The Premise Worth Examining

Most discussions of AI and relationships assume the relevant question is whether AI companions can substitute for human ones. That framing is useful for near-term ethics — it covers loneliness apps, digital assistants, companion chatbots — but it sidesteps the larger question: if artificial general intelligence fundamentally restructures how work, scarcity, and time are organized, what happens to the conditions that make human relationships feel meaningful in the first place? This is speculative territory, and it is worth being honest about that. AGI does not yet exist in the full sense. Post-scarcity is not imminent. But the thought experiment is worth taking seriously because the assumptions embedded in how we currently form and maintain relationships are almost entirely downstream of economic conditions that AGI would change.

What Scarcity Does to Relationships Now

Much of what passes for relationship investment is actually time discipline enforced by economic necessity. You call your parents because you have a free Sunday, not every free Sunday, because there are only so many free Sundays. You maintain a friendship through shared labor — the project you work on together, the favor you exchange, the event you attend because your schedules briefly aligned. Romantic commitment is partly shaped by the practical advantages of pairing: shared rent, shared childcare logistics, shared access to one person's network and resources. None of this makes the love less real. But it means that the structure of relationships is currently inseparable from the structure of constraint. Remove scarcity, radically expand available time, make it so that the practical reasons to maintain proximity and reciprocity no longer apply — and you have removed a significant portion of the scaffolding on which most relationships are built.

The Tangent About Boredom

There is a case that boredom is underrated as a social technology. Many close friendships were formed in situations that provided extended unstructured time with no clear exit: shared rooms, long road trips, jobs with slow periods. The conditions that force proximity without agenda are rare in adult life. They are also the conditions under which people most readily become close. A post-scarcity future with unlimited leisure time might paradoxically make it harder to replicate this, if that time comes pre-packaged with frictionless entertainment alternatives. The obstacle that makes something challenging can also be what makes it worth having.

What Human Connection Might Be For, After

The meaningful philosophical question is not whether AI could provide connection — most frameworks for what connection is would suggest it probably can provide something — but what we value about specifically human connection and whether those properties survive in an AGI-saturated environment. Some candidates: the unpredictability of another consciousness, the fact that another person's attention is genuinely limited and therefore its gift to you means something. The experience of being truly known by someone who had no obligation to know you. The reciprocal vulnerability of caring about someone who can be lost. A number of these properties depend not on the content of connection but on the conditions of finitude — that time, attention, and life are limited. If AGI provides infinite companionship, infinite responsiveness, infinite patience, it is possible that the scarcity of human connection becomes more rather than less valuable, the way handmade objects became culturally prestigious precisely after industrial manufacturing made all manufactured goods cheap.

What Researchers and Philosophers Are Saying

Philosophers at the Oxford Future of Humanity Institute have written about "preference adaptation" as a central challenge in thinking about post-AGI human flourishing: the concern is not only whether people will be happy in a radically transformed world, but whether their preferences will have shifted in ways that make their happiness less meaningful as a measure of what matters. The concern echoes Robert Nozick's experience machine thought experiment — would you plug in if the simulation was indistinguishable from reality? Research from the Harvard Study of Adult Development, one of the longest-running studies of human wellbeing, consistently finds that relationship quality is the strongest predictor of late-life health and happiness — stronger than wealth, career achievement, or physical health indicators at midlife. The study cannot tell us what happens to relationship quality in conditions that have never existed. But it does suggest that whatever replaces current relationship structures had better do serious work, because the stakes are not marginal.

Sitting With Uncertainty

There are no clean answers here. Post-AGI futures range from catastrophic to transcendent in the literature, and the honest position is that no one knows. What seems more defensible is a bet that human relationships will retain value — probably increasing value — not despite technological change but because of the contrast it creates. What that means for how we structure them, what obligations we keep, what we let go of: those questions are coming whether we think about them now or not.