The goal is to detect internal irregularities and breaches that are signaled by divergences in normal patterns of behavior.

I was just talking about this myself. It's very limiting to define behavioral patterns that way, and in fact plain wrong. I would love to see their proof that deviations between normal (which is to say more or less average) behaviour is an indicator of anything negative, especially related to company work.

This has implications for our personal right to privacy, even at work.

I think there is a kind of assumption on the part of some employers that privacy does not exist in the workplace. This is a really toxic idea. The bargain is just too great, but as you say, when faced with accepting privacy intrusion at the cost of the job, I fear most will choose to just accept it.

I think this should be roundly resisted.