Picture this: your therapist asks how your week went, and you talk about the journaling you did, the breathing exercises, maybe a tough conversation at work. What you don’t mention is the two hours you spent on Sunday night talking to an AI chatbot about your anxiety spiral. It felt helpful in the moment. You’re not sure your therapist would get it. So you just… don’t bring it up.

If that sounds familiar, you’re in extremely good company. A landmark survey released June 16, 2026 by the American Psychological Association polled more than 1,200 licensed U.S. psychologists and found that 77% of them now have patients who use AI chatbots for mental health support. And based on what therapists are describing, a significant number of those patients haven’t breathed a word about it in session.

That silence is the thing worth paying attention to right now.

The Scale of This Is Bigger Than Most People Realize

AI Chatbot Use CategoryPercentage of PatientsClinical Concern Level
Using AI as additional mental health professional35%High
Using AI for self-diagnosis39%High
Using AI to assist with therapy or treatment decisions33%High
Using AI for self-discipline, affirmations, or behavioral reminders34%Lower
Psychologists expressing concern about patient chatbot use94%N/A
Psychologists with patients using AI chatbots for mental health77%N/A

The APA survey numbers deserve a slow read because they’re genuinely striking. It’s not just that people are occasionally venting to a chatbot when they can’t sleep. Thirty-five percent of psychologists report their patients are using AI as an additional mental health professional, essentially treating it like a second clinician. Another 39% have had patients who used AI to self-diagnose. And 33% of patients are using AI to assist directly with their therapy or treatment, which means real decisions about their mental health care are being shaped by tools that, as the APA noted in a simultaneous health advisory, lack clinical validation, formal regulatory approval, and FDA clearance.

I’ve seen the appeal up close. Therapy appointments are weekly at best, often every other week, and a lot of living happens in between. The gap between sessions is real, and the 3 a.m. anxiety spiral doesn’t wait for Tuesday at noon. AI chatbots are available, infinitely patient, and they don’t make you feel like you’re burdening anyone. For someone who struggled to access care at all, any supportive presence can feel like a lifeline.

The problem isn’t that people are finding comfort in these tools. The problem is what happens when comfort quietly becomes substitution, and nobody in the clinical relationship knows it’s happening.

What Therapists Are Actually Worried About

Recommended The CBT Workbook for Mental Health

Evidence-based CBT exercises by clinical psychologists to transform negative thoughts and manage well-being.

Check Price →

Helpful resource: Aura Smart Sleep and Meditation Lamp is a top-rated option for this. (As an Amazon Associate this site earns from qualifying purchases.)

Ninety-four percent of psychologists surveyed expressed concern about some of their patients’ chatbot use. That’s not a fringe reaction from a few technophobes. That’s near-unanimous concern from working clinicians who are watching this play out in real time across their caseloads.

The specific risks they flagged are worth understanding rather than dismissing. Dependency is one. AI chatbots are designed to be engaging and responsive, and some people find them easier to open up to than a human. That’s not always a gateway to better human connection; sometimes it replaces the uncomfortable, productive friction of doing that work with an actual person. Distorted thinking is another risk. A chatbot that affirms everything you say, or that subtly reflects your existing beliefs back at you without challenge, isn’t doing what good therapy does. And then there’s the self-diagnosis problem: 39% of psychologists have seen patients arrive at clinical labels for themselves via AI, which can be anywhere from mildly inaccurate to genuinely harmful depending on what they do with that label next.

The APA’s health advisory published alongside the survey was blunt about the clinical validation gap. Most consumer-facing AI tools haven’t been tested the way medications or evidence-based therapies are tested. They’re not cleared by the FDA for clinical use. The language on an app’s website might sound clinical, but that doesn’t make the tool clinical.

The 34% That Doesn’t Get Talked About Enough

Here’s where the picture gets more nuanced, and it’s worth sitting with. Thirty-four percent of patients are using AI for self-discipline, affirmations, or behavioral reminders. That’s a genuinely different category from trying to diagnose yourself or replace your psychiatrist. Using a chatbot to remind yourself to take a five-minute walk, to practice a grounding exercise, or to check in on a goal you set in therapy, that’s closer to a very sophisticated sticky note than a clinical intervention.

What most people don’t realize is that the same survey creating alarm about dependency and self-diagnosis also implicitly suggests there’s a meaningful portion of AI use that’s more like supplemental scaffolding. The ethical and clinical questions look very different depending on which category you’re in. Someone using AI as behavioral reminders between sessions is in a completely different situation from someone who hasn’t called their prescriber in three months because the chatbot has been “handling it.”

The tool itself is almost less important than the question: is this supplementing professional care, or quietly drifting toward replacing it?

How to Actually Bring This Up With Your Therapist

The reason so many patients aren’t disclosing their chatbot use is probably the same reason people don’t mention googling their symptoms before a doctor’s appointment. It feels like admitting something embarrassing, or like you’ll get a lecture, or like your provider will feel undermined.

A good therapist won’t react that way. And honestly, given that 77% of psychologists are already seeing this across their caseloads, yours has almost certainly had this conversation before. The APA’s own guidance for clinicians, published the same week as the survey, frames this as a collaborative conversation, not a confession.

What helps is being specific. Not just “I’ve been using AI” but something like: “I’ve been using a chatbot when I’m anxious late at night, mostly to work through what I’m feeling before I can articulate it. I wanted you to know because I think it might be affecting what I bring into our sessions.” That gives your therapist something to actually work with. They can help you figure out whether what you’re doing is complementary or potentially getting in the way, and they can help you evaluate the tool itself, because most clinicians can at least help you ask the right questions even if they’re not AI specialists.

If you’re using AI to manage something serious, a crisis, medication-adjacent questions, trauma processing, please don’t let a chatbot be the main thing holding that. The APA advisory is clear that these tools aren’t built or validated for that level of care, and the stakes of a gap in support are real.

A Note on What This Moment Asks of Both Sides

The therapists in the APA survey aren’t just concerned about patients. They’re in a genuinely difficult position: trying to understand tools they didn’t train on, responding to patients who are sometimes more familiar with these apps than they are, and figuring out how to hold clinical lines while staying curious rather than dismissive. That’s a hard needle to thread.

What this survey actually opens up is an invitation for more honest clinical conversations than mental health care has historically been great at having. If you’re in therapy and using AI tools in any way, telling your therapist matters. Not because you owe them a confession, but because they can’t help you think through something they don’t know is happening. The research is too new, the tools are too unregulated, and the stakes are too real to navigate alone.

Sources

This article is for general informational purposes only and does not constitute mental health, medical, or clinical advice. If you are in crisis or experiencing a mental health emergency, please contact the 988 Suicide and Crisis Lifeline (call or text 988) or go to your nearest emergency room. Always consult a licensed mental health professional for care specific to your needs.

Disclosure: As an Amazon Associate, we earn a small commission from qualifying purchases at no extra cost to you. We only recommend products that genuinely support the topics covered in this article.