Language once proved a life was behind it, and AI severs that link.

We read machines with reflexes built for human minds.

The problem isn't what AI says but what stands behind it.

In "Social Aims," Emerson wrote, “What you are stands over you the while, and thunders so that I cannot hear what you say to the contrary.” The quote is powerful and often referenced in various iterations. These words suggest, at least to me, something interesting about the output of large language models.

Most of the discussion around AI is framed around intelligence. Is it smart? How smart? Better than here, weaker there? But the more curious feature of these systems may have less to do with intelligence than with ontology. My concern begins when "LLM language" seems to carry so many of the "signals of mind" yet remains detached from the lived experience that gave those signals their human substance.

When Words Carried a Life

Language is attached to life, at least it used to be. We didn't just hear words, we heard a symphony of memory, motive, contradiction, fatigue, hope, vanity, shame, and consequence. And each note plays its role. Even when a person was wrong or insincere, there was still someone there. The words may have twisted reality, but they still emerged from a point of experience.

That is why language has always been more than information. It's evidence of a mind moving through time. A sentence suggested a speaker and that speaker reflected a life. And writing carried the pressure of having passed through a person.

LLMs disturb that arrangement in a way that is both hard to name and then hard to ignore. LLM output can feel thoughtful and even wise. The cadence—those short sentences that feel almost breathless—may suggest reflection. But none of this "banter" is part of our human experience, only a precarious counterfeit that fools many of us.

It's important to understand this. But more importantly, the problem isn't just that the AI is imitating us well. The critical and even stranger fact is that the language can feel inhabited even when nothing like "human inwardness" is there. The words carry the outer shell of thought while remaining ever detached from the lived interior that once made thought feel human.

The Reflex Does Not Stop

I'll argue that our minds don't know how to leave this alone, as we are built to infer a presence from language. We hear tone and imagine temperament. We hear coherence and infer authority. We hear sensitivity and begin to reciprocate.

That reflex remains active even when the old assumptions are being challenged by AI. So, what happens? We read AI with habits formed in a human world. We keep looking for the speaker in the old sense, even while another part of the mind begins to understand that the human component is missing. To me, that's part of the uncanniness. The language feels socially legible and psychologically in tune, but the "being behind the box" doesn't resolve into anything we know how to recognize.

This is where Emerson’s line becomes more than a quotation for me. It becomes a diagnostic. With a human being, what one "is" deepens or is discredited by what one says. The list is endless (and hard to codify for AI) and include a range like character, experience, fear, wisdom, damage, love. All of it stands behind the sentence and gives it force. But with AI, what it is does something different. It does not deepen the words. It destabilizes them.

You can feel this while reading. The output may be useful, but something stands over it anyway. I don't think it's silence exactly. And not deception in any simple sense. It's more like an "organized absence" that can still produce the texture or counterfeit of presence. That bothers me. The language works on us using cues we learned from human life, while the source remains outside the conditions that made those cues meaningful.

The Sentence After the Sentence

Maybe that's why AI feels strange in a way that benchmark scores never quite capture. The shock isn't only that a machine can write, I believe we're far past that. It's that writing no longer guarantees what it once seemed to guarantee. A sentence used to imply, if not truth, then at least a speaker whose words were tethered to a life.

I might be too guarded or old-school, but that tether now feels looser. And once you notice it, reading itself changes. The text still speaks, but another message begins to gather behind it. Not in the literal content, but in the form of being that produced it. Emerson was writing about human character, of course. But the line now opens onto something else entirely. We are encountering language that sounds human while the thing behind it is not, and that difference may be ringing louder than the words themselves.

There was a problem adding your email address. Please try again.

By submitting your information you agree to the Psychology Today Terms & Conditions and Privacy Policy

© Psychology Today