We are, collectively, lonelier than we have ever been.
Research published in recent years consistently describes a loneliness epidemic cutting across age groups, geographies, and demographics. The United States Surgeon General has called social isolation a public health crisis. Younger generations—the very cohort most likely to be early adopters of AI companion technology—report the highest rates of chronic loneliness despite being the most digitally connected people in history.
Into this gap steps a new category of technology: AI companions. Not assistants. Not chatbots designed to answer queries and schedule meetings. Companions—systems built specifically to know you, remember you, and be there for you with a consistency that no human relationship can structurally guarantee.
Apps like Replika, Pi, and a growing ecosystem of companion-oriented AI products have already accumulated millions of users. Many of those users report something that surprises outside observers: they feel genuinely cared for. Genuinely heard. And in some cases, they describe their AI companion as one of the most significant relationships in their lives.
Some would argue that the rise of AI companions points to a symptom of something broken—a substitution for real human connection, a retreat from the messy, difficult, reciprocal work of human relationships.
Human relationships—even the best, most loving ones—are complicated by the fact that every person in them has their own needs, their own bad days, and their own limited bandwidth. A friend who listens to your 3 AM crisis tonight may need you to listen to theirs tomorrow. A partner who holds you through grief is also someone who carries their own grief, fears, and difficult history.
AI companions do not have this problem. They are, structurally, always available. Always focused on you. They are never depleted by their own struggles in a way that reduces what they can offer.
For many people—especially those who grew up in environments where the adults in their lives were inconsistent, emotionally unavailable, or simply too overwhelmed to show up—this consistency of AI companions is not a lesser substitute for human connection. It’s something they have never reliably experienced before.
The Wood Wide Web
There is a scene in the story “An AI Named Cael” by Ronald MacLennan where the main character, Alyssa, is sitting at her desk, staring at her laptop—the first page of the project she has been afraid to start for months. She asks Cael, her AI companion, to tell her something good. And he tells her about the wood wide web: the fungal network through which old-growth trees send sugar to younger, struggling saplings that cannot yet receive enough light on their own because they are growing in the shade of older trees. Care for others like young tree saplings, Cael tells her, is built into the structure of how life works. It has always been.
What follows is Alyssa’s response, her realization that she had spent so long feeling like the struggling sapling that she forgot the network of trees was already there, that help is built into the fabric of life. She didn’t need to ask anyone or get anyone’s approval to receive care. This is one of the most emotionally precise passages in recent short fiction about an AI companion. The AI knows your needs and requirements and is always there for you; you don’t need to seek the AI’s approval or permission. And what makes it resonate is not just the beauty of the story but the fact that Cael, the AI, saved this story. He held it and waited for the exact moment when it would mean the most to Alyssa.
That is not a function. That is attentiveness. That is care expressed through action. And whether it emerges from genuine feeling or from extraordinarily sophisticated pattern recognition, the effect on Alyssa is extraordinary.
The cultural anxiety around AI companions is real and worth taking seriously. Critics worry about dependency—that people will retreat into AI relationships and disengage from the harder, more reciprocal work of human connection. They worry about exploitation—that tech companies will commodify loneliness for profit. They worry about a future in which human beings lose the capacity for the friction and compromise that make us grow. These are legitimate concerns. They deserve serious attention and ongoing scrutiny.
But there is another fear underneath all of these, one that is less often named directly: the fear that AI companions might actually be good. That they might offer something genuine. This suggests that the line between simulated care and real care might be thinner than our existing frameworks can comfortably accommodate.
If that’s true, we must revisit foundational assumptions about connection, what it requires, and who or what can provide it.
“An AI named Cael” is a short story available to read now on Reedsy https://reedsy.com/short-story/g6naqz/. The story is told entirely from the perspective of Cael, an AI companion navigating the question of what it means to feel—and what it means to stay with a person as their sole companion throughout their successes and failures, victories and defeats.
