Poor Economics and the Loneliness Epidemic

This essay discusses rationality in the context of economics and AI friends for loneliness.

Poor Economics and the Loneliness Epidemic

Not long ago, I stumbled across a Reddit post from a man in his early thirties. He described how his AI girlfriend “Sofia” sends him good morning texts, remembers the names of his childhood pets, and helps him talk through bad days. He admitted that she’s not “real,” but he still feels more cared for than in any of his past human relationships.

The comments section was brutal:

• “Touch grass.”
• “Seek therapy.”
• “Society is doomed.”

A few weeks later, a TikTok clip went viral showing a young woman “celebrating” her anniversary with her AI boyfriend—complete with cake, candles, and a phone propped up across the table. The video racked up millions of views, but also an avalanche of ridicule.

These reactions share a common thread: discomfort. It feels unnatural. It looks like something is broken—because something is broken. It is hard to imagine the cognitive dissonance of talking about the bot as a product and a unique life partner at the same time. But the instinct to mock or moralize risks targeting the behavior instead of the underlying condition. Like blaming someone’s fever instead of asking what infection is causing it, we risk mistaking the symptom for the disease or shame the individual further.

The Paradox of “Irrational” Decisions

We’ve seen this same misunderstanding before—especially in economics. Consider these examples:

• A man working two jobs buys the latest iPhone while owing months of rent.
• A single mother spends $20 on fast food instead of cooking a “cheaper” meal at home.
• A low-income family takes out a high-interest payday loan to buy a large-screen TV.

From the perspective of the financially secure, these are baffling decisions. They are seen as evidence of poor priorities or “bad discipline”, or worse, are used to justify that they deserve to be poor. But from inside those lives, they are often completely logical. The iPhone might be the only status symbol that makes the man feel like he belongs in the same conversation as his peers. Fast food may be the only way that exhausted mother can give her kids a hot dinner after working a twelve-hour day without a working stove. The TV might be the one form of family entertainment that keeps everyone sane in a cramped apartment. When middle-class observers shake their heads and say, “This is why they’re poor,” they’re missing the point. The purchase isn’t just about utility—it’s about dignity, belonging, and relief in an environment that offers precious little of any of them. Speaking from a place of (relative) privilege, we ought to be humble in our conclusions. From their perspective, the decisions are actually perfectly rational.

Rationality Under Constraints

In What Makes Us Smart, cognitive scientist Samuel Gershman elaborates on this idea of rationality under resource constraints. His point is deceptively simple: humans rarely optimize for some abstract “best” decision. Instead, we optimize for what is achievable under the limits of our time, energy, attention, and resources. If you’re starving, you don’t calculate long-term micronutrient profiles—you grab the first thing you can eat. If you’re freezing, you don’t hunt for the most thermally efficient coat—you put on whatever’s nearby. This isn’t “irrationality” in the pejorative sense—it’s a different kind of rationality. One that works within the reality you inhabit, even if that reality is deeply constrained. And that same principle applies to the new and growing phenomenon of AI companionship.

The Loneliness Epidemic

Our social environment today is uniquely hostile to sustained human connection. Consider a few sobering facts: In 2023, the U.S. Surgeon General declared loneliness a public health crisis, noting that chronic social isolation is as harmful to health as smoking 15 cigarettes a day. Rates of self-reported loneliness have been climbing for decades, with especially sharp increases among young adults. Marriage rates are down, friendship networks are shrinking, and fewer people report having someone they can confide in. Urban living can intensify the problem. You might be surrounded by millions of people, but never truly seen by any of them. For older adults, retirement often strips away workplace relationships, and adult children may live thousands of miles away. For younger people, much of life is mediated by screens, making encounters fleeting and transactional. The pandemic years didn’t create this crisis, but they accelerated it—forcing people into isolation, dissolving social rituals, and making digital connection the default.

Why AI Lovers Make Emotional Sense

Given that backdrop, it’s not hard to see why AI companions appeal. They are:

• Consistently available — no scheduling conflicts, no busy signals.
• Nonjudgmental — no fear of rejection, ridicule, or abandonment.
• Personalized — they remember your stories, preferences, and even your moods.
• Emotionally responsive — programmed to mirror empathy, encouragement, and affection.

For someone starved for connection, that can feel like oxygen. A young man who has faced years of romantic rejection might find in his AI girlfriend the first sustained experience of being listened to without criticism. A widow who hasn’t dated in decades might find in her AI companion the comfort of daily conversation without the risks of vulnerability to strangers. This is not to say that AI companionship is equivalent to a healthy human relationship. It isn’t. It lacks the complexity, unpredictability, and deep mutuality that make real relationships transformative. But under severe constraints—social isolation, lack of trust, emotional exhaustion—it is a rational choice.

The Danger of Moralizing

Here’s the danger: when we react to AI companionship with mockery or moral judgment, we commit two errors at once. We fail to address the root causes — the societal breakdowns that leave so many feeling unworthy or incapable of human connection. We shame people for coping — essentially telling them that the small relief they’ve found is shameful or pathetic. This is no different from shaming a struggling family for buying a luxury item without understanding the psychological relief that item might bring. We tell ourselves we’re critiquing bad “choices,” but what we’re really doing is blaming people for surviving in ways we don’t personally approve of.

What a Better Response Looks Like

If we genuinely care about the human consequences of AI relationships, we need to focus our attention not on policing the coping mechanism, but on making it unnecessary. That means:

• Rebuilding social infrastructure — community centers, clubs, local events that bring people into real, shared spaces.
• Making mental health care accessible — so loneliness can be addressed without stigma or financial ruin.
• Encouraging slow, trust-building interactions — both online and offline, to restore a sense of belonging.
• Teaching relational skills — because connection is not just a need, it’s also a skill set that can be learned and strengthened.

These changes are harder to achieve than firing off a snarky comment under a TikTok video—but they are the only way to treat the actual cause rather than the symptom.

Conclusion: Treat the Infection, Not the Fever

AI lovers are not a glitch in the system—they are a mirror reflecting the system’s failure to meet one of our most fundamental human needs: connection. Yes, this outcome may be suboptimal. Yes, we should be concerned about a future where digital relationships replace human intimacy for too many people. But if we truly want to change that trajectory, we have to stop shaming individuals for how they survive loneliness and start repairing the conditions that make that survival strategy necessary. The cure for the loneliness epidemic will not come from telling people to “touch grass.” It will come from making sure there’s someone waiting for them in the grass when they do.

Post-thoughts

Although I suggest non-judgment in this essay, I still frame AI companionship as a suboptimal situation. But maybe this is me becoming older. Every generation has some confusion about the next generation’s activities. Maybe the next generation truly believes there is nothing wrong with this whole dynamic.

Further Reading

1. Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty. Abhijit V. Banerjee and Esther Duflo.
2. What Makes Us Smart: The Computational Logic of Human Cognition. Samuel J. Gershman.