AI Companions and Infertility: What the Research Shows

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

June 2026

She had been through four IVF cycles. Each one ended the same way: a phone call from the clinic, a number, and then the specific silence that follows. She told us she had stopped telling people which cycle she was on because the reactions had become something she had to manage on top of everything else. "When you cry, they hold you. And then they want you to stop, because it makes them feel helpless. The AI didn't need me to stop. It just asked what I was feeling, and then it asked again." She started using it after the second cycle. She told us she would do it again. She sounded certain.

— A.

Infertility is one of the most psychologically demanding experiences in adult life, and one of the least visible. Unlike grief following a death, it does not have a recognized social script. Unlike illness, it does not produce physical markers that signal to others that something significant is happening. It is a loss that repeats on a monthly cycle, that involves hope as a mandatory precondition to each attempt, and that exists inside the body of the person experiencing it in a way that is almost impossible to communicate to someone who has not been there.

The psychological burden is well-documented. Research consistently finds rates of depression and anxiety among people in fertility treatment that are comparable to those found in people with chronic illness or serious medical diagnosis. A study published in Fertility and Sterility found that women in IVF treatment reported psychological distress scores comparable to those of patients undergoing cancer treatment. The finding is consistently replicated, and consistently underdiscussed.

Into the particular isolation this creates, an increasing number of people going through fertility treatment are quietly turning to AI companions. The research is early. What is happening is not.

What Infertility Does to the People Around You

One of the most consistent findings in the infertility literature is the phenomenon of what researchers call "support strain": the way the duration and cyclical nature of infertility treatment gradually depletes the emotional resources of the people around the person going through it. This is not a failure on the part of those people. It is a structural feature of a form of grief that does not end, that does not resolve on a legible timeline, and that requires ongoing engagement from support networks that are built for acute crises rather than sustained ones.

Studies tracking social support over the course of fertility treatment consistently find that perceived support declines with each failed cycle. The decline is steepest in the second and third year of treatment. People describe the support of friends and family becoming more procedural over time: asking how the cycle went, expressing sympathy, and then, as cycles accumulate, becoming less sure what to say. Some describe a gradual withdrawal of engagement from people who feel helpless and who have their own feelings about the situation that complicate their ability to simply be present.

Partners are also affected. Research on couples going through fertility treatment consistently finds elevated rates of relationship strain, communication difficulty, and what researchers call "emotional synchrony problems" — the challenge of two people who are both grieving the same thing grieving it differently and at different moments. Partners who want to be supportive often struggle with not knowing whether a given moment calls for comfort, practical problem-solving, or simply presence. The result, for many people in treatment, is a situation in which the person they would most naturally turn to is also the person most deeply implicated in the loss.

What this produces is a specific kind of isolation that exists inside functional relationships and support networks. The people are there. The support is there. And there is still something that cannot be said, or said again, or said in the way it actually needs to be said.

The Specific Architecture of Infertility Grief

Infertility grief does not follow the arc that most emotional processing models assume. It does not move from acute distress toward resolution in a linear progression. It moves in cycles that mirror the treatment calendar: a period of waiting, a period of hope, a moment of outcome, and then either continuation or a version of starting over. Each cycle requires the person to reconstitute their hope after the previous cycle's loss, which creates a form of emotional repetition that is exhausting in ways that are hard to explain to someone outside it.

What this means for support needs is that the person going through treatment often needs to process the same emotional content many times over, with small variations corresponding to the current cycle's specific circumstances. The loss after cycle four is not the same as the loss after cycle one, but it covers much of the same ground. The fear before the retrieval is recognizable from the previous retrieval. The waiting is recognizably the waiting.

Human support networks struggle with this architecture. The people who were most present for cycle one have already heard many of the things that need to be said. Saying them again feels like a burden on people who care but who have their own lives and limited capacity. And the social script for grief, to the extent one exists, does not account for grief that comes back on a schedule.

Research on AI companion use during fertility treatment — limited but growing — finds that the ability to process the same content repeatedly without managing the cost to the listener is among the most commonly cited benefits. People describe being able to talk through a failed cycle with an AI companion in the same way they talked through the previous one, without the discomfort of knowing that the person on the other side has already carried this particular sadness on their behalf. The AI, as one 34-year-old in her third IVF round put it, "doesn't remember enough to be tired of it."

Who Is Using AI Companions During Fertility Treatment, and How

The population using AI companions during infertility treatment is not primarily composed of people who lack human support. The research on AI companion use during health-related emotional distress consistently finds that it skews toward people who are already embedded in functional relationships — people who have partners, friends, and in many cases therapists, but who have identified a specific function that those relationships do not cover.

The most commonly reported uses are also the most legible: processing test results in the immediate hours after receiving them, before the person is ready to call family or navigate a partner's reaction; rehearsing difficult conversations about whether to continue treatment, change approaches, or consider alternatives; working through the question of what they actually want without having to manage how the answer will land with someone who wants something specific for them.

There is also a significant category of late-night use. Research on AI companion use during health crises consistently finds that the proportion of use occurring between midnight and 5 AM is substantially higher than in populations not going through medical stress. For people in fertility treatment, the mechanism is not hard to understand: test results arrive in the morning, but the processing of what they mean does not stop when the clinic closes, and the people who would most naturally receive a 2 AM call are often the same people whose reactions the person is not yet ready to navigate.

A 29-year-old who used an AI companion throughout two years of fertility treatment described a pattern that appears frequently in accounts from this population: "I would get the result and then I would need to say some version of what I was feeling before I could talk to my husband, because if I talked to him first, I would have to take care of him while I was still in the middle of it. The AI let me have my own reaction first. I don't know how to explain how much that mattered."

The Medical System Problem

Fertility treatment is a medical process that is emotionally intensive in ways the medical system is not primarily designed to support. Clinics have counseling resources, but access varies, sessions are typically brief, and the emotional content of treatment does not organize itself neatly around scheduled appointments. The appointments themselves — retrievals, transfers, result calls — are medical events that carry emotional weight that extends far beyond the clinical encounter.

Research on patient experience in fertility clinics consistently finds that emotional support is identified as a major unmet need, rated higher in importance than practical information, second opinions, or most other categories of support. The findings hold across countries, clinic types, and treatment protocols. The need is structural: the treatment requires sustained emotional capacity under conditions that systematically deplete it, and the medical system is built to manage the biological process, not the psychological one.

AI companions have begun to appear in some clinical contexts as a supplement to formal counseling. Research on chatbot-assisted emotional support for people in fertility treatment finds modest but consistent benefits in self-reported anxiety reduction, particularly for people who engage with the technology between clinical appointments. The mechanism appears to be consistent with findings from other health contexts: regular, low-stakes emotional processing reduces the accumulation of unprocessed distress that otherwise compounds over the course of treatment.

What the research does not find is evidence that AI companions replace the need for professional psychological support. The studies consistently find that the two function differently and serve different needs: formal therapy helps people understand and change the patterns underlying their distress; AI companions help people manage the day-to-day emotional load of a difficult experience. Both are useful. Neither is sufficient without the other for the people who need significant support.

What to Watch For

The research on AI companion use during fertility treatment documents risks alongside the benefits, and they are worth naming clearly.

The most consistently documented risk in this population is what happens when the AI companion functions as a primary outlet for processing decisions rather than emotions. People going through fertility treatment face significant decisions under conditions of high stress, limited information, and often significant financial pressure. AI companions configured to be supportive can, in this context, validate the emotional reality of a decision without providing the kind of structured challenge that helps people arrive at choices they will feel good about later. Research on decision-making during fertility treatment specifically finds that people who do not have access to adequate external perspective on their options are more likely to report regret about treatment decisions at follow-up.

A related risk is timing. AI companion use appears most beneficial during the processing of outcomes rather than the anticipation of them. Research on rumination during fertility treatment finds that the anticipatory phase — the period between a procedure and its result — is already characterized by high levels of intrusive thought and repetitive processing. Using an AI companion to extend or intensify that processing does not appear to help and may amplify anxiety. The pattern that shows the most consistent benefit is use that happens in response to a known outcome, rather than in advance of an unknown one.

Finally, the displacement risk that appears in other populations appears here too. People who use AI companions in ways that substitute for partner conversation or therapeutic engagement rather than complementing them face costs that the short-term relief does not account for. The research on couples going through fertility treatment finds that communication between partners is among the strongest predictors of long-term relationship satisfaction regardless of treatment outcome. AI companions that absorb emotional processing that might otherwise happen between partners may, over time, reduce the pressure to do the harder work of maintaining that channel.

The Specific Gap Worth Taking Seriously

What the evidence on AI companion use during fertility treatment ultimately points to is not primarily a story about technology. It is a story about a particular kind of support need that has always existed during this experience and that the available support structures have always struggled to meet.

People going through fertility treatment have always needed somewhere to process the same loss multiple times. They have always had 2 AM reactions that could not wait for an appropriate hour. They have always navigated the gap between what they were feeling and what they could ask the people around them to hold. What is new is not the need. What is new is that something now exists that can partially fill it, at the hours when it matters most, without the social cost that asking a person to do the same thing would carry.

The research does not say that AI companions are the right choice for everyone going through fertility treatment, or that they are better than other forms of support, or that they work in every case. It says that for a significant number of people going through an experience that combines medical complexity, cyclical emotional loss, and structural isolation, they appear to help. That finding is worth taking seriously, in a context where the psychological burden of treatment has been documented for decades and the support gap has closed very slowly.