How does this happen? Somehow the software that converts and transmits the scanner's audio instead pulls the input from the computer's mic, rather than the input providing the scanner audio.
It actually happens occasionally with other feeds here, too. Hopefully the volunteer providing this particular feed sees your note and takes action before someone in that apartment says or does something they don't want transmitted to the millions listening to the feed. (That latter part was sarcasm, in the event it wasn't obvious.)