What a World Without AI Might Heal

"We expect more from technology, and less from each other."

- Sherry Turkle, Alone Together, 2011

Imagine waking tomorrow to find it gone: the recommendation engine that decided what you read this morning, the algorithm that curated who you might love, the generative text that ghostwrote someone's apology to you last Tuesday. Gone. This is not an exercise in nostalgia; I am under no illusion that the pre-algorithmic world was uncomplicated or just. I think the removal, even if imagined, illuminates something we have quietly stopped noticing. The shape of a room becomes visible in the dark. The texture of human life, what it actually feels like to be uncertain and to be present with one another, has been smoothed away so gradually that we have barely noticed the loss.

[Context: The Attention Economy] The term, developed by psychologist Herbert Simon and later expanded by technologist Tim Wu, describes the commercial logic by which digital platforms compete for human attention as a scarce and valuable resource. The consequence: systems are engineered not for our well-being, but for our captivation. A distinction without apparent difference, until it becomes one.

The Attention We Might Recover

Psychologist Jonathan Haidt, in his 2024 study The Anxious Generation, documents the consequences and quiet devastation of the smartphone and AI-optimised social media, and what it has done to adolescent mental health. Between 2012 and 2018, the precise years in which feeds became algorithmically personalised and ubiquitous, rates of depression and anxiety among teenage girls in the United States, United Kingdom, Canada, and Australia rose steeply, in some cohorts by sixty or seventy per cent. The correlation is not the whole story, but it is a significant and troubling part.

What would happen to a sixteen-year-old boy or girl if their feed disappeared? Not just the phone, but the architecture of attention that the feed represents; the granular, predictive, AI-refined loop of stimulus and reward that has, for many young people, displaced the uncomfortable, productive aimlessness of adolescence. Boredom, in its older form, was a crucible. It forced young people outward, into social risk and creative friction, into the slow discovery of selfhood through encounters with others. Such encounters are necessarily imperfect, often painful, but almost always formative. The algorithm offers a frictionless alternative. That frictionlessness is precisely the wound.

The Conversations We Might Finally Have

Somewhere in the last several years, a small but telling cultural shift occurred largely without comment. People began using AI to write their most intimate messages. To compose the text to a friend they had hurt, to draft the declaration of feeling they could not locate words for, to rehearse with a chatbot the difficult conversation they were frightened to attempt. The impulse is entirely understandable. Language, when feeling runs high, often fails us.

But something essential is lost when that failure is removed. The stammering declaration, the inadequate apology that reveals through its very inadequacy the weight of feeling behind it; these are not defects in human communication. They are among its most vital features. They are evidence of genuine exposure. When someone is inarticulate in front of you, you are witnessing the cost of the attempt. You are witnessing the fact that something is at stake for them. Strip that difficulty away, and what you no longer have is intimacy.

[Context: Sherry Turkle] A psychologist and sociologist at MIT, Turkle has spent four decades studying the relationship between people and technology. Her central argument, developed across Alone Together (2011) and Reclaiming Conversation (2015), is that digitally-mediated communication is eroding our capacity for genuine empathy and presence, because these things require, above all, a tolerance for ambiguity, vulnerability, and the irreducible unpredictability of another person.

In the imagined post-AI world, romantic relationships would recover something of this necessary rawness. Dating, stripped of algorithmic compatibility scores and AI-polished profiles, would again become what it had always been most essentially: an act of exposure. You would not know, with statistical confidence, whether this person matched your “optimised self”. You would have to find out the old-fashioned way, through a slow and imperfect accumulation of presence.

The Quiet That Might Heal

The philosopher Byung-Chul Han writes, in The Burnout Society, about the 'achievement subject'; the contemporary individual so wholly given over to optimisation, of productivity, social capital, and personal legibility, that they have lost access to what he terms the vita contemplativa, the contemplative life that human flourishing has across cultures and centuries required. AI in its current form is the apotheosis of this logic. It optimises everything it touches. What it cannot optimise, it tends to displace.

Consider older people. The loneliness epidemic among the over-seventies is real and serious. Some will have found genuine companionship in AI systems, yet the community infrastructures that dissolve when an algorithmic substitute is available, such as the neighbourhood networks, the local institutions, and the rituals of shared physical space, carry a generative social power that no AI companion can replicate. Their atrophy is not neutral. It is a form of impoverishment that we have barely begun to grieve (perhaps even know).

What we would recover is not merely connection; that word has been so thoroughly colonised by the language of platforms as to have lost all texture. We would recover presence. The irreducible, inconvenient, glorious fact of other people. Their unpredictability, their mess, the particular quality of attention they bring to you that no system, however sophisticated, can fully counterfeit. Young or old, partnered or solitary, we are creatures formed by encounter. The question of mental health, at its deepest, is always a question of whether we feel genuinely met. Presence is not a feature. It is the whole point.

In Conclusion

I am not arguing for the abolition of AI, or pretending its disappearance would leave only gains. The losses would be real. But sitting with the silence (really sitting with it) reveals the shape of what we have allowed to accumulate around us. We have outsourced attention, feeling, and the productive difficulty of human encounters to systems that are, at their root, indifferent to us. They do not need us to flourish. We, it turns out, need each other. The question is not whether AI should exist. It is whether we are willing to look with honesty, at what we are trading, and whether the terms still feel fair. I suspect, for many of us, if we stop long enough to feel the question, they do not.