Also, I would like to point out three things to folks who are asking "What about personal responsibility? What about agency?" here and on the article comments.
Facebook's shadow profiles take away any 'agency' and 'personal responsibility' out of the equation. While I don't yet know Facebook using shadow profiles to make someone's life worse off, their penchant for saying "Screw it" to any ethical considerations makes me feel its only a matter of time before even that crutch goes away.
Secondly, there is the complexity. You cannot take personal responsibility for something that you don't fully comprehend. At the same time, we are at a point where AI produces data insights that even the AI builders cannot reason about. For example, how does someone take responsibility for clicking the Like button on a post and then finding out that most of the people who liked the post were (say) closet anarchists which was then correlated with them being the best targets for some kind of subversive advertising from (say) the NRA?
Thirdly, there is the asymmetry involved in "moderate use". It is possible for Facebook to create "automated notifications" implemented at a very tiny cost to them in terms of their resources/profits while the effort needed on your part to make sure you don't turn into a Pavlovian "notifications" dog comes at an extremely high time cost to you.
With shadow profiles, they are being forced to give their data to Facebook or live like an outcast.
How intrusive are the shadow profiles actually? We will know only if their data collection process (as well as the data collection process for all these ad-first companies) is completely and thoroughly investigated.
Facebook's shadow profiles take away any 'agency' and 'personal responsibility' out of the equation. While I don't yet know Facebook using shadow profiles to make someone's life worse off, their penchant for saying "Screw it" to any ethical considerations makes me feel its only a matter of time before even that crutch goes away.
Secondly, there is the complexity. You cannot take personal responsibility for something that you don't fully comprehend. At the same time, we are at a point where AI produces data insights that even the AI builders cannot reason about. For example, how does someone take responsibility for clicking the Like button on a post and then finding out that most of the people who liked the post were (say) closet anarchists which was then correlated with them being the best targets for some kind of subversive advertising from (say) the NRA?
Thirdly, there is the asymmetry involved in "moderate use". It is possible for Facebook to create "automated notifications" implemented at a very tiny cost to them in terms of their resources/profits while the effort needed on your part to make sure you don't turn into a Pavlovian "notifications" dog comes at an extremely high time cost to you.