Technology like this is stuck in a no win situation.
Human police officers are allowed to take action based on reasonable suspicion.
AI, if it is worse than humans, will be criticized for false positives.
And if it is better than humans, without being able to explain why it will be criticized for that as well even if it is in general better than humans, it will be criticized for the false positives.
But as the saying goes, this is the worst it will ever be. At some point this thing or something like it will have a precision over 95%. At what point will the debate become awkward because the thing is superhumanly efficient at sorting out the baddies?
Technology like this is stuck in a no win situation.
Human police officers are allowed to take action based on reasonable suspicion.
AI, if it is worse than humans, will be criticized for false positives.
And if it is better than humans, without being able to explain why it will be criticized for that as well even if it is in general better than humans, it will be criticized for the false positives.
But as the saying goes, this is the worst it will ever be. At some point this thing or something like it will have a precision over 95%. At what point will the debate become awkward because the thing is superhumanly efficient at sorting out the baddies?