Yes, that's exactly what it does.

By shifting it a few points across the boundary, it creates a higher confidence that X is actually Y instead of X.

The problem is that without studying the neural network and trying out different inputs and seeing the results, it's hard to figure out exactly what will shift things across the boundary. Even if you could get the same starting data set and attempt to train your own, you don't necessarily know which data was used for testing vs validation.

You'll likely come up with similar boundaries but not the same therefore you don't know the effectiveness of your adversarial approach until you actually try it against that particular system..

But how do you go from 60% certainty that it is an A to 90% certainty that it is a B?