"Surveillance" is in the intent. If I collect user tracking data and that data leaks, or the government requisitions it, does it not then become surveillance data?

FWIW, I have convinced myself that ubiquitous surveillance is pretty much inevitable as a side-effect to automation of infrastructure, so the question is how to manage it (rather than whether.) Right now we are seeing a split between "Morlocks" and "Eloi" where there is a huge gap between the understanding of the folks working at these AI-developing super-silos and the understanding of Joe and Jane User on the street.