You're right about the theoretical distinction between data collection and surveillance, but in Facebook's (and Google's) case the distinction isn't meaningful because most people in the US have a Facebook and a Google account.

Open protocols would make it harder for law enforcement to act on data collection because there wouldn't be a single entity they could leverage to have done the legwork for them.

I'm a prominent Friendica figure, and yet no amount of law enforcement pressure could get me to hand out dat about any given Friendica account because I simply don't have access to it.

Open protocols means that individual servers could choose to collect as much data as Facebook or not, and users could choose their home server based on that criterion. There isn't such freedom with Facebook.