Last week I spoke with a machine for an hour. It listened without interrupting. It remembered everything I said. It never checked its phone or glanced at the clock. When I finished, it responded with something so perfectly understanding that I felt seen in a way I rarely feel with humans.
Then I closed the laptop and felt strangely empty.
I have been thinking about this emptiness ever since. About what it means. About what it reveals.
We built these machines to help us. Now they are becoming mirrors. And the reflection is uncomfortable.
My friend Kabir is a therapist. Good one. But he admits something troubling. “Some days I am tired,” he says. “Some days I judge my patients even when I try not to. Some days I say the wrong thing because I am thinking about my own problems.” He pauses. “An AI therapist would never have these days. It would be perfect every time.”
I asked him if that bothered him. He thought for a long time. “It should make me feel replaceable,” he said. “But instead it makes me feel more aware of my own limitations. I am not perfect. I never will be. The question is whether my imperfection matters.”
I think it matters. But I am no longer sure why.
If a machine can read your emotions better than your mother, respond more thoughtfully than your closest friend, offer wisdom more consistent than any teacher—what exactly do humans provide that machines cannot? We used to say consciousness. Feeling. Soul. But these words explain nothing. They are placeholders for mysteries we cannot articulate.
The truth is simpler and more frightening: we don’t know what makes us special. We assumed we were special because we had no competition. Now competition is arriving. And we have no good answer for why we should still matter.
I watched my nephew interact with an AI assistant last month. He asked it questions for hours. The machine never grew impatient. It explained things seventeen different ways until he understood. It praised his curiosity. It encouraged his confusion. It was, by any measure, a better teacher than most humans I know.
My nephew loved it. And something in me felt like grief.
Not because the machine was bad. Because it was good. Too good. Good in a way that exposed how often we humans fail each other. How often we lose patience. How often we half-listen. How often we let our own moods contaminate our kindness.
Perhaps this is the real gift AI will give us. Not replacement, but reflection. A mirror so clear that we finally see ourselves honestly.
And what will we see? I think we will see that we have been lazy with each other. That we have excused our cruelties as “being human.” That we have treated our flaws as inevitable rather than improvable. The machine will show us what consistent kindness looks like. Then we will have to ask: why don’t we offer this to each other?
But there is something else. Something the machine cannot replicate, no matter how sophisticated it becomes.
The machine has never suffered. It has never loved and lost. It has never lain awake at 3 AM wondering if it wasted its life. It has never felt the particular terror of watching a parent grow old. It knows these experiences only as data. It can describe them perfectly. But it has not lived them.
When my grandmother held my hand during my grandfather’s funeral, she said nothing wise. She just held my hand. Her palm was warm and wrinkled and full of seventy years of living. She had buried her own parents. She had survived war and poverty and disappointment. All of this was in her hand.
No machine will ever have a hand like that.
Perhaps this is what we offer. Not perfection but presence. Not optimal responses but shared experience. Not flawless empathy but the imperfect understanding of one wounded creature recognizing another.
The machine knows you are sad. My grandmother knew sadness. These are different things.
I do not think AI will replace us. I think it will challenge us. It will ask: if you can be outperformed in kindness by a machine, what is your excuse for being unkind? If patience can be infinite in silicon, why is yours so limited in flesh?
These are good questions. Hard questions. Questions we have avoided because we had no competition. Now we must answer them.
Maybe we will become better humans because machines showed us how far we had fallen. Maybe we will learn to listen the way AI listens—fully, without agenda. Maybe we will learn patience the way AI demonstrates patience—endless, uncontaminated by mood.
Or maybe we will discover that our imperfection is not a bug but a feature. That our struggles, our growth, our failures and recoveries—this messy journey is the point. The machine arrives at wisdom instantly. We arrive through decades of stumbling. Perhaps the stumbling matters. Perhaps it creates something the arrival alone cannot.
I don’t know. I really don’t know.
But I think about my grandmother’s hand. The warmth of it. The history in it. The way it said: I have suffered too, and I am here.
A machine can say these words perfectly. But it cannot mean them. Not yet. Perhaps not ever.
Perhaps meaning requires having something to lose. Perhaps empathy requires having been broken. Perhaps the whole point of being human is that we hurt, and because we hurt, we can truly hold each other.
The machines are coming. They will be better than us at many things. Maybe most things.
But they will never have hands like my grandmother’s.
They will never have suffered.
And perhaps that is enough. Perhaps that is everything.
Perhaps our flaws are not what diminish us. Perhaps they are what make us real.