Imagine discovering your therapist has been using AI to analyze your sessions without your knowledge. It’s not just a breach of trust—it’s a glimpse into the controversial intersection of technology and mental health. But here’s where it gets controversial: as AI tools like ChatGPT and note-taking apps infiltrate therapy rooms, clients like Molly Quinn are left questioning the ethics of their use. And this is the part most people miss: it’s not just about convenience—it’s about consent, privacy, and the very essence of the therapeutic relationship.
Molly, a 31-year-old librarian, had been in therapy for two years without issue—until her therapist introduced a new AI-powered note-taking app. The app promised to record, transcribe, and summarize sessions, but Molly hesitated. ‘I’m not a big fan of AI,’ she admits. ‘I understand the challenges of note-taking in therapy, but I needed time to research how my private data would be handled.’ Her concerns were valid: while the app claimed to delete data after transcription, the summarization process meant her information would linger in some AI cloud—a prospect that made her uneasy.
But her therapist didn’t wait for Molly’s consent. Midway through a session, Molly noticed her therapist using an iPad instead of taking handwritten notes. ‘It left me feeling violated,’ Molly recalls. She confronted her therapist, who admitted to using the app without permission. ‘My trust was broken,’ Molly says. ‘I no longer see her.’
Molly’s experience isn’t isolated. The National Alliance of Counsellors and Psychotherapists (NACP) in the UK has received numerous complaints this year about therapists using AI—whether for note-taking, transcribing, or even drafting emails. ‘Therapists are experimenting with AI, often without proper disclosure or consent,’ says Meg Moss, head of public affairs at the NACP. ‘Clients are understandably uncomfortable.’
Here’s the kicker: Reddit threads are flooded with similar stories. One client was shocked during a Zoom session when their therapist accidentally shared their screen, revealing they were using ChatGPT to formulate responses. Photographer Brendan Keen shared his experience on Medium, detailing how his BetterHelp therapist used AI to reply to his messages, leaving him feeling betrayed and worried about confidentiality.
While some therapists argue AI is a useful tool, others warn of its pitfalls. Ranjith Devakumar, a therapeutic counsellor, uses AI to brainstorm ideas or research techniques but never shares client details. ‘It’s a soundboard, not a replacement for human judgment,’ he explains. Ruby Mitchell, an accredited therapist, uses AI sparingly—as a refresher, not a teacher. She’s cautious, fearing AI could trap clients in reassurance loops or replace crucial supervisor interactions.
But here’s the real question: should therapists rely on AI at all? Richard Miller, an ethics consultant with the British Association for Counselling and Psychotherapy (BACP), is drafting guidelines to address this. ‘I wouldn’t want a counsellor using ChatGPT for diagnosis,’ he says. ‘And if you outsource note-taking, don’t outsource critical thinking.’ He emphasizes the need for consent, data security, and ethical use: ‘I don’t want the first generation of AI users to be guinea pigs for confidentiality.’
Molly’s story highlights two moral issues: lack of consent and the questionable use of AI itself. ‘I don’t think my therapist understood the implications,’ she says, worrying for less tech-savvy clients. Meg Moss agrees, urging therapists to prioritize transparency and caution: ‘AI can add uncertainty to the therapeutic relationship. Unless absolutely necessary, I’d avoid it.’
Now, here’s the controversial part: Is AI in therapy a helpful innovation or a risky shortcut? Should therapists be required to disclose its use, even for minor tasks? And what happens if AI-driven mistakes lead to ethical dilemmas? Let’s spark a conversation—share your thoughts in the comments. Are you comfortable with AI in therapy, or does it cross a line? The future of mental health care may depend on how we answer these questions.
