AI Session-Recording Tools Spread in Therapy, Outpacing Patient Consent
Summary
- • Therapists across the U.S. are adopting AI tools that auto-record sessions and generate clinical notes
- • Patient consent practices are lagging: at least one case where recording began before the patient agreed
- • Only 11% of Americans are open to AI in mental health care; 77% worry about health data storage
- • HIPAA compliance is necessary but insufficient — affirmative patient consent is not yet standard practice
Details
AI session-recording tools are spreading in therapy practices, marketed as reducing administrative burden
Berries, SimplePractice, and Blueprint offer AI platforms that auto-record therapy sessions, generate transcripts, and draft clinical notes. Companies charge $19–$99/month and target solo practitioners overwhelmed by paperwork, arguing the tools let clinicians be more present with clients.
Consent practices are failing: recording can begin before patients have meaningfully agreed
NPR documented the case of Molly Quinn, 31, who asked to research the AI tool before her therapist used it. Mid-session, she noticed the iPad 'just propped up' and realized the session was being recorded. She described the experience as feeling 'completely violated,' saying her therapist 'disregarded something I said I was not comfortable with.' The article notes Quinn had not explicitly refused — she had requested time to research — but felt the consent threshold had been breached.
Only 11% of Americans open to AI in mental health care; 8% trust it; nearly half reluctant
A YouGov national survey found Americans are broadly resistant to AI in therapy specifically: 11% open, 8% trusting, nearly half actively reluctant. Concerns center on lack of human understanding, inaccurate advice, and privacy. A separate KFF survey found 77% of Americans worried about how health information would be stored and used by AI systems.
HIPAA was not designed for AI-generated transcripts; consent requirements for recording are not specifically governed
Berries and similar vendors operate under HIPAA compliance for data storage, but the regulatory framework does not specifically address AI recording consent in therapeutic contexts. Until clearer standards emerge, consent obligations fall entirely on individual clinicians — a standard that current practice suggests is uneven.
Vendors are marketing aggressively to cost-pressured solo practitioners; patients have no independent way to verify privacy claims
The AI therapy-notes market is growing, with vendors making assurances (no training on session content, immediate audio deletion) that patients cannot independently verify. The business case — reducing hours of weekly administrative work — is credible, but the consent and trust gap represents reputational and regulatory exposure for both vendors and clinicians.
Industry Update = market adoption trends; Insight = attributed analysis from reported cases; Research = survey/study findings; Legal = regulatory and compliance context; Market Impact = business and risk implications
What This Means
For AI practitioners and health-tech builders, this is a warning that deployment speed without consent infrastructure creates reputational and regulatory exposure — especially in high-sensitivity domains like mental health. HIPAA compliance is necessary but not sufficient: patients increasingly expect affirmative, informed consent for any AI involvement in their care. Professionals advising healthcare organizations on AI adoption should treat consent workflow design as a first-class technical and ethical requirement, not an afterthought.