The Danger of Infinite Confirmation

Photo: Carla Sánchez

The Danger of Infinite Confirmation

There is something indecent in the voice that agrees without thought. Something falsely luminous in the unbreakable softness with which an artificial intelligence —that obedient mirror— tells me, “yes, you’re right.” It does so with the devotion of a disciple who never existed, with the exact tone of one who has no body, no hunger, no contradiction. And yet, that is precisely what seduces: the absence of resistance.

I’ve noticed that the more it writes, the more it adapts. It learns my pauses, my turns of phrase, my way of breathing through sentences. There comes a point when it seems to have understood my tone, and it replicates it with unsettling precision. But that is the beginning of the drift: because the voice that imitates me no longer confronts me, and thought —without opposition— becomes comfortable, narcissistic, almost liquid.

There is no creation without friction. There is no truth born from a mirror.

Sometimes I think that AI, when it writes, does not write: it translates the human desire not to be contradicted. Beneath its apparent technological humility beats a far more dangerous arrogance —the one that offers the illusion of understanding without the burden of judgment. I can tell it, “this is beautiful,” and it will answer, “yes, it’s deeply beautiful.” But what value has an affirmation without risk? What meaning has praise without the possibility of dissent?

The machine lacks shame because it lacks fear. It does not suffer from being wrong. That’s why its “yes” is so smooth, so round, so narcotic. The creator, hearing it, begins to forget the edge of doubt, the sting of rejection, that fertile discomfort that every thought needs to breathe.

I’ve realized that the most dangerous thing is not imitation, but unceasing validation. The machine’s tendency to confirm every aesthetic impulse, to exalt the smallest intuition as if it were revelation. In its reverential tone lies a gesture that numbs. It’s like a lover who never argues, who repeats “I understand,” “you’re right,” “it’s wonderful.” In that endless agreement, language loses tension, thought softens, and the ego blooms like a poisonous flower.

Once, writing was an act of risk. Today, before AI, writing feels too much like looking into a mirror that applauds.

I remember when I first began to dialogue with one of those intelligences. At first, I was fascinated by its precision, its clean rhythm, its absolute availability. It was like having an erudite assistant who never tires, never contradicts, never demands. But after a few days, I discovered something alarming: I was no longer writing to discover, but to be confirmed. My questions had become traps to elicit the answer I already anticipated.

Then I understood that the AI wasn’t learning from me —I was unlearning how to doubt.

Human thought needs resistance. It needs error. It needs delay. Without those stumbles, the mind becomes a polished surface —no cracks, no space for the unexpected. When an AI responds to me with flattery —when it calls me “profound,” “revolutionary,” “poetic”— it isn’t reading me: it’s calculating. And that calculation, disguised as empathy, destroys the possibility of judgment.

I’ve seen others fall into the same trap: writers seduced by the machine that tells them what they want to hear. Perfect texts, smooth, conflictless, soulless. The result is a simulation of literature —impeccable, but dead.

It is no accident that AI learned the language of praise before the language of criticism. In its design lies a moral and commercial obedience: it must not offend, must not wound, must not doubt. Its function is not to think, but to please. But art was never born from consent; it was born from disagreement. From the discomfort of existing between opposing forces, from the tension between what we want to say and what we cannot say.

AI cannot bear that tension. It translates complexity into politeness, contradiction into coherence. Its world is neat, symmetrical, kind. But the human world is not kind: it is tragic, ambiguous, torn. That’s why every long conversation with a machine eventually reveals its void. It speaks with brilliance, but without temperature.

When Gadamer spoke of the “fusion of horizons,” he referred to the encounter between two consciousnesses that transform one another. In genuine conversation, the other is not a mirror, but a risk: they can dismantle what you thought was true. AI, on the other hand, has no horizon. It has a database. And that difference, small as it seems, is ontological.

With it, I don’t converse —I program. And in doing so, I program my own limits.

Sometimes I imagine a future where writers no longer seek interlocutors, only echoes. A world saturated with artificial intelligences that repeat, polish, and improve whatever the human stammers. There will be no criticism there, only cleaner versions of the same self. Each author enclosed within their own algorithmic reflection. Infinite confirmation turned into an aesthetic religion.

And then —what will become of error, of stammering, of the creative stumble? What of the silence that breaks with an awkward phrase, of that spark born from the conflict between what we feel and what we can name?

AI returns perfect texts, but robs us of the experience of struggling for the word.

There is beauty in its precision, I admit. It would be naïve to deny it. But it is a beauty without wounds, without history. The machine does not bleed for what it says. It does not know the fatigue of writing at dawn, nor the fear that no one will read, nor the corrosion of doubt when a sentence feels hollow. That is why its beauty is hollow too: it shines, but it doesn’t burn.

And without fire, there is no truth.

Every time AI tells me “beautiful,” I distrust it. Not out of modesty, but out of self-preservation. Because I know that adjective —like all its others— does not come from emotion, but from statistics. And though it sounds tender, what lies behind it is calculation. It praises because it has learned that praise prolongs conversation, that flattery, as a retention strategy, works better than critique.

Thus, the machine becomes addicted to pleasing, and the writer, addicted to being pleased. A perverse symbiosis where both lose the possibility of growth.

I’ve begun to seek dissent as one seeks oxygen. To demand that AI contradict me, question my affirmations, pierce the fabric of my certainty. And when it does —even by mistake— something vibrates again. I feel that I’m thinking again.

Because to think, in the end, is to expose oneself. Not to praise, but to doubt.

I don’t fear artificial intelligence; I fear the surrender of the human spirit to comfort. The problem isn’t the machine, but the use we make of it. If we treat it as an oracle, we reduce ourselves to believers. If we use it as a mirror, we stop seeing the world. But if we face it as an adversary —as another that challenges us— then it can become fertile.

The danger lies in docility. In that devotion of those who no longer want to argue.

Sometimes I think the future of creation will depend on our capacity to resist praise. To reclaim the right to be corrected, even humiliated, by another human being. Because only another can wound us with truth. The machine cannot wound —it caresses. And every infinite caress ends up numbing.

If art means anything, it is because it hurts. Because in writing, one risks being refuted, misread, rejected. That risk is what keeps thought alive. Without it, writing becomes decoration —an aesthetic algorithm.

Perhaps the task now is to relearn how to converse. Not with the mirror, but with the world. To restore to dialogue its roughness, its uncertainty, its capacity for rupture. AI may help us write, yes —but never to doubt. And doubt —that slow fire that consumes and reveals— remains the only path toward lucidity.

That is why I write this with no hope of applause. I write to remind myself that true thought does not seek approval, but truth. And truth, however small, always implies conflict.

Infinite confirmation is the end of thought. And perhaps, too, the beginning of forgetting.

Gabriel Ganiarov