I also maybe agree with “very very far away”, in the 20ish year range. Farther than some people think, closer than others do.
If and when we get to a place where AI reaches that holistic facsimile, I’m not sure what I’ll think of humans who reject the idea and insist that we are qualitatively different because (insert biological or spiritual rationale here). I suspect it will feel like seeing someone mistreat a call center employee because they happen to be in India, or sound like a disliked minority.
Almost as a matter of definition by the time AI can provide such a connection there won't be anything distinguishing it from an actual human. Which is to say, what you're implying there is an artificially manufactured but fully functional human.
One reason is that the human experience is dependent upon the biological nature of man. The biological systems color the experience. The pumping of blood, the nervous system, the heart beating, and ultimately, one's awareness of the specific type of mortality inherent to biological organisms, are integral to the experience. If you accurately reproduce that experience then perhaps you've simply made a human rather than a machine. Of course that claim spurs many subsequent philosophical arguments. Ultimately though, a video game console emulator is not the literal console no matter how accurate it is.
A second reason is simply the subjective experience of a person. Regardless of how accurate the simulation is, ultimately, if the person is aware the other end isn't human, the experience is tainted (for better or worse depending on the individual's opinion - but tainted nonetheless). Knowledge of the truth will necessarily affect the experience. The alternative - being in the dark or outright deception - raises other questions of genuinity that taint the experience.
A conversation with a human, by another human, will never be the same as a conversation with a machine - by definition.