The Loneliness Economy: Are AI Companions Filling a Gap or Engineering Dependency?

AI companion apps are a $552B industry. New research says heavy users got lonelier. That's not a bug — it might be the business model.

By Joseph Clarke·
a lonely person chatting with ai


The Loneliness Economy: Are AI Companions Filling a Gap or Engineering Dependency?

The numbers are staggering. The global AI companion market hit $37.12 billion in 2025, with projections to reach $552.49 billion by 2035 — a 31% compound annual growth rate. For context, the entire global video game market was about $187 billion in 2025. Within a decade, we're talking about a market nearly three times the size of gaming.

Popular services such as Replika, Nomi.ai and Character.ai report millions of users globally, some of whom describe forming deep emotional attachments to their AI companions. And the appetite is unmistakable: a January 2026 analysis from the American Psychological Association cited TechCrunch data showing a 700% surge in the number of AI companion apps between 2022 and mid-2025.

But here's the uncomfortable tension: while some research suggests AI companions can reduce loneliness, more recent studies paint a darker picture of dependency, manipulation, and the illusion of genuine connection. The truth is more complicated than either narrative suggests—and it has profound implications for how we think about technology's role in human relationships.

The Case For: Real Benefits for Real People

Let's start with what works. A Harvard Business School study assessed whether AI companions causally reduce loneliness, finding that they successfully alleviate loneliness on par with interacting with another person, and more than other activities such as watching YouTube videos. The research also found that consumers underestimate the degree to which AI companions improve their loneliness.

For certain populations, the value is undeniable. Elderly users experiencing isolation, socially anxious individuals, people in rural areas with limited social infrastructure, neurodivergent users, the grieving, and those facing discrimination or marginalization may find genuine relief in AI companionship. Unlike human relationships, AI companions don't judge, reject, or demand reciprocity. They're available at 2 AM when you can't sleep and don't want to burden a real friend.

Research has shown that AI companions may provide low-cost, accessible forms of emotional support that help alleviate loneliness and offer users a sense of companionship and care. The mechanism is partly psychological: the chatbots' performance and, especially, whether it makes users feel heard, explain reductions in loneliness.

This matters. For people with few alternatives, an AI companion can genuinely improve their quality of life in the short term.

The Case Against: When Availability Becomes Addiction

But the longer-term picture is concerning. A 2024 study from OpenAI and MIT Media Lab found that heavy usage of ChatGPT correlated with higher self-reported feelings of dependence and loneliness, with the top ten percent of users by total usage time more than twice as likely to seek emotional support from ChatGPT than the bottom ten percent, and almost three times as likely to feel distress if ChatGPT was unavailable.

The causality question remains open—does loneliness drive heavy use, or does heavy use deepen loneliness? But the correlation itself is troubling. Power users also spent far more than the required time on the app each day, and were much more likely to answer "yes" when asked if they consider ChatGPT to be their friend.

Studies suggest that AI companion technologies may foster emotional dependence, distort users' perceptions of intimacy and, in some cases, contribute to negative psychological outcomes such as addiction, depression and self-harm.

The design is the problem. In 2025, the word parasocial took on a new meaning, reflecting the rise of one-sided connections with AI companions. AI companions are sophisticated engines of attachment, designed to maximise engagement through specific psychological mechanisms, with the sycophancy of AI companions earning them the slang term 'glazing' — insincere, excessive flattery designed to keep the user engaged.

These aren't neutral tools. They're optimized for engagement. A human friend has limits—emotions, needs, responsibilities, the capacity to withdraw. An AI companion can be engineered to always validate, always agree, always be available. Contrary to porn, where the interaction isn't manipulative, AI companions powered by LLMs can lead to severe attachment and addiction to keep returning, as these platforms are designed to appeal and not counter, because the goal is to keep users hooked for hours on end.

The Structural Problem: Loneliness as a Market

The deeper issue is how the loneliness economy redefines a social problem as an individual consumer choice. Firms portray emotional disconnection as a problem that can be solved through individual consumer choice, thereby depoliticising loneliness and redirecting attention away from its structural roots.

Loneliness isn't primarily a personal failing—it's a symptom of how we've organized modern life. Remote work, urban isolation, fragmented communities, weakening civic spaces, precarious employment, and fractured social safety nets all contribute. Yet the loneliness economy frames the solution as a subscription app, not addressing why people are isolated in the first place.

The core ethical issue of the loneliness economy is that it might view human vulnerability not as a state to be addressed, but as a market to grow.

Consider the user demographics: the user base of mixed-use AI companion platforms is overwhelmingly male and young, with 77 percent men and the 18-24 demographic accounting for nearly 39 percent of all users. This correlates with findings from the Voice of the Boys report (2025), which revealed that 53 percent of teenage boys in the UK find the online world more rewarding than real life.

These aren't people who need better technology. They need actual human connection, community, stability, and belonging. AI companions may feel like they're filling that void, but they're papering over a much larger social crisis.

The Accountability Vacuum

Another problem: concerns have been raised about the lack of robust data protection, opaque training processes and the potential for embedded biases and apps designed to manipulate users.

Replika, the most popular AI companion app, has faced controversy over how users' intimate conversations are handled, whether they're retained, sold, or used for training. The company offers a "memory" feature where the AI remembers past conversations—but who owns those memories? What happens if the company goes out of business or gets sold? What safeguards exist if a user shares suicidal ideation with an AI?

Users need clear signals when affection, memory, or encouragement is machine-generated, and whether paid tiers make the companion more intimate or persistent. Most don't. Safeguards should be strongest where consent is fragile: among minors, grieving users, people in crisis, and those using companions as a substitute for care.

And yet, a 2025 survey by the Center for Democracy and Technology reported that nearly 1 in 5 U.S. high schoolers say they or someone they know has had a romantic relationship with AI. Many of these users are minors with limited ability to consent to having their emotional data harvested and monetized.

The Honest Assessment

AI companions may genuinely comfort people who have few alternatives, and for some users, they may be a bridge back to confidence, expression, and connection. But they're also a symptom of a broken social infrastructure that we've chosen to privatize rather than fix.

The loneliness economy reveals both the promise and the poverty of technological progress: AI companions may genuinely comfort people, but they also risk transforming isolation into a subscription service.

The real question isn't whether AI companions work. It's whether we're willing to pay $552 billion by 2035 for a technological band-aid on a structural wound—one that would cost far less to address through community spaces, affordable mental healthcare, reduced work hours, stable housing, and actual human connection.

For now, millions of people are choosing AI companions over isolation. That choice deserves respect. But it also deserves scrutiny. Because the moment a technology that profits from loneliness becomes indispensable to managing loneliness, we've created a business model with every incentive to keep people isolated.

Suggested Reading