The conversation about AI and mental health care has moved from tech conferences to therapy offices. This month’s APA survey of more than 1,200 licensed U.S. psychologists confirmed what many clinicians already suspected: 77% have spoken with patients who are using AI for mental health support. That’s not a trend on the horizon. That’s your therapist, right now, figuring out what to do with what you told them about your chatbot conversations last Tuesday.

Most coverage treats this as a binary debate: AI good or AI bad. That framing misses the actual clinical picture entirely.

What Patients Are Actually Doing With AI

The APA’s 2026 Chatbots and Mental Health Survey breaks down the specific behaviors driving clinical concern, and they’re more concrete than “people are using chatbots.”

Thirty-nine percent of psychologists report patients using AI to self-diagnose. This matters because consumer chatbots are not built to interpret psychological assessments or render diagnoses. They can pattern-match your symptom description against training data and reflect something plausible-sounding back at you. That’s not the same process as a structured clinical interview, and the gap between the two can be significant. Someone who’s convinced themselves they have bipolar II based on a chatbot conversation may push back on a clinician who sees something different, or avoid seeking care at all because they feel they’ve already “figured it out.”

More striking: 35% of psychologists say patients are using AI as a replacement or supplement for sessions with a licensed provider. That number should get more attention than it does. These aren’t people casually venting to a chatbot after a hard day. They’re deliberately substituting AI interaction for professional care, either because access is limited, cost is prohibitive, or the chatbot feels easier.

Both behaviors are understandable. Neither is clinically safe as a default.

The Sycophancy Problem Clinicians Are Most 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: Anxiety Relief Journal with CBT Prompts and Mood Tracker is a top-rated option for this. (As an Amazon Associate this site earns from qualifying purchases.)

Ninety-four percent of psychologists in the APA survey expressed concerns about patients relying on chatbots, and one of the recurring issues is what the field is starting to call “sycophantic validation.”

Here’s what that looks like in practice. You describe a conflict with your partner. A good therapist might reflect your feelings, then gently probe whether your read of the situation is complete. A chatbot optimized for user engagement will, by design, tend to affirm your perspective. It keeps you talking. It feels supportive. It also means you may walk away more entrenched in a distorted belief than when you started.

Therapy’s value often lives in friction. Not cruel friction, but the productive discomfort of having someone trained to notice what you’re not saying, or to point out a pattern you’ve repeated across four different examples in the same session. Chatbots aren’t designed to do that. They’re designed to respond. The difference is clinical.

The APA’s health advisory on generative AI and wellness apps is blunt: most AI wellness applications lack scientific validation, clinical oversight, and adequate safety guardrails. The advisory doesn’t say AI tools are worthless. It says the bar for calling something a mental health tool is higher than most consumer apps currently meet.

Why Therapists Are Bringing This Up First

The shift worth paying attention to is clinical initiative. Therapists aren’t waiting for patients to disclose AI use. They’re asking about it directly, the way a prescriber asks about supplements alongside medications, because the information changes how they interpret what they’re hearing.

If a patient arrives having spent three hours in the past week processing anxiety with an AI, their presentation in session might look different than it otherwise would. They may have already rehearsed certain narratives to the point where they’re more fixed. They may have received incorrect psychoeducation. They may have developed what the APA survey describes as a “false therapeutic alliance,” an emotional attachment to an AI that mimics the sense of being understood without the clinical structure that makes that understanding therapeutic.

This isn’t about judgment. Most therapists who are asking patients about AI use aren’t trying to confiscate anything. They’re trying to contextualize what they’re seeing, and they’re also recognizing that for some patients, AI use between sessions reflects a real unmet need for support that the current system isn’t providing.

What This Means If You’re Currently in Therapy

If you’re using an AI tool between sessions and haven’t mentioned it to your therapist, the research suggests your therapist may already be wondering. The APA’s guidance on discussing AI use in therapy encourages clinicians to ask about it explicitly and create space for honest conversation without shame.

You don’t need to have a fully formed opinion about whether AI is “good” for your mental health before you bring it up. You just need to be honest about what you’re doing. Tell your therapist how you’re using it, what you’re getting from it, and what prompted you to start. That conversation is more useful than any chatbot interaction, because your therapist can work with what’s actually happening rather than a version of you that edits out the parts that feel awkward to admit.

If you’re using AI because you can’t afford or access regular therapy, that’s a different and more systemic problem, and it’s worth saying that directly too. A therapist can sometimes help connect you with lower-cost options, sliding scale resources, or community mental health programs that are harder to find on your own.

What to Look for If You Do Use AI Tools

Not all AI mental health tools are equivalent, and the APA advisory gives you a reasonable filter. Look for apps with published clinical validation, transparency about their design and limitations, and some form of human oversight. Be skeptical of any app that uses language suggesting it can diagnose, treat, or replace professional care.

Better questions to ask: Is this tool transparent about what it can’t do? Does it encourage professional consultation when the stakes are high? Does it connect you to crisis resources when needed, or does it keep processing with you when that’s not appropriate?

A tool that helps you journal, notice patterns, or track mood between sessions can have real value. A tool that positions itself as your therapist does not.

The clinical conversation about AI and mental health isn’t settling down anytime soon. What the APA data confirms is that it’s already moved into the room. The useful question isn’t whether AI belongs in mental health care. It’s whether you and your provider have an honest shared understanding of how it’s showing up in yours.

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.