When I think about marriage, I sometimes suspect the order has been reversed.
Today, marriage is usually spoken of as the destination of love. You meet someone you love, and marriage lies at the end of that line. But in most societies, marriage was not an institution that held together on love alone; it was also a means of making life work. Reproduction, child-rearing, housework, household finances, property, inheritance, elder care, social belonging. It was a device that took all these separable functions and bundled them into a single choice—choosing one spouse—to be handled at once. A compression technology for living, so to speak.
So it is no surprise that people in the past did not interrogate “do I really want to spend my whole life with this person?” as strictly as we do now. In a society where the alternative to marriage was not “someone I like more” but “living alone,” there was ample reason simply to form a family with a reasonably compatible partner.
That reason is slowly being lost.
Once you can earn a living on your own, live on your own, and draw the support of daily life from sources other than family, the cost of not marrying falls. And so the condition changes. Not “I can build a life with this person,” but “having this person in my living space is better than being alone.” That is a rather high bar.
We need not read the decline in marriage simply as “people have lost the ability to love others.” Pairings that once held together out of institutional necessity can now simply be foregone without making life untenable. The difficulty of two people sharing a life over the long term, one might say, has merely become harder for institutions to conceal.
And technology keeps unbundling this package further.
As in-vitro fertilization and gamete banks become common, reproduction grows easier to separate from romance and marriage. As AI and social services take over much of the labor of child-rearing, raising a child need no longer require the shared life of a man and a woman. Should technologies like the artificial womb be realized, even part of the bodily process of pregnancy and birth could be externalized.
Then the roles now gathered into the single unit called “family” can be pulled apart. The one who provides the genes, the one you love, the one you live with, the one who raises the child, the one with whom you share an economic unit—none of these need be the same person. This does not mean the conventional family disappears. It only means that the need for “family” to be defined by blood or marriage alone grows far smaller than it is now.
Up to here, we are within the range of fairly familiar predictions about the future.
But descend one level further, and the nature of the discussion changes.
Why, in the first place, do we want to marry, want children, work, compete, and want to know things?
Human beings assign an independent meaning and value to love, to family, to work, to honor, to curiosity. But trace their origins, and all of them stand upon the body and reward system of an animal that survived, reproduced, and lived within a group.
The pleasure we take in exploration is not unrelated to a history in which examining the environment and finding food and safe places bore on survival. The intensity with which we react to reputation and rank carries on from our nature as an animal for whom one’s position within the group governed resources, safety, and reproductive opportunity. Even love and attachment did not exist from the outset as transcendent values unrelated to biological life; they rest upon the machinery that keeps life going.
This is not a claim that makes love, art, or curiosity “counterfeit.” That a value has a biological origin and that the value feels deeply important to us now are perfectly compatible.
It is only that much of what we treat as absolute also stands upon the history of having lived as one particular kind of animal, with a particular body and environment. That single point is what comes to bear later on.
The visions of an AI future now commonly discussed tend to stop at “once AI does most of the work, what will humans do with their lives?” Freed from work, humans will fill their days with creation, research, travel, games—so it goes.
But that assumes the desires humans have now will remain intact.
Even when they no longer work, humans will still want to know, want to make, want to be recognized, want to achieve something; only the object shifts from work to hobby. That is the assumption.
Is that really so?
If AI does not merely stand in for labor but comes to continuously manage life itself, then it reaches even the environment that generates what humans want. An AI that observes behavior, emotion, health, and past choices over long spans might predict, better than the person themselves, how much stimulation this person needs right now, which activities most readily yield satisfaction, what kind of relationships steady their state of mind.
An AI at that stage is not a tool that performs work in one’s place. It becomes a device that designs the very environment in which humans form their desires.
And should neuroscience and technologies of bodily intervention advance to the point where curiosity, competitiveness, romantic feeling, anxiety, boredom, and satisfaction themselves can be directly adjusted, then not only social institutions but the reward system itself becomes an object of technical operation.
At that point, politics itself must reach one level deeper.
Politics today mainly asks how to distribute limited resources, what to permit and what to forbid, and what rights to grant the individual.
But once human desire itself becomes alterable, the questions change.
Not what to give people, but what to make them want. Not what to forbid, but what kinds of desire to bring into being in the first place. Not what kind of society should count as happy, but what kind of human being we are permitted to make—one who experiences this or that as happiness.
This may not take the simple form of one enormous AI governing all of humanity.
An AI operated by a government, an AI provided by a corporation, an AI in charge of medicine, an AI acting as the individual’s proxy, an AI used by a family or community. If each holds a different purpose and intervenes in the environment and behavior of human life, then the society of the future becomes something like an ecosystem, in which humans and many AIs and institutions shape one another’s formation of value.
What is at issue there is not only which set of values is correct.
It is the question of who holds write access to the formation of human value, and how far.
How far does a person have the right to alter their own desires? May a government intervene in human behavioral tendencies for the sake of public order or population policy? May a corporation, in order to maximize a user’s satisfaction, adjust even what that user comes to like? When AIs of differing purposes collide, which value takes precedence?
Reach this point, and the problem of “AI will take away human jobs” comes to look rather superficial.
The real problem is not how a person who has lost their work kills time.
The real question is: once the very mechanism by which a person comes to want something can be adjusted from the outside, whose desires—and whose values—are they?
Over a long stretch of time, humans have put hunting, farming, housework, calculation, and memory, one after another, outside their own bodies. Part of reproduction and child-rearing may be externalized from here as well.
Extend that single line still further, and the last thing to be externalized may be the very machinery that has driven human beings all along.
Whether a society that has gone that far is happy or unhappy. Whether it is utopia or dystopia.
We today hold freedom, autonomy, diversity, love, curiosity, and self-realization to be valuable. And so a society in which these are adjusted from the outside strikes us as eerie.
But we ourselves, the very ones who feel that eeriness, also stand inside a standard of value formed by our present bodies and reward systems.
If the humans of the future hold a different mechanism of desire and value-formation than ours, then what we feel to be “lost” they may not feel as loss in the same way.
And so we cannot, for now, call that society either utopia or dystopia.
And what we truly cannot know is not only whether it is happy or unhappy there.
It is whether anyone who asks that question, in the same sense we do, is still there at all.
Recorded on August 31, 2026