On the other hand, mining more entities with the same set of attributes decreases the risk, because the rules learned from them will have stronger support.
Ten Myths About Machine Learning
Pedro Domingos

While it is a general truth that more instances for learning leads to better predictions, I think the problem that people are actually concerned about is that in a model of NSA-type mass surveillance, even small false positive rates would lead to unworkable numbers of false positives in a problem like terrorism detection, where the number of true positives is going to be vanishingly small on a population level. Beyond (the major) concerns about privacy, it’s also a problem of utility.

Show your support

Clapping shows how much you appreciated Tom Northey’s story.