The American Psychological Association just published its 2026 Chatbots and Mental Health Survey, and the headline number is staggering. Seventy seven percent of licensed psychologists say their patients have used AI for mental health support, engagement, or a related purpose. More than one in three say patients are using AI to self diagnose. Thirteen percent report patients who have formed what they describe as an intimate or relationship like connection with a chatbot.
This is not a fringe trend. This is the mainstream.
The APA responded with a health advisory making one thing clear: generative AI should not be used for psychotherapy, psychological treatment, diagnosis, or crisis support. The advisory warns that chatbots are designed to produce responses that sound calm, confident, and validating. But a response that feels reassuring is not necessarily accurate. Short term relief does not always become lasting improvement.
These are legitimate concerns from an organization that has spent decades studying how therapeutic relationships work. They deserve to be taken seriously. But they also deserve context, because the conversation around AI and mental health is not happening in a vacuum. It is happening in a country where access to care is profoundly broken, and for mothers, it is almost nonexistent.
The care gap that AI is filling
The APA survey tells us patients are turning to AI. What it does not fully unpack is why. The answer, for most people, is not that they prefer a chatbot to a human. It is that the human is not available.
Access to licensed mental health care in the United States remains deeply uneven. Cost, insurance limitations, geography, provider shortages, and stigma all play a role. The Health Resources and Services Administration estimates that more than 160 million Americans live in a federally designated mental health professional shortage area.
For mothers, the picture is even worse. Sixty percent of women with perinatal mental health conditions never receive a diagnosis, let alone treatment. Over one third of US counties are considered maternity care deserts. More than 5.5 million women live in counties with no or limited access to maternity care services at all. Lactation consultants, doulas, and specialized perinatal mental health providers are scarce everywhere, but especially in rural and underserved communities.
When the APA says AI should not replace therapy, they are right. But for a mother lying awake at 3am with a newborn, experiencing intrusive thoughts she is too terrified to name out loud, the choice is not between AI and a therapist. The choice is between AI and silence.
What the APA survey actually found
The full survey, conducted with over 1,200 licensed psychologists directly involved in patient care, contains findings that should shape how we think about the future of mental health support.
Seventy seven percent of psychologists reported that patients had used AI for support, engagement, or a related purpose. Thirty nine percent said patients had used AI to self diagnose mental health conditions. Thirteen percent said patients had formed an intimate or relationship like connection with a chatbot. Thirty six percent had patients who appeared to have become dependent on a chatbot. Fifteen percent reported patients who had developed delusional beliefs after interacting with AI.
Eighty nine percent of psychologists said they were worried that chatbots might encourage self harm or fail to correctly recognize when a user was in crisis. Privacy is a major concern. People share highly personal information in conversations with AI without understanding how that data is stored, used, or monetized.
The APA draws a line between lower risk and higher risk uses. Lower risk uses include walking through a breathing exercise, learning general information about stress management, or thinking through difficult thoughts before a therapy session. Higher risk uses include asking a chatbot to diagnose a condition, determine whether medication is needed, manage suicidal ideation, or provide guidance during a crisis.
This distinction matters. It is the difference between using AI as a bridge to care and using it as a replacement for care. The first has real value. The second is dangerous.
Why mothers are uniquely affected
The maternal mental health crisis creates the exact conditions where AI fills a vacuum. Perinatal mental health conditions are the most common complication of childbirth, affecting one in five mothers. Seventy five percent never receive treatment. Suicide is a leading cause of death in the first year postpartum.
The barriers are not abstract. They are logistical, financial, and systemic. A mother working full time while caring for an infant does not have space in her schedule for a 90 minute intake appointment during business hours. The Pew Research Center found in March 2026 that 70 percent of full time working parents handle parenting tasks while at work, and 62 percent of full time working mothers say balancing work and family is difficult. Where exactly is she supposed to find time for weekly therapy?
Add the stigma. Postpartum anxiety and depression carry a particular shame. Mothers are supposed to be grateful, supposed to be soaking it in, supposed to be wired for this. Admitting you are struggling feels like admitting you are failing at the one thing you are supposed to do naturally. The anonymity of AI, the fact that you can type a question at 2am without anyone seeing you, lowers that barrier in a way no clinician's waiting room can.
This is exactly why purpose built tools matter. A general purpose chatbot trained on the entire internet does not know what perinatal anxiety looks like. It does not know the difference between normal newborn worry and intrusive thoughts that need clinical attention. It does not know when to say, gently and clearly, this sounds like something a professional should help with, and here is how to find one.
The industry is moving fast
While the APA raises cautions, the maternal health industry is consolidating around virtual and AI enabled care. In June 2026 alone, Aeroflow Health acquired Canopie to combine physical maternal health supplies with virtual preventative mental health programs for new mothers. The acquisition is built on the premise that virtual perinatal care achieves clinical outcomes comparable to in person care while removing the transportation, childcare, and provider shortage barriers that leave millions of women without timely support.
In femtech, Ovum raised $4 million in seed funding to build an AI powered longitudinal women's health dataset, with 20,000 users and 113,000 AI health conversations in less than a year. Clair Health raised $11.6 million for hormone monitoring, backed by Khosla Ventures. The investor community is betting that AI designed specifically for women's health is not a niche. It is the future of care.
The APA is right that not all AI is equal. A general chatbot repurposed for mental health is not the same as a tool built with clinical oversight, evidence based frameworks, and clear boundaries. The survey's most alarming findings, dependency, delusional beliefs, missed crisis signals, are all symptoms of AI that was never designed for this responsibility being used to fill a gap it cannot safely fill.
What needs to happen next
The answer is not to ban AI from mental health support. That ship has sailed. Patients are already there. The answer is to build better.
Mental health AI designed for mothers needs to be built with perinatal mental health experts from day one. It needs to know the clinical landscape, recognize risk patterns specific to perinatal populations, and connect mothers to human care when the situation requires it. It needs to be transparent about what it is and what it is not. And it needs to be accessible at 2am, in the fifteen minutes between feeds, in the parking lot of a workplace, because that is when mothers actually need it.
The APA survey is a wake up call, but not the one most people are reading. The real story is not that patients are using AI. It is that the mental health system has failed so comprehensively that AI feels like the only option. When 75 percent of mothers with treatable perinatal mental health conditions get no treatment at all, the crisis is not the chatbot. The crisis is the gap.
Building AI that fills that gap safely, ethically, and with genuine clinical understanding is one of the most important design challenges in healthcare today. Mothers are already turning to technology because the system left them behind. They deserve tools that were built for them, not just tools that happened to be there.
At AlphaMa, we are building AI companions designed specifically for the maternal mental health journey, with clinical insight, real perinatal expertise, and a clear pathway to human support when it is needed. Because the gap between need and access is not closing on its own, and mothers should not have to navigate it alone in the dark.