Once AI is used as the buffer between all forms of communication (making sure the emails I send put me in the best light, filtering incoming email to make me more productive, generating web content, reading web content in a way that makes me more efficient, even voice and video calls, the nation state can not only monitor, but also subtly alter all communications, in real time, and in context to guide/enforce how people interact/communicate and think.
This is already happening in areas such as search, getting "answers", and content generation. Your information is constrained/mapped into what the the AI deems appropriate.
We already know that the NSA taps into the communication backbone of the internet. Why not add a real-time AI monitor/filter to best manage how we all communicate ?