Even if one were to agree that this would be reasonable, would you say facebook would be liable if the tool failed in some cases? And how do you enforce that, one way or the other?
Algorithms that recognize people in photos aren't (and probably never will be) 100% accurate, especially in this context (people in the background of photos will often be partially obscured, poorly lit, maybe really small/far away, etc.).
Algorithms that recognize people in photos aren't (and probably never will be) 100% accurate, especially in this context (people in the background of photos will often be partially obscured, poorly lit, maybe really small/far away, etc.).