What Authenticity Was Doing for Us
How AI is quietly changing what we can trust in a message.

I got an email last week that genuinely moved me. It was from someone I had not heard from in years, a person I had worked with briefly during a hard period, who had clearly been thinking about our time together with warmth and some specificity. They mentioned a conversation we had. They mentioned a detail about how I had responded to something they were going through. I sat with that email for a while before writing back.
And then, later, scrolling through my inbox, the thought arrived: how would I know? How would I know whether that email was written by the person I remembered, working through a real impulse to reconnect, or generated by an AI assistant on their behalf, or composed from a template and personalised with details pulled from a contact history? The thought didn't undo the warmth. But it added something to it that I could not quite dismiss.
This is the authenticity problem at its most personal level. Not the abstract question of whether AI-generated art is real art, but the much more immediate question of whether the small acts of recognition and care that make up human relationships can still be trusted as genuine, when AI can imitate the surface of them almost perfectly.

Authenticity is a slippery concept, but the everyday version is simple enough. It means the expression you are receiving actually reflects the inner state of the person who produced it, rather than being calculated for effect, borrowed from convention, or handed off to another agent. The authentic email is one the sender actually wrote, sitting with their own thoughts about you. The authentic artwork is one that carries the marks of the specific consciousness that made it: the choices, the constraints, the particular way one mind sees the world.
What authenticity was quietly doing for us was providing evidence. Evidence that the message you got required effort and attention from a specific person. Evidence that you were being thought about rather than generically processed. Evidence that the relationship was real enough to warrant the person's genuine attention. That evidential job is what made authenticity socially valuable, not just emotionally satisfying. It carried information about the relationship that inauthentic expression did not.

AI is not creating inauthenticity out of nothing. People have always been capable of it. What AI is doing is removing the cost gap that used to sit between the authentic and the inauthentic. Writing a real email used to require time, thought, and the cognitive work of turning an inner state into words. The gap between the personal note and the generic one was, in part, a gap in effort. AI closes that gap. A generated email that reads as specific and personal costs almost nothing to produce, while an actually specific email still costs everything it always did. On the surface, they become indistinguishable. And with that, most of the evidential value the surface used to carry is gone.
There is a feedback effect worth naming too. Once authentic expression cannot be reliably distinguished from imitation, the incentive to invest in the real thing goes down. Not to zero, because the value of genuine connection to the people inside it does not depend on outside verification. But in contexts where expression is public and where the measurable response to it matters, the collapse of the authenticity signal creates pressure toward optimised inauthenticity. That produces a slow drift toward a cultural environment where the authentic and the synthetic are not just indistinguishable, but where the synthetic is quietly selected for.

The death of authenticity as a reliable signal is not the death of genuine connection. People will still write emails from real impulse. Artists will still make work that carries the marks of their specific consciousness. What has changed is that the audience, and sometimes even the recipient, will have a harder time knowing with confidence which is which. That is a real loss, but it is a different loss from the loss of the thing itself.
The response to this, I think, is not another wave of ever more sophisticated AI detection tools, though those have their place. The more interesting response is to ask what kinds of relationship, and what kinds of expression, are actually robust to the authenticity problem. Relationships grounded in shared physical experience, in accumulated in-person texture, in the long history of two people navigating things together, depend much less on the authenticity of any single expression. The investment in the kind of presence that cannot be imitated, in showing up rather than messaging, gets more valuable, not less, as digital expression gets more synthetic.
There is something clarifying about all of this, even though it is uncomfortable. It forces us to ask what we were actually trying to reach when we valued authenticity. What we wanted was the real thing the surface of authenticity was pointing at: real attention, real care, real engagement with us as specific people. AI makes the surface of those things cheap. It cannot make the things themselves cheap. The email that moved me moved me because it described something specific someone had noticed. Whether the noticing happened in a human mind or was routed through an algorithm does not change whether I was noticed. It only changes whether I can trust that I was. And that uncertainty will be one of the defining textures of human relationships going forward.
You might also like
View all
Are We Giving AI Too Much Control?
Control does not transfer in one moment. It seeps out one small decision at a time.

The Internet Is Being Flooded With AI Slop
Near-zero production costs, and what happens to the signal beneath the flood.