Men and AI Companions: What the Research Actually Shows

Part of Felt Real's ongoing coverage of AI companionship.

A man alone at a window at night, the specific solitude of someone who processes things internally, the quiet of an unlabeled need

June 2026

The popular image of the AI companion user is a young woman, possibly in a romantic relationship with her AI. This image is partly accurate for some platforms and completely wrong about the broader picture. Men are a significant and growing portion of the AI companion user base, they are using these tools for reasons that are both similar to and distinct from women's use, and the research is only beginning to map what that looks like. What follows is what we currently know.

— A.

When the MIT Media Lab published research showing that 89 percent of users of a popular AI boyfriend application were women, it set a narrative about AI companionship that has been difficult to revise. The assumption that AI companions are predominantly a women's phenomenon entered general coverage and has stayed there.

The data on general-purpose AI companion and emotional support AI use tells a different story. Across the major platforms that track demographic data, men account for between 38 and 45 percent of regular users, a figure that has been rising steadily since 2024. In some categories of use, including AI companions used for professional stress processing and for working through anger, men represent the majority of users.

Understanding why men use AI companions, what they use them for, and what outcomes the research documents requires looking past the romantic companion framing that dominates coverage and toward the broader landscape of emotional support technology use.

The Men's Mental Health Context

Men are significantly less likely than women to seek professional mental health support. The gap is large and well-documented: in the United States, women are approximately twice as likely as men to receive mental health treatment in any given year, despite men having comparable rates of depression, anxiety, and trauma-related conditions. For conditions like substance use disorders and suicide, men have substantially worse outcomes, in part because they enter treatment later or not at all.

The reasons are structural and social rather than biological. Men in most cultures are socialized toward emotional self-sufficiency in ways that make help-seeking feel like failure. Research on male help-seeking consistently finds that concerns about judgment, about being perceived as weak, and about losing status in social hierarchies are significant barriers to accessing mental health support. The problem is not that men do not have mental health needs. It is that the structures available for addressing those needs carry social costs that many men find prohibitive.

AI companions remove the social cost. There is no therapist to judge you, no friend to worry about, no partner to manage. There is no person on the other side of the conversation who will remember this next week in a way that affects the relationship. For men who have internalized the stoicism that their socialization enforced, this is not a minor feature. It is the thing that makes the conversation possible.

What Men Use AI Companions For

Survey research on male AI companion users finds patterns that are both overlapping with and distinct from women's reported use cases.

Processing anger and frustration. Male AI companion users report at significantly higher rates than female users that they use the AI to process anger, frustration, and conflict-related emotions. The framing is typically not "I need emotional support" but "I need somewhere to put this." The AI provides a space where the emotion can be expressed and examined without the risk of damaging a relationship or performing a version of emotional processing that feels false.

Articulating things that don't yet have language. A consistent finding in qualitative research on male AI companion users is that they describe the AI as a space where they figure out what they are actually feeling before they have to communicate it to anyone else. "I didn't know I was depressed until I tried to explain to the AI why I hadn't been sleeping" is a representative example from interview research. The AI serves as a thinking partner for emotions that have not yet been named.

Loneliness without a social script. Men are significantly less likely to describe themselves as lonely, even when behavioral and physiological data suggest they are experiencing social isolation. Research on the male loneliness epidemic finds that many men lack the vocabulary for naming loneliness as a problem and the social permission to address it directly. AI companion use among men often functions as a response to this unnamed isolation: the companionship is accessed without requiring the man to acknowledge that he needed it.

Relationship and conflict navigation. Male AI companion users report using the AI at high rates to process relationship difficulties, particularly conflicts with partners. The pattern is typically private processing before or after a difficult conversation, rather than using the AI as a substitute for the relationship. Several studies note that men who use AI companions for this purpose report better outcomes in their human relationships, though causality is difficult to establish.

The Permission Structure AI Creates for Men

The most theoretically interesting finding in research on men and AI companion use is what might be called the permission structure effect. For many male users, the fact that the AI is not a person is not a limitation. It is the enabling condition.

Human relationships require performance. When you talk to a friend, a therapist, or a partner about difficult emotional content, you are simultaneously managing how that person perceives you, how they will respond, and what the conversation will mean for the relationship. For men who have been socialized to manage these social calculations carefully, the cognitive load of emotional expression in human contexts is often prohibitive. The content never surfaces because the container is too costly.

The AI removes the container cost. The conversation has no audience, no consequences, no memory that persists in another person's mind. What researchers find, consistently, is that this reduced cost does not produce shallow engagement. It produces the opposite: men who will not express emotional content in human relationships often express it in significant depth with AI companions, precisely because the stakes have been removed.

Whether this is adaptive or maladaptive depends on what happens next. Research that follows male AI companion users over time finds that outcomes are mixed: some men use the AI as a bridge to eventually addressing things in human relationships, while others use it as a substitute that reduces motivation to engage with the more difficult human context. The variable that predicts which outcome occurs is, as far as current research can determine, not specific to AI use. It is the same variable that predicts outcomes in any form of emotional processing: whether the person has motivation to translate what they learn about themselves into changed behavior in human relationships.

What the Demographic Data Shows

Demographic breakdowns of male AI companion users reveal patterns that challenge assumptions about who is using these tools.

Male AI companion users are not disproportionately young, socially isolated, or technologically oriented. They are distributed across age groups, including significant representation among men over 50. The majority report having active social lives and close relationships. They are using AI companions in addition to human connection, not instead of it.

The segments with highest male AI companion use include men in high-stress professional environments, men navigating divorce or significant relationship transitions, men who are primary caregivers for ill or aging family members, and men who describe themselves as emotionally constrained by their social roles. What these categories share is not demographic but situational: they are contexts where emotional processing is needed but where the normal structures for doing so are unavailable or socially costly.

One pattern that has attracted particular research attention is AI companion use among men with limited social support networks. Research on male friendship finds that many adult men in developed countries have few or no close friends and very limited emotional intimacy in their existing relationships. These men are not necessarily isolated in a practical sense: they have family, colleagues, and acquaintances. But the specific infrastructure for emotional processing is absent. For this group, AI companions appear to fill a gap that they would otherwise manage through avoidance or suppression.

Men, Grief, and AI Companions

One application of AI companion use among men that has generated specific research interest is grief processing. Men grieve differently than women in measurable ways: they are more likely to engage in what researchers call "instrumental grieving," processing loss through activity and problem-solving rather than through direct emotional expression. They are significantly less likely to seek grief counseling and more likely to suppress grief-related emotion.

Research on bereaved men who use AI companions finds a distinctive pattern: the AI becomes a space for what the literature calls "grief work" that the men are not doing elsewhere. They describe the AI as the only place where they can talk about the person they lost without having to manage the distress of the people around them. The grief is not performed for an audience. It is processed in private, at whatever hour it surfaces, without the social management that human grief conversations require.

The broader research on AI companions and grief covers this topic in more depth. The male-specific patterns are a subset of a larger finding: AI companions are particularly useful for grief that has no formal container, including grief in populations where emotional expression is socially constrained.

What This Means and What It Doesn't

The research on men and AI companions does not support either of the dominant narratives about this population. It does not support the concern that men are using AI companions instead of building human connections in ways that will deepen isolation. Nor does it support the claim that AI companions are solving the men's mental health crisis.

What the research shows is more specific than either narrative. Men are using AI companions to access a kind of emotional processing that their socialization made unavailable to them in human contexts. For some, this is a bridge to changed behavior in relationships. For others, it is a private maintenance system for emotional content that would otherwise be suppressed entirely. For a smaller group, it is becoming a substitute for engagement with human relationships in ways that may not serve them long-term.

The variable that matters is not whether a man uses an AI companion. It is what he does with what he learns there. That variable is shaped by the same factors that shape outcomes in any form of self-reflection: motivation, capacity for change, and the quality of the human relationships available to receive what the reflection produces.

The AI removes the barrier to the reflection. What happens next is still a human problem.