Is NLP almost at human-level accuracy in picking up on subtext? Can it formulate a human-level response to sarcasm? Could it respond in kind? Does it know whe

Is NLP almost at human-level accuracy in picking up on subtext? Can it formulate a human-level response to sarcasm? Could it respond in kind? Does it know when a question is rhetorical? :)