A neural network classifier would surely get less of both false positives, false negatived, ans exploitable vulnerabilities.
But at the end of the say, the question is just if Amazon has any interest at all in stop being accessory to large scale fraud. And the answer is pretty obvious; any of the things people are discussing here is viable, they could pick any or even spend more than 3 seconds thinking about the problem.
But at the end of the say, the question is just if Amazon has any interest at all in stop being accessory to large scale fraud. And the answer is pretty obvious; any of the things people are discussing here is viable, they could pick any or even spend more than 3 seconds thinking about the problem.