A growing number of therapists are adopting artificial intelligence tools to automate clinical note-taking, creating tension between administrative relief and patient privacy concerns. These systems record sessions, generate transcripts, and draft clinical notes automatically—offering potential time savings but raising trust issues for patients who expect absolute confidentiality.
Companies marketing AI note-taking platforms, such as Berries, SimplePractice, and Blueprint, charge between $19 to $99 per month. They emphasize convenience for solo practitioners overwhelmed by documentation. Berries’ Co-CEO Tal Salman told NPR the system is designed to reduce administrative burden without interfering with therapy, with audio processed in real time and deleted immediately afterward. However, the experience of patient Molly Quinn illustrates the trust hazard. After discussing her therapist’s plans to use AI and expressing uncertainty, Quinn discovered during her next session that recording had begun without her explicit consent. “I felt completely violated,” she said afterward, noting that the breach of trust fundamentally damaged her therapeutic relationship.
Privacy protections alone do not guarantee safety. Kellie Owens, an assistant professor of medical ethics at NYU Grossman School of Medicine, cautions that HIPAA compliance does not eliminate breach risk. “Regardless of what protections we have in place, that doesn’t mean data can’t be breached,” she said. Healthcare systems and major corporations have experienced repeated data breaches in recent years. Additionally, written consent forms may create a false sense of agreement—research shows people often scroll through consent documents without reading them or feel pressured to agree.
Clinical concerns extend beyond cybersecurity. Marisa Cohen, a couples and sex therapist in New York, observes that introducing AI fundamentally alters the therapeutic experience: “Clients know or feel like something else is listening to them. That awareness can subtly alter their disclosure.” She also flags accuracy risks; if AI-generated errors go undetected and notes are later subpoenaed, inaccurate information becomes part of a patient’s permanent record.
Some clinicians report substantial benefits. Kym Tolson, a fully remote therapist, says the AI system reduced her documentation time from 15-20 minutes per client to about 2 minutes. She emphasizes thorough client discussion beforehand and careful review of AI-generated notes for hallucinations. Research in other healthcare settings supports time savings—one large analysis found clinicians saved nearly 16,000 hours of documentation annually. However, only 11% of Americans surveyed by YouGov said they would be open to AI in mental health care, and just 8% trust the technology, highlighting the persistence of patient hesitation.
This article is an AI-assisted summary. All facts and figures are drawn from the original report: https://www.npr.org/2026/05/26/nx-s1-5826943/talk-therapy-mental-health-ai-artificial-intelligence-privacy-trust