There are parallel issues that muddy this debate where it concerns Facebook specifically (CA's self-reported extortion/blackmailing scheme) and issues that conc

There are parallel issues that muddy this debate where it concerns Facebook specifically (CA's self-reported extortion/blackmailing scheme) and issues that concern Facebook, but were/are standard practice, and widely requested by FB publishers, including media companies[0] (CA seeding the social graph with users of a non-election related app) This tweet is sneaky. Users of the OFA app were voluntary; their friends weren't, and microtargeting of friends was of much greater value to a campaign (true believers are guaranteed, it's the undecided/uncaring who can move the needle) But that's speaking of ethics; under the terms of service all these actions were voluntary even for people who were never notified of what their friends had agreed to. The one violation of FB terms may have been CA continuing to use the data after FB requested its deletion. Still, as in parallel construction, once you have a good data set there are many ways in which you can reconstruct it and safely delete the original data. Even if OFA did as FB asked and didn't keep data after the terms of service changed, they still had the knowledge, and it can be reconstructed (then that's another debate, which we may never get to until something flashy happens involving parallel construction) [0]: Quiz time - if you wanted to collect as many third-party web trackers, what websites would you visit? To make it harder, you can only visit 5 sites - which ones?