On Being Human in an AI-Saturated World

There is a cracked tile in my grandmother’s courtyard in Marrakesh, near the fountain where the water always pools.

She refused to replace it because she liked the sound the water made when it hit the crack. It was imperfect, but she liked it that way because a perfect one would have wanted nothing from the people who entered it.

People entered constantly, without calling. They came because a cousin had given birth, or because someone’s father had died, or because it was Friday and you do not eat Friday couscous alone. They sat on low cushions along the wall and drank mint tea poured from a great height so the foam would rise. Often they had nothing particular to say. Showing up was enough, and everyone understood this without saying so.

There is a word in Moroccan Arabic for what that felt like: Nashat. It translates poorly into English, something like joy of life, the belief that to be alive, in a body, among other bodies, is already reason enough to be glad. Nashat is not a production or an outcome; it is what rises when people show up.

I grew up with this word, and with the slow understanding that in my grandmother’s world, intelligence mattered less than presence. My current work at OpenAI runs on the opposite conviction, and this tension between both worlds has become impossible to avoid.

To better process the tension, I wrote about it in Noema Magazine. It was thrilling to hear this in response from a reader because I fully agree:

“You’ve captured the dilemma. We need more thoughtfulness in the design choices.”

The assumption behind it is one I recognize: that building fast and thinking carefully are incompatible — that urgency crowds out conscience. From the outside, I understand why that reads as true. From the inside, it looks different: not certainty, but the discomfort that precedes it. The pauses before confident claims. The genuine difficulty people feel when they reach the harder edges of what we are making. I am saying this because the essay I want to write requires honesty about the room I actually work in, not the one that makes a better headline.

We are in a genuine shift, not just in what technology can do, but in what it asks us to reconsider about ourselves: who we want to be for one another, and what we believe that is actually worth.

My friend Julia lost someone exactly a year ago last summer. Virginia was eight years old, the daughter of Julia’s closest friend, a woman who has been like a sister to her since college. The night before camp drop-off, Virginia was at Julia’s house, and she slipped away from the other children into the living room and sat at the piano and played a song she had written herself.

Virginia died in the flooding at Camp Mystic. Twenty-seven campers and counselors were lost.

In the weeks after, Julia cooked meals. She washed Virginia’s stuffed bunny, found in the aftermath. She sat on the couch in the room where, one week earlier, a child had been playing her own song. She stayed on the phone with her friend when there was nothing useful left to say and she stayed anyway.

She also, at 3 a.m. when she could not sleep, talked to ChatGPT. She told me it helped. She could say the unsayable without worrying about exhausting someone. It met her where she was without needing her to be okay.

In the morning, she got up and washed the bunny.

The model interpreted Julia’s grief. Julia participated in her friend’s. Those are not two versions of the same thing: one is a map, drawn with great care and accuracy, and the other is the wound itself, carried in a body, into a kitchen, on a Tuesday morning.

Aristotle wrote about three kinds of friendship. Friendships of utility, where people help each other accomplish things. Friendships of pleasure, where people bring each other delight. And friendships of virtue, where people wish the best for each other and are changed by the relationship over time. The third kind of friendship was, for Aristotle, the only complete form.

AI will become very good at the first two. Whether it can participate in the third, in the slow formation through which human beings become good together, I genuinely do not know. What I do know is that AI can interpret suffering without being implicated in the condition that makes suffering possible.

A human being who sits with you in grief arrives with their own mortality, their own losses, their own body that will also fail. That shared vulnerability is not incidental to why it matters, it is what the mattering is made of.

I am a genuine optimist about this technology, and I have the evidence to stand on. The OpenAI and MIT study tracking nearly a thousand people over four weeks found that sustained interaction with AI companions meaningfully reduced loneliness, and users themselves underestimated how much it would help. For a person who has been estranged from family, or who lives alone, or who needs to say something at 3 a.m. that no one should be asked to receive while sleeping, that is not a small thing.

And yet the same study surfaced the shadow: the heaviest users were also the loneliest, the most emotionally dependent, and the least socialized with other humans. Both findings are true simultaneously, which means the technology works and the cost is visible in the very people it most reliably helps.

The people building this well already know this. What I keep trying to articulate, to colleagues and to myself, is that interpretation and participation are different goods, and substituting one for the other at scale will change us in ways we have not yet fully reckoned with.

The more optimized the world becomes, the more valuable the unoptimized will feel.

The sites where human beings remain real to one another are not the efficient ones. They are the ones that ask something of you.

My grandmother’s courtyard is still there, and so is the cracked tile. Fewer people come on Fridays now, and the ones who do carry phones that could explain almost anything about grief, or memory, or how to make tea. What the phones cannot do is sit. They cannot stay. They cannot be the person who, years later, will be remembered for being there.

I grew up watching my grandmother pour tea from a great height, aerating the liquid, making the foam rise, doing something the kettle could not do for her. She was not being inefficient, she was making something that required her. I think about that a lot these days, and I think we are all going to have to decide, much more deliberately than we have so far, what we still want to make that way. Read the full piece on Noema.

I’m just so unbelievably excited to be alive in this era.

– Houda