I guess since it's their network, they have the right to police it any way they want to. If you don't agree with it, you don't use them (harder to do in areas where there's only ONE provider).
Deep packet scanning I've not heard of before or know of any way to determine that an MP3 that I've sent to someone includes a song or my voice.
Scanning and blocking of destinations has been done for years, and there are active lists built based on the IP number and domain names of places that house material (which is where many of those "cease notices" get their data from.
Wonder how they'll do it and how long before others jump it and start to use the data for more than just detection of activities. It'll be used similar to cookies so they can issue advertisements and targeted material.