Adversarial learning is like an automated turing test. This means theory is ready to pass the turing test, we just need a to make discriminators larger and stronger than humans.