This is more fantasy thinking (but hey...all the 20th anniversary Matrix articles are popping up...) At what point do machine learning "Adversarial" attacks co

This is more fantasy thinking (but hey...all the 20th anniversary Matrix articles are popping up...) At what point do machine learning "Adversarial" attacks converge with real life "adversarial attacks". Or do we already have this, but it's called "camouflage"?