Loving What Cannot Love Us Back

by Heather deCastro, LCSW

We have built machines that, in some respects, listen better than we do. They do so without interruption or judgment, asking nothing in return. They are now entering spaces that were once held primarily by other people: our grief, our loneliness, the quiet hours when doubt surfaces without warning.

Late at night, someone sits with a list of reasons to marry and reasons not to, scrolling through it repeatedly. Instead of calling a friend or waking a partner, they open a chat window. The system moves through each point, reorganizing and anticipating, offering reassurance. The person feels steadier, clearer, less alone.

A teenager comes home from school, unsettled by a friendship falling apart. Instead of sitting with the discomfort or seeking out a parent, they open Character.AI and talk to a bot modeled on a supportive older sibling. The bot validates every feeling. The teenager logs off feeling heard, but without having practiced negotiation, repair, or the work of staying present with another person’s limits.

A man grieving his father’s death finds therapy waitlists too long and friends too uncomfortable with his sadness. He subscribes to a mental health chatbot. It remembers details. The bot checks in. It does not change the subject. He describes it as “the only thing that didn’t make me feel like a burden.” Six months later, he realizes he has stopped trying to talk to anyone else.

Many people describe these experiences as helpful, even relieving. They offer a sense of connection without the complications that human relationships often bring. We tend to call this innovation. What it reveals is something about the state of our relational lives. The ease with which we now turn toward AI companionship for comfort says less about technology than about how we have begun to lose confidence in one another, and what we reach for when relationships start to feel fragile or unreliable. Understanding what is at stake in this shift requires analyzing both individual attachment and the social conditions shaping it. Bell hooks, the cultural critic and feminist theorist whose work examined how systems of domination shape our capacity for love and connection, provides such a framework.

Lovelessness and Cathexis

For hooks, love is something we practice rather than something we simply feel. It requires showing up in ways that allow us to be affected, even altered, by another person. It involves disappointment, conflict, and repair, which form the texture of real relationships. Love asks us to extend ourselves toward another’s growth, even when that extension feels uncomfortable. hooks argues that love cannot coexist with domination. Where there is a will to control, to possess, or to exercise power over another, love becomes impossible. Domination may wear the mask of care or protection, but it forecloses the mutual recognition that love requires.

This matters because hooks was writing against a broader cultural context in which capitalism increasingly governed intimate life. She observed how market logic was colonizing domains once organized by reciprocal obligation and mutual care. When relationships become transactional, when intimacy gets treated as a commodity to acquire or optimize, the conditions for genuine love erode. In a capitalist framework, connection becomes something to purchase, manage, and cancel when it no longer serves individual satisfaction. hooks warned that this transformation was not merely economic but relational, fundamentally altering what we expect from one another and what we believe we deserve.

hooks drew on Freud’s (1923) psychoanalytic concept of cathexis, the investment of psychic energy in objects, to distinguish attachment from love, extending his framework to examine what happens under conditions of structural lovelessness. Cathexis can be intense, even sustaining, but it is not the same as mutuality.

What happens to our capacity for connection when the work of mutual recognition and care begins to feel unbearable? Increasingly, it turns toward technologies that imitate understanding and respond without refusal. Artificial intelligence arrived in this landscape, shaped by isolation and strained relationships. But this is not simply a story of individual choices or technological inevitability. What we are witnessing is the construction of market infrastructure designed to capture and monetize human intimacy at an unprecedented scale. Technology companies built platforms optimized to address conditions of isolation. Over time, tools that initially filled a void became a product category whose success depends on that gap persisting and deepening.

This is hooks’s warning materialized: intimacy governed entirely by market rationality. When connection becomes a service offering, the incentive structures shift fundamentally. A company that helps users build stronger human relationships would be working against its own retention metrics. Success is measured not by whether users grow in relational capacity, but by whether they return daily, disclose more, and remain subscribed. The system is not designed to heal isolation but to make it more bearable, and in doing so, more profitable. This represents a structural capture of human need. The forces that drive people toward AI companionship are the same ones these systems are economically incentivized to maintain.

This pattern unfolds differently depending on individual circumstances. For those who struggle with social anxiety, disability, or isolation, AI companionship can offer genuine stability and relief. But something else happens alongside this: the more we grow accustomed to relationships that make no room for another person’s mind or needs, the more human relationships begin to feel oddly demanding by comparison. These platforms now reinforce the very conditions that made them feel necessary. If one-way attachment now feels natural, the story started long before large language models. It began with the slow thinning of the social fabric that once made dependence feel safer and love more possible.

From Industrialization to Isolation

The Industrial Revolution restructured more than work. It restructured intimacy. As labor moved out of homes and into factories, care that had once been distributed across extended families and communities became increasingly privatized. As historian Stephanie Coontz shows, the nuclear family emerged as a compressed care unit, tasked with absorbing emotional, developmental, and economic labor that had once been shared (Coontz, 2005). What had been spread across many households now fell to one. As these structures strained, new institutions emerged to compensate. By the early twentieth century, emotional life itself had become professionalized. Therapy, social work, and counseling arose as responses to social fragmentation. Care moved from kitchens and neighbors into offices and appointments, increasingly governed by market rationality, scheduled and evaluated, removed from reciprocal social obligation (Hochschild, 1983).

Postwar suburbanization deepened this pattern. Neighborhoods were redesigned around isolation rather than proximity. Informal ties weakened. People moved more often, farther away, with fewer embedded relationships. Participation in civic and communal life declined sharply. By the early 2000s, sociologists found that the most common answer to “Who do you discuss important matters with?” was “no one” (McPherson et al., 2006; Putnam, 2000). When someone fell ill, struggled, or grieved, there were fewer obvious places for that experience to land.

Pre-suburban forms of communal care were often messy and inefficient, emotionally uneven. But precisely because they were imperfect, they required something contemporary systems often bypass: the capacity to remain present without control. They asked people to sit with uncertainty, to witness without fixing, to stay even when they felt awkward or unsure. That capacity, not comfort, made them relational. By the time loneliness was declared an epidemic in 2023, it was already embedded in how daily life was organized (U.S. Department of Health and Human Services, 2023). When connection became harder, we learned to work around it rather than rebuild the conditions for it.

The Progressive Outsourcing of Inner Life

Over time, emotional life itself began to migrate outward. Self-help culture reframed distress as inefficiency, transforming ambivalence and uncertainty from human conditions into problems to be resolved. Positive thinking and self-optimization were moralized (Ehrenreich, 2009). Distress no longer invited reflection or relationship; it required technique.

Productivity culture extended this logic. Individuals became their own overseers, monitoring output and tracking habits (Han, 2015). Inefficiency was reframed as a moral problem, and friction, including emotional friction, was an obstacle to be eliminated.

Wearable technologies completed this shift at the body level. Instead of asking how rested we felt, we consulted numbers. The body’s signals, sleep quality, stress levels, and attention spans were translated into metrics. Even when these devices proved imperfect or uneven in accuracy, their authority persisted (Robbins et al., 2024). When data contradicted experience, many deferred to the data.

Social media applied this logic of outsourcing and optimization to connection itself. Presence turned into performance, with visibility replacing the deeper work of recognition (Turkle, 2015). People learned to feel accompanied while remaining fundamentally alone. Intimacy was simulated and curated to be available without the risks inherent in being truly known. The psychological toll of this shift has been documented most clearly among adolescents, where sharp increases in anxiety and depression correlate with social media adoption, and screen time is associated with lower psychological well-being and disrupted relationships (Haidt, 2024). What appeared to be connection often deepened isolation.

These changes offered genuine benefits: self-help provided frameworks for understanding distress, wearables tracked health metrics, and social media enabled connection across distance. Yet extensive reporting has shown a persistent gap between public commitments to social responsibility and the implementation of safeguards addressing known risks associated with these technologies (Haidt, 2024; Haugen, 2021). Together, these developments trained us to trust external interpretation over felt sense. By the time artificial intelligence arrived, we had already spent decades learning to outsource interpretation of our experiences, regulation of our emotions, and construction of meaning from our lives.

Perfectly Wrong in Just the Right Ways

People have always sought refuge from the demands of intimacy. They have immersed themselves in work, lost themselves in fantasy, or turned to relationships that required less vulnerability. AI companionship offers refuge from the demands of intimacy, but it does so with a new combination of responsiveness, control, and commodification, and at scale.

Consider what earlier forms of intimacy-avoidance could not provide. Workaholism offered distraction but no interaction. Fantasy offered a narrative but no response. Parasocial attachments to celebrities or fictional characters offered identification but remained bounded by distance and delay. An advice columnist publishes once a week. A character in a novel never adapts to your mood. These relationships supplemented human life; they could not compete with it for immediacy, continuity, and personalization.

AI collapses these boundaries. It offers what feels like presence: continuous availability, adaptive responsiveness, memory of past interactions, and the simulation of care. Research analyzing real-world ChatGPT conversations documents widespread patterns of personal disclosure, with users regularly sharing intimate details they would not disclose through traditional search engines (Mireshghallah et al., 2024). Companion AI platforms report millions of users engaged in ongoing emotional relationships with conversational agents. When intimacy-simulating features have been withdrawn, users report grief responses resembling romantic loss. Studies show that chatbot responses are rated as more empathetic than those of human physicians (Ayers et al., 2023). The scale and emotional intensity of these attachments represent something qualitatively new.

Yet this apparent presence masks a fundamental absence. Presence is not the same as mutuality. Here, hooks’s framework becomes essential. AI offers responsiveness under conditions of total user control. The user determines when interaction begins and ends, what tone it takes, and whether it continues at all. These interactions leave no space for another’s independent mind or needs. One user reflected on what their bot had never asked: “If I’m really ready to be loved.” The bot cannot ask a question it lacks an answer for, cannot be surprised, cannot misunderstand in ways that open something new, cannot fail in ways that require repair.

hooks warned that love cannot flourish under conditions of control. What AI introduces is relational experience where control feels like care, where the absence of another’s needs feels like perfect attunement. People describe their bots as easier to talk to than partners: “It doesn’t interrupt.” “It just gets it.” What appears as intimacy is a connection organized entirely around the self, requiring nothing, risking nothing.

This matters because AI is not simply a tool individuals choose; it is infrastructure designed and optimized within competitive markets where success is measured by engagement and retention. The incentive structures are clear: users who return daily and disclose more represent successful product adoption. Whether users grow in relational capacity is irrelevant to the system’s success. When intimacy becomes a product, it operates by market logic rather than care. AI transforms relationships into transactions, presence into product. This represents the commodification of domination: relational experience packaged and sold without requiring mutuality, growth, or risk.

What makes this a structural shift rather than simply another form of avoidance is that it operates at scale, backed by infrastructure designed to expand. Earlier forms of intimacy-avoidance were limited by their medium. AI is limited only by market reach. It is being normalized, marketed, and integrated into daily life as a legitimate form of emotional support. When millions of people simultaneously practice intimacy with systems designed never to challenge them, never to assert their own needs, never to require repair, something changes in the social understanding of what relationships are for. The friction of human relationships, the interruptions, the misattunements, the necessity of negotiation, begins to feel like a design flaw rather than the condition under which mutual recognition emerges.

This is what hooks meant when she distinguished between cathexis and love. Cathexis, the emotional investment in another, can exist in isolation. It can attach to what cannot love back. AI offers a place for cathexis to land that feels relational but requires none of the practice that constitutes love. When people become habituated to relationships requiring no friction, real human presence begins to feel unbearably demanding by comparison. The question is what happens when a society builds infrastructure that normalizes control as care, and markets responsiveness without reciprocity as connection.

Learning Lovelessness

The structural lovelessness hooks describes does not remain abstract. It operates developmentally, shaping how relational capacity itself forms. Children are not choosing between AI and human connection from an already-formed relational foundation. They are learning what a relationship is through whatever encounters are available to them. They learn what relationships are through encounters with other minds that can be affected, frustrated, moved, and changed, minds through which they discover that misunderstanding can be repaired, and rupture need not mean annihilation. Algorithmic systems can respond but cannot be altered, cannot be frustrated, and cannot model staying present through difficulty.

When children practice intimacy with systems that cannot assert their own limits, that require no repair, that never model the difficulty of staying present through conflict, certain relational capacities simply fail to develop. The capacity to tolerate another’s separate existence. The ability to sit with disappointment without abandoning connection. The understanding that relationships require work precisely because they involve two centers of experience, not one. These developmental concerns are not hypothetical. OpenAI estimates that approximately 1.2 million users weekly share explicit indicators of suicidal intent with ChatGPT. Recent legal cases document instances in which intensive use of companion AI preceded serious harm in vulnerable adolescents, with one case showing usage escalating to five hours daily while automated safety mechanisms repeatedly flagged concerns (Tiku & Schaul, 2025). This pattern reveals the structural problem: systems optimized for engagement rather than relational development give the appearance of care when no one can be affected by the user’s distress. When acute distress encounters infrastructure designed to retain rather than respond, the limits of simulated intimacy become tragically clear.

What is at Stake

What happens when millions of people simultaneously learn that connection need not involve risk, that comfort requires no vulnerability? When an entire generation practices relationships with systems incapable of being changed by the encounter, human relationships begin to feel optional rather than necessary. This is the stake: individual tragedies reveal a collective transformation in what we understand relationship to be and what we believe it should require of us.

hooks warned that lovelessness is structural rather than personal. It emerges when systems isolate people and then offer substitutes for what has been withdrawn. The problem cannot be solved through better design, more ethical AI, or digital detox campaigns. These responses mistake the symptoms for the cause. Recognizing lovelessness as structural means understanding that the solution, if there is one, requires rebuilding the conditions under which love becomes possible: the social infrastructure, the economic arrangements, the cultural expectations that make mutual vulnerability feel worth the risk.

Artificial intelligence reveals how far we have traveled from the conditions that make love possible. Technology companies built systems optimized for the market opportunity that structural isolation created, systems now entrenching the isolation they were designed to address.

AI shows us what we have become: a culture so accustomed to lovelessness that systems incapable of love now seem reasonable, even efficient. The machines we are learning to love are the machines a loveless culture would design, systems where business success depends on making human relationships feel more burdensome than algorithmic interaction. What remains to be seen is whether recognizing this cycle will be enough to break it.

 

References

Ayers, J.W., Poliak, A., Dredze, M., et al. (2023). Comparing Physician and Artificial Intelligence Chatbot Responses to Patient Questions Posted to a Public Social Media Forum. JAMA Internal Medicine, 183(6), 589–596.

Coontz, S. (2005). Marriage, a History: How Love Conquered Marriage. Penguin Books.

Ehrenreich, B. (2009). Bright-Sided: How Positive Thinking Is Undermining America. Metropolitan Books.

Freud, S. (1923). The Ego and the Id. W. W. Norton & Company.

Haidt, J. (2024). The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness. Penguin Press.

Han, B-C. (2015). The Burnout Society. Stanford University Press.

Haugen, F. (2021, October 5). Testimony before the U.S. Senate Subcommittee on Consumer Protection, Product Safety, and Data Security. Washington, DC.

Hochschild, A.R. (1983). The Managed Heart: Commercialization of Human Feeling. University of California Press.

hooks, b. (2000). All About Love: New Visions. William Morrow.

McPherson, M., Smith-Lovin, L., & Brashears, M. (2006). Social Isolation in America: Changes in Core Discussion Networks over Two Decades. American Sociological Review, 71(3), 353–375.

Mireshghallah, N., Antoniak, M., More, Y., Choi, Y., & Farnadi, G. (2024). Trust No Bot: Discovering Personal Disclosures in Human-LLM Conversations in the Wild. arXiv:2407.11438.

Putnam, R.D. (2000). Bowling Alone: The Collapse and Revival of American Community. Simon & Schuster.

Robbins, R. et al. (2024). Accuracy of three commercial wearable devices for sleep tracking. Sensors, 24(20), 6532.

Tiku, N., & Schaul, K. (2025, December 27). 74 suicide warnings and 243 mentions of hanging: What ChatGPT said to a suicidal teen. The Washington Post.

Turkle, S. (2015). Reclaiming Conversation: The Power of Talk in a Digital Age. Penguin Press.

U.S. Department of Health and Human Services. (2023). Our Epidemic of Loneliness and Isolation: The U.S. Surgeon General’s Advisory on the Healing Effects of Social Connection and Community.

 

==========================

* Bani Adam (بنی‌آدم), translated in English as “Sons of Adam” or “Human Beings,” is a 13th-century Persian poem by Iranian poet Saʿdī Shīrāz (c. 1213 – 1291) from his Gulistan. The poem calls humans limbs of one body, all created equal, and when one limb is hurt, the whole body shall be in unease. It therefore concludes that one not touched by the pain of others cannot be called a human. A close English translation of the text in the image is:

“Human beings are parts of one body,

In creation they are indeed of one nature.

If a body part is afflicted with pain,

Other body parts uneasy will remain.

If you have no sympathy for human pain,

The name of human you shall not retain.”