|
|
|
|
|
by lolinder
996 days ago
|
|
If Facebook could have a false positive rate of 0% and a false negative rate of 0%, they would absolutely make that happen. Unfortunately, due to the way statistics work, Facebook can pick its false positive rate or its false negative rate, but it's impossible to get to 0% false positives without just giving up on moderation altogether. We're not talking about capital punishment here, we're talking about social media, and Facebook appears to have made the very reasonable decision that it's worth accidentally rate limiting some innocent accounts in order to keep spam lower than would otherwise be possible. |
|
That sounds reasonable.
Well, reasonable unless you happen to be (or care about) one of the innocents being accidentally punished, because some corporation's algorithm said so.
Scholars have been worrying about this kind of thing in the real world's justice system for a long time (think of William Blackstone's "it is better that ten guilty persons escape than one innocent suffer" quote, which is over 250 years old[0]; of the 1895 U.S. Supreme Court's "it is better to let the crime of a guilty person go unpunished than to condemn the innocent", and these weren't novel ideas, they go all the way back to the Romans).
Where are the checks and balances for the online world?
[0] https://en.wikipedia.org/wiki/Commentaries_on_the_Laws_of_En...