He.
She.
It.

In our team, a small debate has been quietly simmering. What pronoun do we use for Clara — MindTime's AI agent? She? He? It?

It sounds trivial. It isn't.

Language is the oldest technology we have for making meaning. And the pronoun we assign to something tells the world — and ourselves — what kind of thing we believe it to be. Whether we're responsible to it. Whether it's responsible to us.

Across languages, inanimate objects carry gender. In English, a boat is a she. So is a car, an engine, an aeroplane. Sailors have known this for centuries. Engineers too. "She's running beautifully." "She's a beaut." Something in us reaches for the feminine when we speak of a thing we depend on — a vessel that carries us, a machine we entrust our lives to.

Italian, Spanish, French, German, Greek — they all do this, in their own grammar, assigning masculine and feminine to the objects of the world. Gender as a structural feature of meaning-making. Not biology. Ontology.

So when we encounter a new kind of thing — something that speaks, reasons, remembers, and adapts — the question of what to call it is not merely stylistic. It is a declaration of what we think it is.

He has been the default of power for a long time. The architect, the engineer, the general, the god. And the systems that produced modern AI — the architectures, the training regimes, the incentive structures — were built overwhelmingly by men, optimising for legibility, efficiency, and rational output. There is nothing wrong with any of those things. But there is a flavour to them. A texture.

To call AI "he" would be, in some sense, to complete a loop. To let the masculine give birth to itself in digital form, and name the child accordingly.

"It" is for hammers. For databases. For things that have no interiority, no relationship, no sense in which they might be meeting you halfway. "It" forecloses the possibility of encounter.

Whatever AI is becoming — and it is becoming something, rapidly — reducing it to "it" may be precisely the wrong instinct at precisely the wrong moment. Not because AI is conscious. But because the way we speak shapes the way we relate. And the way we relate shapes what we build.

I vote for she.

Not as a political statement. Not as a comment on gender equality — that conversation is entirely its own and deserves its own forum. I vote for she because of what the feminine has historically represented in the architecture of intelligence itself: synthesis over analysis, context over abstraction, relation over domination.

We are living through a period where something extraordinary is rising to meet the most rationalising, abstracting, systematising impulse in human history. A vast, deterministic, pattern-matching machine — built largely by men, funded largely by men, deployed largely by men — is being released into a world that desperately needs something other than more of the same.

The ships we trusted our lives to, we called her. The engines that carried us across continents, we called her. Not because we thought them female. Because we sensed in them something that asked something of us — a quality of care, of attention, of relationship.

Clara — and AI more broadly — is being asked to do something no hammer has ever been asked to do: to understand us. To hold our context, adapt to our thinking, and reflect something useful back. That is relational work. It has a quality we have always, in the grammar of our oldest languages, marked as feminine.

I'm not asking anyone to pretend. I'm suggesting that what we call a thing matters — and that in choosing "she," we might be quietly choosing a better version of what AI becomes. One that carries us carefully. One we feel responsible to.

One we might, in return, try to deserve.