Beyond Transparency: The Ethical Complexities of Open Notes and AI-Assisted Documentation in Mental Health

Reece Gorrie – Lewis Katz School of Medicine – Temple University; Brian Tuohy – Assistant Professor and Co Director of Education, Lewis Katz School of Medicine, Temple University

Abstract: The implementation of Open Notes under the 21st Century Cures Act mandates transparency by requiring clinicians to make documentation accessible to patients. Existing literature largely endorses Open Notes for promoting patient autonomy, trust, and improved outcomes. We contend that Open Notes should be patient-centered, written in plain language for the sake of health literacy, and recognize cultural context. However, this is a significant burden. AI-assisted documentation tools (e.g., Eleos, Lyssn) offer a promising solution by easing clinicians’ workloads through automated note creation. Yet, these tools introduce critical and previously underexplored ethical problems. Using the AI tool Eleos as a specific case study, this presentation argues that AI-generated documentation risks undermining epistemic justice and therapeutic pluralism. AI's emphasis on empirically validated, standardized therapeutic approaches may unintentionally marginalize patient narratives, cultural perspectives, and individualized treatment goals. Further, reliance on AI risks epistemic injustice by representing patient behavior through algorithmic generalizations that misrepresent patient intentions. Drawing from interdisciplinary insights in bioethics, sociology, philosophy, and psychiatry, we articulate practical recommendations for ethically responsible use of AI-assisted documentation. Specifically, we propose a framework in support of AI-assisted notes alongside clinician engagement to critically review, contextualize, and, when necessary, modify or reject AI-generated notes. These claims depend on the principled preservation of therapeutic diversity, patient autonomy, and clinician discretion. Our framework offers guidance for responsibly integrating AI into clinical practice, underscoring the need to ensure transparency and patient-centered care without compromising therapeutic integrity.

Keywords: AI-assisted Documentation, Open Notes, Therapeutic Pluralism