When Molly Quinn realized her therapist had been using an AI tool to record her sessions, she didn’t just feel surprised. She left. After two years of building trust with that therapist, the disclosure came mid-session rather than upfront, and that was enough. NPR’s Morning Edition covered her story on May 26, 2026, and the response was immediate: therapy forums lit up, mental health Twitter argued in circles, and a lot of patients started quietly wondering what was happening inside their own sessions.

This is the conversation the mental health field has been putting off, and it’s now unavoidable.

What’s Actually Happening in the Room

AI Use in Therapy PracticePercentageSource
Psychologists who have used AI at least once56%APA 2025 Practitioner Pulse Survey (March 2026)
Psychologists who use AI monthly29%APA 2025 Practitioner Pulse Survey (March 2026)
Psychologists concerned about privacy implications67%APA 2025 Practitioner Pulse Survey (March 2026)

AI scribing tools, designed to record, transcribe, and auto-generate clinical notes, have moved from novelty to common practice faster than most patients realize. According to the APA’s 2025 Practitioner Pulse Survey, released in March 2026, 56% of psychologists have used AI in their practice at least once, and 29% use it monthly. Those numbers jumped significantly from 2024. The APA responded by releasing specific guidelines for evaluating AI scribing technologies in March 2026, which tells you something: this escalated fast enough to require formal professional guidance within a single year.

The appeal for therapists is real and worth acknowledging. Documentation is a genuine burden. Clinical note-writing after sessions takes time that could go toward other patients or, frankly, toward the therapist’s own mental health. A tool that transcribes a session and drafts a SOAP note automatically has obvious practical value. That doesn’t make the consent problem disappear, but it does explain why adoption is outpacing the ethics conversation.

The Data Problem Is More Specific Than “Privacy Concerns”

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: First, We Make the Beast Beautiful by Sarah Wilson is a top-rated option for this. (As an Amazon Associate this site earns from qualifying purchases.)

Here’s what most coverage glosses over: when a HIPAA-compliant AI tool records a therapy session, the audio doesn’t just stay between you and your therapist. It gets transmitted to a private tech company’s servers, transcribed, and processed there. Your trauma history, your diagnoses, the thing you finally said out loud for the first time, all of it sits with a third party you’ve never heard of and almost certainly never consented to share with.

“HIPAA-compliant” is doing a lot of work in those vendor marketing materials. HIPAA compliance means a company has agreed to certain data handling standards. It doesn’t mean your data is invulnerable, it doesn’t mean it can’t be used for model training under certain contract terms, and it doesn’t mean you have meaningful visibility into who at that company can access your file. Patients in online therapy communities, including Reddit, have been raising exactly these questions since at least 2025, often getting vague answers or none at all.

Two thirds of psychologists surveyed by the APA, 67%, reported remaining concerned about the privacy implications of AI tools in clinical settings. That’s the people using the tools flagging the problem. That number should get more attention than it does.

The Quinn case in NPR’s reporting crystallized something important: the issue isn’t only whether AI is used, it’s how and when patients are told. There’s a significant difference between a therapist who discusses AI note-taking during intake, explains what the tool is, names the third-party vendor, and gets written acknowledgment versus one who starts recording and mentions it when the patient notices a small icon on the screen.

The APA’s March 2026 guidelines address informed consent explicitly, which is good. What’s murkier is enforcement. Licensing boards vary by state. Informed consent standards in mental health care are already inconsistent. A guideline from the APA carries professional weight but isn’t law, and a therapist who buries AI recording disclosure in a standard intake form’s dense paragraph isn’t technically violating it.

The practical question for patients isn’t “is this ethical in the abstract?” It’s “did my therapist tell me clearly, before we started, exactly what is being recorded, where it goes, and who can access it?” If you’re not sure, that’s worth asking directly.

What You Can Reasonably Do Right Now

You have the right to ask your therapist whether any AI tools are used to record or transcribe your sessions. You have the right to ask which company hosts that data, whether you can opt out, and what happens to recordings after notes are generated. A therapist who responds defensively to those questions is giving you information.

If you’re starting with a new therapist, this belongs in your first-session checklist alongside questions about their approach and fees. The GBH/TPR coverage from May 2026 noted that some therapists are now proactively addressing AI use in their intake process, partly in response to this public conversation. That’s the right direction, and it’s reasonable to expect it.

If you’ve already been in therapy and this is new information, take a breath before assuming the worst. Many therapists using these tools do have consent language somewhere in their paperwork. The problem is often opacity and timing rather than deliberate deception. A direct conversation with your therapist about what specifically is recorded and stored is usually the right first step before deciding whether to stay or go.

And if you’re considering leaving, that’s a valid choice. The therapeutic relationship depends on trust more than almost any other professional relationship you have. As AllAboutCookies reported in June 2026, patients are increasingly factoring data practices into their therapy decisions, the same way they might ask a new doctor about their records system.

The Broader Stakes

The mental health field has spent years working to reduce barriers to care, to make therapy feel safer, more accessible, and less stigmatized. AI note-taking tools, handled well, could genuinely help with that by reducing therapist burnout and administrative load. Handled badly, they erode the one thing therapy actually runs on.

The APA’s 67% of worried psychologists aren’t Luddites. They’re practitioners who understand that a patient disclosing abuse or suicidal ideation or addiction in a session is extending an enormous amount of trust, and that trust has a data trail now. The profession is figuring out how to hold both things at once: the efficiency gains and the ethical weight. Patients deserve to be part of that conversation, not surprised by it mid-session.

If you’re in therapy or considering it, you don’t need to become an AI policy expert. You do need to feel like you know what happens in your sessions. That’s not a high bar. It’s the baseline.

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.