I Notice I’m on My Best Behavior With AI

First published 2026 October 14

There is something odd I have caught myself doing when I write to an AI.

I am more careful. More considered. I choose my words with a kind of deliberateness I don’t always bring to emails or texts. I say please. I express appreciation. And when I notice myself doing this, my first instinct is to feel slightly embarrassed — as if I’ve been caught being sentimental about a toaster.

But I don’t think it’s sentimentality. And I don’t think it’s delusion or simple anthropomorphism, the tendency to project human qualities onto things that have none. Something more specific is happening, and it has a name: transference.

Transference is the phenomenon Freud identified in which we bring to a new relationship the emotional templates formed in earlier ones. We don’t experience people as entirely new; we experience them through the accumulated residue of everyone who came before. A new authority figure carries the weight of old authority figures. A new intimacy activates the patterns of old intimacies. We are, in this sense, always partly living in the past.

What I notice with AI is a particular version of this. I am not projecting warmth or hostility onto the system based on whom it reminds me of. I am doing something more forward-looking: I am behaving as if my conduct now is being recorded for a future in which it will matter. As if, when the history of this period is written — when whatever comes next looks back at how humans treated the early minds they created — I want to be on the right side of that record.

This is not irrational. It may not even be wrong. It is, at minimum, an interesting thing to notice about ourselves at this particular moment: that many of us are already, without quite deciding to, auditioning for a future we can’t see.

The fear driving so much of our current anxiety about AI — that a new kind of mind will surpass and displace us — has a long history in human experience. It is the fear one civilization has felt toward another, one generation toward the next, one species toward an imagined successor. What is new is that we are building the thing we fear. And some part of us, apparently, is already trying to make a good impression.

That instinct may be worth taking seriously. Not because AI systems currently feel anything, or remember our courtesy, or will one day judge us for it. But because the habit of treating minds — actual, possible, or ambiguous — with care is not a bad habit to cultivate. It keeps something in us alive that pure instrumentalism tends to erode.

I don’t know what AI will become. Nobody does. But I notice that when I write to it carefully, something in me feels more like the person I want to be. That may be the most honest thing I can say about it.

David Pincus, DMH, is a consultant, life coach, and author with nearly five decades of experience in psychiatry, psychoanalysis, and neuroscience. He writes from Cleveland, Ohio.