of this article thinks this could be applied to an audience or protestors. For this to work, the audio output to each individual person has to be exactly what each individual person is currently saying. If it's a combined audio mess of an entire audience's voices, it's not exactly going to screw up anyone's speech patterns. So the only way this could work in an audience would be if the back of every chair in front of people had one of these devices on it which could filter out the noise of everyone else around that person and only output their voice.